FEASIBILITY AND DESIGN CONSIDERATIONS

FOR THE USE OF LIFTS AS AN EMERGENCY

EXIT IN APARTMENT BUILDINGS

BY

Than Singh Sharma

BSc, BE (Fire), MSc (Disaster Mitigation), CFES, MIFE (UK)

Queensland University of Technology

Brisbane, Australia

A THESIS SUBMITTED TO THE SCHOOL OF URBAN DEVELOPMENTS

QUEENSLAND UNIVERSITY OF TECHNOLOGY IN PARTIAL

FULFILLMENT OF REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

March 2008

iii

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted for a degree or

diploma for any other higher education institution to the best of my knowledge and

belief. The thesis contains no material previously published or written by another

person except where due reference is made.

Signed:

THAN SINGH SHARMA

Date: 20.03.2008

v

ABSTRACT

Emergency evacuation in high-rise apartment building is a challenge for fire safety

professionals. Lift evacuation is a controversial issue because the safe operation of

lift is not ensured under the existing design and operating conditions. Lifts are not

permitted for public evacuation in apartment buildings during fire emergencies as per

the provisions of building codes and regulations. However, the concept of using lifts

for emergency evacuation has been gaining considerable attention during recent

years.

The lift evacuation can be considered as an alternative facility if it is efficient,

reliable and readily accessible. It can also provide a safer means of evacuation for the

aged and disabled persons, who may not be able to evacuate promptly, efficiently

and unassisted using the exit stairs during fire emergencies. Moreover, lifts can

enable building corporate management to easily and promptly access the fire-

affected floor and commence fire fighting.

The work on the use of lifts for emergency evacuation was initiated in the early

1990s at the National Institute of Standards and Technology (NIST, USA) in which

pros and cons were analysed in order to develop suitable guidelines. This research

project examines the feasibility of using lifts along with design modifications as an

alternative facility for a safer and more efficient emergency evacuation. The scope of

this research is limited to apartment buildings where occupant load is low and fire

load is generally confined to dwelling compartment units.

This research project analysed the important issues in relation to the use of lifts for

emergency evacuation. The issues were divided into three categories: human

behavioural response, fire hazards and lift operational mechanism. Output variables

relating to human behavioural response were modelled and analysed as a stochastic

process. Residents’ choice for using evacuation routes was determined using a

survey. The issues of fire hazards (fire, smoke and toxic gases) were analysed for

occupant safety under variable conditions using the concept of fire safety index. The

issues of lift operational mechanism such as lift malfunctioning due to excessive

vi

temperature, electric power failure and water damage were considered for developing

probabilistic models. An integrated approach of risk assessment for the issues of

human behavioural response and fire hazards (such as ‘decision uncertainty’, ‘panic’,

‘nonfatal and fatal injuries’) was developed based on the Multi-Objectives Decision

Analysis method. The results for lift and stair systems were compared and the

feasibility of using lift with design modifications was analysed for alternative designs

and evacuation strategies.

The outcomes of this research have shown that using lifts with a protected lobby for

up to one-fourth of the building population (who may be aged and disabled) has huge

potential as an alternative evacuation facility with enhanced level of safety. Lifts

with protected lobby for one-fourth of the building population showed an improved

level of fire safety from exposure to fire effluents. The reliability of lift operational

mechanism is also improved in protected lift shafts. Lifts with protected lobby for up

to one-fourth of the building population and stairs for up to three-fourth of the

building population showed an improved evacuation safety. The risks in combined

evacuation systems (protected lifts and stairs) are found to be lower when compared

to using stairs or protected lifts. Lifts with double lobby protection (for example, two

levels of compartmentation with fire and smoke doors for lift lobby) showed further

improvements.

This research has proposed alternative designs for lifts and developed models for

analyzing evacuation effectiveness based on risks related to human behaviour, fire

hazards and operational mechanism. It has shown that a combined use of lifts and

stairs has significant advantages. The performance based lift evacuation system is

achievable in apartment buildings. These research findings are based on uncertainty

analysis, which can be further extended to other types of buildings in the future.

vii

ACKNOWLEDGEMENT

I take this special opportunity to express my deep sense of gratitude to Prof. Mahen

Mahendran, Dr. Yaping He, A/ Prof. Gopi Chattopadhya and Mr. B. (Jack)

Williamson, my supervisors, whose invaluable guidance, support and encouragement

have nourished this research project. My wholehearted thanks to my supervisors.

The time spent with Dr. Yaping He at University of Western Sydney is also a

priceless treasure for me, who so patiently and cheerfully goaded and helped me to

complete this research project.

I am particularly thankful to Mr. Christopher R. Odgers, Principal Fire Safety

Engineer from Fire Check Consultants, Queensland for his guidance and immense

assistance related to computer modelling and literature in all phases of this research

project. I also convey my special thanks to Mr. David Mason, University of Southern

Queensland for providing guidance on developing ARENA simulation models. The

officers from NSWFS and QFRS involved in the survey and interviews were fully

supportive at all times.

I wish to acknowledge Mr. R. C. Sharma, Chief Fire Officer for approving my study

leave from Delhi Fire Service for undertaking this research project. I also

acknowledge Mr. S. K. Dheri, Retired Chief Fire Officer, Delhi Fire Service, for

providing me support and guidance.

I express my gratitude and appreciation to my father and my mother for their

unlimited love, support and sacrifices. I acknowledge and recognize my wife Manju

for her moral support and encouragement in all phases of my life and my sons

Aayush and Archit for whom I couldn’t spare my time on many important occasions

during this research tenure.

viii

CONTENTS

STATEMENT OF ORIGINAL AUTHORSHIP .................................................... III

ABSTRACT ................................................................................................................V

ACKNOWLEDGEMENT ...................................................................................... VII

CONTENTS ............................................................................................................VIII

LIST OF FIGURES................................................................................................XIII

LIST OF TABLES..................................................................................................XVI

ABBREVIATIONS/ DEFINITIONS ....................................................................XIX

NOMENCLATURE ...............................................................................................XXI

1. INTRODUCTION .................................................................................................1

1.1 BACKGROUND ................................................................................................1 1.2 PROBLEM STATEMENT....................................................................................3 1.3 RESEARCH STATEMENT (HYPOTHESIS)...........................................................5 1.4 OBJECTIVES OF RESEARCH .............................................................................5 1.5 SCOPE .............................................................................................................6 1.6 RESEARCH METHODOLOGY ............................................................................7 1.7 THESIS CONTENTS ..........................................................................................8

2. LITERATURE REVIEW ...................................................................................10

2.1 CURRENT REGULATIONS IN RELATION TO THE USE OF LIFTS FOR

EVACUATION .............................................................................................................11 2.2 RECENT DEVELOPMENTS IN THE SUBJECT AREA ..........................................13 2.3 STATISTICS ...................................................................................................15 2.4 HUMAN BEHAVIOUR DURING FIRE ...............................................................19 2.5 FIRE AND SMOKE HAZARDS..........................................................................22 2.6 TOXICITY OF FIRE EFFLUENTS ......................................................................29

2.6.1 Mass Loss Models ...................................................................................29 2.6.2 Toxic Gas Models....................................................................................30 2.6.3 Human Incapacitation Model ..................................................................32 2.6.4 Visibility ..................................................................................................33

2.7 LIFT OPERATIONAL SYSTEMS .......................................................................33 2.7.1 Overview .................................................................................................33 2.7.2 Lift Dispatch Control...............................................................................34 2.7.3 Concerns of Lift Operational Mechanism ...............................................35 2.7.4 Lift Protection Measures .........................................................................36 2.7.5 Sprinklers in Residential Buildings .........................................................36 2.7.6 Evaluation of Lift Evacuation Time ........................................................38

2.8 STAIR EVACUATION SYSTEM........................................................................39 2.8.1 Overview .................................................................................................39 2.8.2 Evaluation of Stair Travelling Time........................................................39

2.9 RISK ASSESSMENT METHODS .......................................................................41 2.10 APPLICATION OF COMPUTER MODELS IN THE RESEARCH STUDY .................44 2.11 DISCUSSION AND SUMMARY.........................................................................46

ix

3. RESEARCH METHODOLOGY....................................................................... 49

3.1.1 Risk Identification ................................................................................... 50 3.1.2 Analytical Hierarchical Process for Risk Priorities................................. 53 3.1.3 Acceptable Level of Risk ........................................................................ 62 3.1.4 Hypothetical Building and DTS Provisions ............................................ 63 3.1.5 Selection of a Fire Scenario..................................................................... 65 3.1.6 Concept Design Options.......................................................................... 68 3.1.7 Risk Quantification.................................................................................. 71 3.1.8 Risk Assessment...................................................................................... 71 3.1.9 Selection of Design Options.................................................................... 72

3.2 RESEARCH WORK......................................................................................... 72 3.3 CONCLUSION ................................................................................................ 75

4. STOCHASTIC MODELS OF BUILDING EVACUATION .......................... 76

4.1 INTRODUCTION ............................................................................................. 76 4.2 PILOT SURVEY.............................................................................................. 77

4.2.1 Pilot Survey Overview ............................................................................ 77 4.2.2 Pilot Survey Results ................................................................................ 78 4.2.3 Discussion and Conclusion ..................................................................... 81

4.3 INTERVIEWS ................................................................................................. 83 4.4 ANALYSIS OF BUILDING EVACUATION PERIODS........................................... 85

4.4.1 Methodology ........................................................................................... 85 4.4.2 Discrete Event Simulation....................................................................... 89

4.5 MODEL FRAMEWORK ................................................................................... 91 4.5.1 Hypothetical Building and Parameters.................................................... 91 4.5.2 Lift Supervisory Controller for Lift Simulation Model........................... 93 4.5.3 Lift Simulation Variables ........................................................................ 95 4.5.4 Stair Simulation Variables..................................................................... 100

4.6 SIMULATION MODELS ................................................................................ 103 4.6.1 Lift Simulation Model ........................................................................... 103 4.6.2 Stair Simulation Model ......................................................................... 104

4.7 SIMULATION RESULTS ................................................................................ 105 4.7.1 Lift Simulation Model ........................................................................... 106 4.7.2 Stair Simulation Model ......................................................................... 108

4.8 ANALYSIS OF RESULTS ............................................................................... 110 4.8.1 Lift Waiting Time.................................................................................. 110 4.8.2 Lift Transportation Time and Stair Travelling Time............................. 112 4.8.3 Lift Pre-Evacuation Time and Stair Pre-Evacuation Time.................... 114 4.8.4 Lift Evacuation Time and Stair Evacuation Time................................. 116 4.8.5 Number of Evacuees in Queue .............................................................. 118 4.8.6 Findings ................................................................................................. 119

4.9 MODEL VERIFICATION ............................................................................... 120 4.10 CONCLUSION .............................................................................................. 121

5. FIRE HAZARD MODELS OF LIFE THREATENING CONDITIONS .... 123

5.1 INTRODUCTION ........................................................................................... 123 5.2 ANALYSIS OF FIRE HAZARDS ..................................................................... 124

5.2.1 Effects of Fire Effluents and Evaluation Criteria .................................. 124 5.2.2 Safety Index........................................................................................... 129 5.2.3 Safety Index for Three Locations.......................................................... 131

x

5.2.4 Field Model ‘FDS’.................................................................................133 5.3 MODEL FRAMEWORK AND VARIABLES.......................................................133

5.3.1 Hypothetical Building Model ................................................................133 5.3.2 Concept Designs ....................................................................................136 5.3.3 FDS Model Boundary Conditions .........................................................136 5.3.4 Fire Simulation Scenarios......................................................................138

5.4 FDS MODEL SET UP...................................................................................139 5.4.1 Conventional Domain and Grid System................................................139 5.4.2 Smoke Leakages/ Openings...................................................................141

5.5 FDS OUTPUT ..............................................................................................142 5.6 FDS RESULTS .............................................................................................149

5.6.1 Concept Design A (Unprotected Lift Lobby)........................................149 5.6.2 FDS Results Analysis ............................................................................165

5.7 FED OF SMOKE, ASPHYXIANT AND HEAT ..................................................170 5.7.1 Concept Design A (Unprotected Lift Lobby)........................................170 5.7.2 Summary of FED Results and Analysis ................................................177

5.8 DETERMINATION OF SAFETY INDEX ...........................................................181 5.8.1 Strength Variables ASET ......................................................................181 5.8.2 Load Variables RSET............................................................................182 5.8.3 Safety Index...........................................................................................183 5.8.4 Safety Index Analysis............................................................................183

5.9 CONCLUSION ..............................................................................................184

6. RELIABILITY OF LIFT OPERATIONAL MECHANISM ........................185

6.1 INTRODUCTION ...........................................................................................185 6.2 ANALYSIS OF LIFT OPERATIONAL MECHANISM..........................................186 6.3 METHODOLOGY ..........................................................................................187 6.4 LIFT MALFUNCTIONING DUE TO EXCESSIVE TEMPERATURE RISE...............189 6.5 ELECTRIC POWER FAILURE.........................................................................195

6.5.1 System Descriptions ..............................................................................195 6.5.2 System Boundary Conditions ................................................................196 6.5.3 Data and Statistics .................................................................................196 6.5.4 Fault Tree Analysis................................................................................198 6.5.5 Analysis of Results ................................................................................201

6.6 PROBABILISTIC ANALYSIS OF WATER DAMAGE .........................................202 6.6.1 Water Spread from Fire Protection and Fire Fighting Measures...........203 6.6.2 Complex Parallel and Series System for Water Spread ........................205 6.6.3 Water Spread Result Analysis ...............................................................209

6.7 OUTCOMES .................................................................................................210 6.8 INFLUENCE ON HUMAN BEHAVIOURAL RESPONSE .....................................210 6.9 CONCLUSION ..............................................................................................211

7. RISK ASSESSMENT OF EVACUATION ROUTES....................................212

7.1 INTRODUCTION ...........................................................................................212 7.2 RISK ANALYSIS OF BUILDING EVACUATION SYSTEM .................................213

7.2.1 Assumptions ..........................................................................................213 7.2.2 Methodology – Multi-Objectives Decision Analysis ............................214

7.3 RISK ASSESSMENT......................................................................................216 7.3.1 Identify Concept Design Options and Evacuation Strategies................216 7.3.2 Evaluation Considerations and Evaluation Measures ...........................216

xi

7.3.3 Specify Weights .................................................................................... 218 7.3.4 Value Functions..................................................................................... 222 7.3.5 Sensitivity Analysis ............................................................................... 229

7.4 ANALYSIS OF RESULTS ............................................................................... 231 7.5 CONCLUSION .............................................................................................. 232

8. FEASIBILITY AND DESIGN CONSIDERATIONS.................................... 234

8.1 INTRODUCTION ........................................................................................... 234 8.2 FEASIBILITY OPTIONS................................................................................. 235

8.2.1 Lifts with Protected Lobby to evacuate 25% of the Building Population235 8.2.2 Double Protected Lift Lobby for the Entire Building Population ......... 236

8.3 REDUNDANCY MEASURES .......................................................................... 237 8.3.1 Common Lift and Stair Lobby .............................................................. 237 8.3.2 Refuge Area........................................................................................... 238 8.3.3 Scattered Design of Lift System............................................................ 239 8.3.4 Pressurization ........................................................................................ 240 8.3.5 Smoke Seal in Lift Landing Door ......................................................... 240

8.4 FIRE PROTECTION MEASURES FOR LIFT SYSTEM........................................ 241 8.5 STRATEGIC PLANNING................................................................................ 242 8.6 CONCLUSION .............................................................................................. 242

9. CONCLUSIONS................................................................................................ 244

9.1 SUMMARY .................................................................................................. 244 9.2 RESEARCH FINDINGS .................................................................................. 247

9.2.1 Advancements in Systematic and In-Depth Risk Analysis ................... 248 9.2.2 Contribution to Building Evacuation Strategy ...................................... 248

9.3 SCOPE FOR FUTURE WORK ......................................................................... 249

REFERENCES ........................................................................................................ 251

APPENDIX A ............................................................................................................ 263 International Listing of Major Fires, where Lifts were used............................. 263

APPENDIX B ............................................................................................................ 264 Risk Priorities and Matrix Consistency Ratio ................................................... 264

APPENDIX C ............................................................................................................ 266 Survey Questionnaire ........................................................................................ 266

APPENDIX D ............................................................................................................ 271 Interview............................................................................................................ 271

APPENDIX E............................................................................................................. 275 Occupant Response and Coping Times............................................................. 275

APPENDIX F............................................................................................................. 276 SIMAN Language ............................................................................................. 276

APPENDIX G ............................................................................................................ 283 ARENA Results ................................................................................................ 283

APPENDIX H ............................................................................................................ 304 Verification of ARENA Model ......................................................................... 304

APPENDIX J ............................................................................................................. 308 Building Characteristics, HRR and Temperature.............................................. 308

APPENDIX K............................................................................................................ 311 Occupants’ Movement Causing Door Opening and Closing ............................ 311

APPENDIX L............................................................................................................. 313 Visibility Determination at a Focal Point.......................................................... 313

xii

APPENDIX M............................................................................................................315 Species Concentration and Fractional Effective Doses of Smoke, Gases and Heat ...............................................................................................................315

APPENDIX N ............................................................................................................367 Calculation of Fractional Effective Doses of Smoke, Gases and Heat .............367

.

xiii

LIST OF FIGURES

Figure 1-1 – Fire and Rescue from Pallister Plaissance Apartments, USA .................3

Figure 2-1 – Fire Shaft and Stairs Protected with Lift Lobby (NBS, 2000) ..............12

Figure 2-2 – Fire Fighter Lift is utilized up to one level below the Fire-Affected

Floor (CEN, 2003)..............................................................................................13

Figure 2-3 – Factors Causing Death during Residential Fires (Miller, 2005)............18

Figure 2-4 – Stress Model (Proulx, 1993) ..................................................................20

Figure 2-5 – Fear induced Panic Behaviour ...............................................................21

Figure 2-6 – Piston Effect: Lift acts as a Piston for Smoke Movement .....................25

Figure 2-7 – Smoke Infiltration and Heat Exposure to Lift Landing Door................26

Figure 2-8 – Lift Door after the Fire Test (Bennetts et al., 1999) ..............................26

Figure 2-9 – Temperature inside the Lift Shaft (Bennetts et al., 1999) .....................27

Figure 2-10 – LMR Protected with Sleeves (smoke dissipates through vent) ...........28

Figure 3-1 – Risks involved in the Lift Evacuation System.......................................52

Figure 3-2 – Risks at Three Hierarchical Levels........................................................57

Figure 3-3 – Hierarchical Relationship for the Evaluation of Risk Priorities ............59

Figure 3-4 – Global Risk Priorities ............................................................................61

Figure 3-5 – Typical Floor of a Hypothetical Building (57 m ×××× 20 m)......................63

Figure 3-6 – Positive Pressurisation and Smoke Lobby in a Fire Isolated Exit .........65

Figure 3-7 – Event Tree Analysis for a Worst Possible Path (or Fire Scenario)........67

Figure 3-8 – Comparison of Stairs and Lifts ..............................................................69

Figure 3-9 – Three Concept Designs for risk analysis ...............................................70

Figure 3-10 – Research Work Flow Diagram ............................................................73

Figure 4-1 – Residents’ Awareness of Emergency Evacuation Procedure ................80

Figure 4-2 – Residents’ Experience during Fire Drill ................................................81

Figure 4-3 – Flow Diagram for Analysing the Output Variables of Models .............88

Figure 4-4 – SIMAN Flow Diagram ..........................................................................91

Figure 4-5 – Typical Floor of a Hypothetical Building (57m ×××× 20m)........................92

Figure 4-6 – Lift Controller Logic Diagram used for ARENA Simulation Model....94

Figure 4-7 – Poisson Distribution for Occupant Arrival (Lifts).................................99

Figure 4-8 – Poisson Distribution for Occupant Arrival (Stairs) .............................101

Figure 4-9 – Triangular Distribution for Occupants’ Stair Travelling Time............102

xiv

Figure 4-10 – Various Zones in Lift Simulation Model (two floors only)...............103

Figure 4-11 – Various Zones in Stair Simulation Model (three floors only) ...........104

Figure 4-12 – Animation of Lift and Stair Simulation Models at 300 seconds........105

Figure 4-13 – Lift Waiting Times during the Fire Occurrences at Three Levels.....110

Figure 4-14 – Lift Waiting Times.............................................................................112

Figure 4-15 – Lift Transportation and Stair Travelling Times.................................114

Figure 4-16 – Lift Pre-Evacuation Times.................................................................115

Figure 4-17 – Lift and Stair Evacuation Times ........................................................117

Figure 4-18 – Number of Evacuees in Queue in Lift System ..................................119

Figure 5-1 – Flow Diagram for Calculating Safety Index........................................130

Figure 5-2 – Load Variables for Safety Index for the locations of Lift Lobby, Lift

Shaft and Stair Shaft .........................................................................................132

Figure 5-3 – Typical Floor of a Hypothetical Building for a Generalised Fire

Scenario (dimensions not to scale) ...................................................................134

Figure 5-4 – Wind Speed Profile..............................................................................138

Figure 5-5 – Computational Domain for FDS Model ..............................................140

Figure 5-6 – Three Grid Sizes used in the FDS Model (38th floor view).................141

Figure 5-7 – Smoke Leakage Openings in the Lift Shaft Wall ................................142

Figure 5-8 – Output points in the Fire Compartment, the Lift Lobby, the Lift Shaft

and the Stair ......................................................................................................143

Figure 5-9 – Snapshots of Smoke View and Temperature Contour (Fire Scenario 1)

..........................................................................................................................144

Figure 5-10 – Snapshots of Temperature Contour and Vector slice (Fire Scenario 3)

..........................................................................................................................145

Figure 5-11 – Slice Snapshot of Visibility in the Lift Lobby (Fire Scenario 5).......146

Figure 5-12 – Snapshot of Temperature Contour at 720 seconds (Fire Scenario 5) 147

Figure 5-13 – Slice Snapshots in a Vertical Plane in the Lift Lobby at 600 seconds

(Fire Scenario 5) ...............................................................................................148

Figure 5-14 – Smoke, Gases and Heat in the Lift Lobby, the Lift Shaft and the LMR

(Fire Scenarios 1 to 6) ......................................................................................164

Figure 5-15 – Temperature and CO on the 38th Floor for Unprotected Lift Lobby .166

Figure 5-16 – Temperature in the Lift Shaft (with and without wind).....................167

Figure 5-17 – Temperature in the LMR (without wind) ..........................................168

Figure 5-18 – FED in Fire Scenarios 1 to 6..............................................................176

xv

Figure 6-1 – Probability of Excess Temperature Rise in LMR (Fire Scenarios 1 to 6)

..........................................................................................................................192

Figure 6-2 – Probability Distribution and Consequences of Excess Temperature in

LMR .................................................................................................................194

Figure 6-3 – Typical Electrical System for Essential and Non-Essential Supplies..195

Figure 6-4 – Fault Tree Analysis for Electric Fire in SOU ......................................198

Figure 6-5 – Fault Tree Analysis for Electric Power Failure in Unprotected Lift

Lobby................................................................................................................199

Figure 6-6 – Fault Tree Analysis for Electric Power Failure in Protected Lift Lobby

..........................................................................................................................200

Figure 6-7 – HRR during Three Stages of Water Application.................................203

Figure 6-8 – Complex Parallel-Series System for Probability of Water Spread......207

Figure 6-9 – Complex Parallel-Series System for Quantity of Water Spread..........208

Figure 6-10 – Quantity of Water Spread and Building Evacuation .........................209

Figure 7-1 – Multi-Objectives Decision Analysis Methodology .............................214

Figure 7-2 – Influence Diagram of Building Evacuation Risk Model .....................217

Figure 7-3 – Disadvantages of High Rise Living (Mori and UHK, 2002)...............220

Figure 7-4 – A Value Tree for the Parametric Global Weights ...............................222

Figure 7-5 – Value Functions for Building Evacuation Times ................................224

Figure 7-6 – Sensitivity Analysis for Different Weights Placed on Panic...............231

Figure 7-7 – Risk Values for Concept Design Options............................................232

Figure 8-1 – Evacuation Option 1: One-Fourth of the Building Population using

Protected Lifts and the rest using Stairs ...........................................................235

Figure 8-2 – Evacuation Option 2: Double Protected Lift Lobby............................236

Figure 8-3 – Rectangular Arrangement for a Common Lift Lobby .........................238

Figure 8-4 – Lifts located in the Refuge Area..........................................................239

Figure 8-5 – Smoke Seal in the Lift Landing Door and Wall Frame.......................241

xvi

LIST OF TABLES

Table 2-1: Summary of High-Rise Building Fires in the US for 1998 (Hall, 2001) ..15

Table 2-2: Extent of Fire and Smoke Damage in Apartment Buildings for 1994-98

(Hall, 2001).........................................................................................................16

Table 2-3: Locations of Fires in Apartment Buildings for 1994-98 (NFPA, 1999)...16

Table 2-4: Australian Fires, Fatalities and Nonfatal Injuries for 1993-94 (King, 1997)

............................................................................................................................17

Table 2-5: New Zealand Fire Fatalities for 1997-2002 (Miller, 2005) ......................17

Table 2-6: Causes of Death in New Zealand for 1997-2002 (Miller, 2005) ..............17

Table 2-7: People living in High-Rise Apartment Buildings, Australia (ABS, 2004)

............................................................................................................................18

Table 2-8: Effects of Visibility distance on Turning Back during Smoke in Corridors

(Pauls, 1995).......................................................................................................23

Table 2-9: Performance of Automatic Sprinkler System in Residential Buildings

(Marryatt, 1988)..................................................................................................37

Table 2-10: Risk Assessment Methods ......................................................................44

Table 3-1: The 9-Point Scale (Saaty, 1980) ...............................................................54

Table 3-2: Random Inconsistency Indices (Source: Saaty, 1980)..............................56

Table 3-3: Matrix (2 × 2) for Priorities of Lift Evacuation ........................................60

Table 3-4: Matrix (3 × 3) for Priorities of Psychological Impact...............................60

Table 3-5: Matrix (3 × 3) for Priorities of Physiological Impact ...............................60

Table 3-6: Global Risk Priorities................................................................................61

Table 3-7: Consistency Tests of Matrices ..................................................................62

Table 3-8: Lift and Stair Systems for Comparison.....................................................69

Table 4-1: Residents’ Age Distribution......................................................................78

Table 4-2: Residents’ Use of Lifts as Normal Access and Egress Routes .................79

Table 4-3: Residents’ Preferred Exit Routes during Fire Emergencies......................80

Table 4-4: Model Parameters .....................................................................................93

Table 4-5: Lift Time Periods and Number of Evacuees in Queue (2nd floor fire)....106

Table 4-6: Lift Time Periods and Number of Evacuees in Queue (19th floor fire) ..106

Table 4-7: Lift Time Periods and Number of Evacuees in Queue (38th floor fire) ..106

xvii

Table 4-8: Lift Time Periods and Number of Evacuees in Queue (2nd floor fire) –

25% population.................................................................................................107

Table 4-9: Lift Time Periods and Number of Evacuees in Queue (19th floor fire) –

25% population.................................................................................................107

Table 4-10: Lift Time Periods and Number of Evacuees in Queue (38th floor fire)–

25% population.................................................................................................108

Table 4-11: Stair Time Periods (2nd floor fire).........................................................108

Table 4-12: Stair Time Periods (19th floor fire) .......................................................109

Table 4-13: Stair Time Periods (38th floor fire) .......................................................109

Table 4-14: Stair Time Periods (2nd

floor fire) – 75% population ...........................109

Table 4-15: Stair Time Periods (19th floor fire) – 75% population ..........................109

Table 4-16: Stair Time Periods (38th floor fire) – 75% population ..........................109

Table 4-17: Means and Standard Deviations of Output Variables...........................120

Table 4-18: Verification of ARENA model for Lifts and Stairs..............................121

Table 5-1: Description of Fire Simulation Scenarios ...............................................139

Table 5-2: Time to Exceed Tenability Limits in Lift Lobby....................................177

Table 5-3: Time to Exceed Tenability Limits in Lift Shaft and LMR .....................178

Table 5-4: Time to Exceed Tenability Limits in Stair Shaft ....................................179

Table 5-5: Means and Standard Deviations of ASET (lift lobby)............................181

Table 5-6: Means and Standard Deviations of ASET (lift shaft) .............................181

Table 5-7: Means and Standard Deviations of ASET (stairs) ..................................182

Table 5-8: Means and Standard Deviation of RSET................................................182

Table 5-9: Safety Indices for Lift and Stair Evacuation...........................................183

Table 6-1: Temperatures and their Impact on Lift Systems.....................................190

Table 6-2: Probability of Excess Temperature Occurrence in LMR in Fire Scenarios

..........................................................................................................................193

Table 6-3: Unavailability of Lifts to Building Population .......................................194

Table 6-4: System Boundary Conditions .................................................................196

Table 6-5: Impact on Lift System.............................................................................201

Table 6-6: Probability of Water Spread at Three Levels..........................................206

Table 7-1: Concept Design Options and their illustrations ......................................216

Table 7-2: Risk related Parameters ..........................................................................218

Table 7-3: Parametric Values and Weights relating to Building Evacuation...........221

xviii

Table 7-4: Matrix (3 × 3) for Priorities Risk Factor (p3) .........................................225

Table 7-5: Parameter Strengths for Concept Design Options ..................................228

Table 7-6: Summary of Assigned Values.................................................................228

Table 7-7: Total Risk Values for Concept Design Options......................................229

Table 7-8: Total Risk Values from Analyses based on a Zero Weight for Panic.....230

Table 7-9: Total Risk Values from Analyses based on 100% Weight for Panic......230

xix

Key Words: Apartment building, Aged, Disabled, Emergency exit, Evacuation, Fire,

Fire safety, High-rise building, Hazard, Injury, Life safety, Lift, Modelling, Panic,

Risk analysis, Reliability, Smoke, Stochastic model, Tenability limits, Toxicity,

Uncertainty.

ABBREVIATIONS/ DEFINITIONS

Aged: Occupants more than 65 years in age.

Alternative evacuation facility: Alternative evacuation facility such as lift is to

comply with the performance requirements of BCA, other than the DTS provisions

such as stairs.

Apartment building: A building containing two or more sole occupancy units, each

being a separate dwelling (BCA classification).

AHP: Analytical Hierarchy Process.

ASET: Available Safe Evacuation Time.

BCA: The Building Code of Australia 2005. {A reference of BCA 2005 is quoted in

this thesis as this research was undertaken in the year 2004. However, BCA 2007 is

currently enforced and there is no substantial change in evacuation facilities.}

Coping time: Time for occupants’ coping activities those resulting from an occupant

perceiving that fire poses an actual threat to the point where the occupant initiates

evacuation or avoidance activities.

Disabled: Occupants physically incapable of evacuating building using stairs.

DTS: Deemed-to-satisfy Provision of BCA 2005.

EEP: Emergency evacuation procedure.

Evacuation and egress: The term “evacuation” is used for emergency escape from

the building whereas “egress” is used for non-emergency exit only.

Fractional effective dose (FED): FED refers to incapacitation, lethality or endpoint.

When not used with reference to a specific asphyxiant, the term FED represents the

summation of FEDs for all asphyxiant in a combustion atmosphere.

Fire door (FD): Door designed to contain fire for a nominated period while

facilitating emergency egress. Fire doors are rated in terms of structural adequacy,

integrity and insulation.

High-rise buildings: High-rise buildings are 25 m in effective height or above. The

effective height is measured from the lowest level of fire brigade access to the

xx

highest occupied level (BCA classification). The term “mega high-rise” is used for

buildings having 50 storeys or more.

Incapacitation: State of physical inability to accomplish a specific task.

Lift and elevator: Lift is the UK nomenclature; Elevator is the US nomenclature.

LMR: Lift Machine Room.

MODA: Multi-Objectives Decision Analysis.

PDF: Probability Distribution Function.

POE: Probability of Occurrence.

Protected shaft and unprotected shaft: Protected shaft is provided with protected

lift lobby having fire resistive construction and fire door whereas unprotected shaft is

unprotected lift lobby without fire resistive construction and fire doors.

Refuge area: Refuge area is provided in high-rise building as a temporary shelter for

occupants during fire emergencies.

Response time: Time for occupants’ response from occurrence of detectable cue to

activities involved in responding to those cues.

Round trip time (RTT): RTT is the time taken by lift after stopping at entrance

lobby for performing the entire trip after serving the most probable number of floors.

RSET: Required Safe Evacuation Time.

Smoke alarm: A smoke alarm complying with Australian Standard AS 3786. This

alarm is self-contained including an audible warning device and is not connected to

the building fire detection and alarm system.

Smoke detector: A smoke detector complying with AS 1670 and connected to the

building fire detection and alarm system.

Smoke door (SD): Door with a closely fitting leaf or leaves, furnished with smoke

seals and an approved sensing device, release mechanism and closing mechanism, to

protect openings in partitions against the passage of smoke during a fire.

SOU: Sole Occupancy Unit – a unit in an apartment building.

Uncertainty: Two type of uncertainties – uncertainty due to lack of fundamental

knowledge about specific factors or pathways; uncertainty due to variability

(stochastic uncertainty, randomness) in data.

Tenability limits: Tenability limits are the safe environment conditions during

which occupants can perform designated actions. When the limit exceeds, occupants

are considered incapacitated due to hazardous conditions or exposure.

Vertical transportation: Vertical transportation means lift transportation.

xxi

NOMENCLATURE

AE Total leakage area from the lift shaft m2

aij Element (or attribute) of matrix -

cp Specific heat of gas kJ/kg.K

Cs Extinction coefficient 1/m

D Physical fire diameter m

D* Characteristic fire diameter m

Fs Nominal evacuee flow in stairs person/m/s

g Acceleration due to gravity m/s2

h Height above the ground for wind m

J Number of lifts -

k Exponent for the wind velocity -

L Stochastic parameter for load variable -

m Number of lift round trips -

n Number of floors -

N Number of people per floor person

p Parameter measure -

PE Pressurisation level in the pressurized space Pa

q Radiant heat flux kW/m2

Q Heat release rate kW

Q* Heat release rate kW

QE Air supply to the pressurized space m3/s

Q Water discharge of fire protective system litre

RA Acceptable Risk -

S Stochastic parameter for strength variable -

t1 Egress time (congestion) second

ta Elevator evacuation start up time second

tn Egress time (free walk) second

to Travel time from the elevator lobby to the outside second

tr,j Time for round trip second

ts Walking time between adjacent floors second

xxii

tLPM Lift pre-movement time second

tLW Lift waiting time second

tLT Lift transportation time second

tLI Lift intermittent floor movement time second

tFD Fire detection time second

tLM Occupant movement time for lift second

tLPE Lift pre-evacuation time second

tSPM Stair pre-movement time second

tST Stair travelling time second

tSI Stair intermittent floor movement time second

T Air temperature C

T∞ Ambient temperature K

TLMR Temperature in LMR C

vi Parametric value -

V Visibility distance m

Vh Wind speed at a height m/s

Vr Wind speed at the reference height m/s

W Effective width of the stairs m

wi Parameter weight -

x Number of evacuees person

Greek symbol

Mµ Mean of safety margin -

Mσ Standard deviation of safety margin -

β Safety index -

volβ Degree of voluntariness -

α Fire growth coefficient kW/s2

η Lift trip inefficiency %

λ Occupant arrival rate person

λ Eigenvalue -

ρ∞ Density at ambient temperature kg/m3

1

1. INTRODUCTION

1.1 Background

In the past stairs was the only emergency evacuation facility in high-rise buildings.

Alternative safe evacuation facility has been of interest to many researchers. Over the

past two decades, the works of Klote (1982) and many others (Pauls, 1977, Pauls et

al., 1991) have led to a better understanding of many problems associated with the

use of lifts during fire emergencies. Subsequent to the WTC twin tower collapse on

September 11, 2001, researchers have focused their efforts on the use of lifts as an

alternative evacuation facility in buildings (Kuligowski, 2003, Kuligowski and

Bukowski, 2004). Analysis of 480 first-person accounts in WTC showed that 81%

used the stairs, 6% used the lifts, and 13% used a mix of lifts and stairs (Bill, 2002).

Two-third of the survivors from floors above the 78th floor sky lobby (an area where

people changed lifts) and one-third from floors between the sky lobbies on the 44th

and 78th floors used lifts. A total of 2605 lives was lost. Some researchers believe

that if adequate emergency escape lifts were available at the WTC, more occupants

could have been evacuated in a shorter period of time and the extent of the fatality

might have been greatly reduced (So and Yu, 2003). There were incidents where

stairs became unsafe due to the presence of smoke, and occupants therefore used lifts

as a faster mode of evacuation (Bill, 2002). This caused a great concern among the

fire safety community and generated the need to review the relevance of existing

evacuation strategy for high-rise buildings.

Use of lifts is not recommended for building occupants for evacuation purposes due

to some of the major incidents. However, past records show that the occupants do

use lifts for emergency evacuation (see Appendix A). In some cases, evacuating

occupants were trapped in the lift car and exposed to hot smoke resulting in

unconsciousness and deaths (Barnett et al., 1992). One of the worst cases was in the

MGM Grand Hotel, Las Vegas on November 21, 1980, where fire from ground floor

restaurant rapidly spread to the upper floors through air systems, stairways and lift

shafts. The fire caused 85 fatalities due to smoke inhalation and many were found in

lifts (Bryan, 1982). The occupants’ decision to use lifts can be perceived as an

alternative during life threatening conditions as the occupants may feel that their

2

main exit routes are closing rapidly due to smoke and toxic gases and their search for

main exit route would be a waste.

Whilst alternative evacuation facilities are explored for safe evacuation from

buildings, evacuation using lifts during fire emergencies is a controversial issue

because the safe operation of normal lift is not always ensured. The major concerns

are believed to be loss of power supply, smoke infiltration, lift passing through the

fire-affected floor, lift car being exposed to heat radiation, water run off from fire

fighting operation, malfunctioning of lift equipment, electric short circuit of re-call

button, occupant panic and overloading (Klote, 1982). The use of lifts for an

emergency exit is viewed as an alternative in high-rise buildings since the occupants

can have options for an alternative evacuation route. Building management will also

use the lifts for prompt access to the fire-affected floor.

The benefits of using combined lift and stair evacuation have been demonstrated by

many researchers. In a study conducted by Klote et al., (1993a), it was found that the

use of both lifts and stairs can reduce evacuation times by as much as 50% over the

use of stairs alone. If lifts are used in mega high-rise buildings during an emergency

situation, evacuation times can be reduced to 15-30 minutes instead of 2-3 hours

(Siikonen and Hakonen, 2003). Greater evacuation efficiency occurs as the height of

the building increases. A similar result was found by Andersson and Wadensten

(2000) in their simulation of the One Canada Square building at Canary Wharf in

London. Evacuation time was reduced with the use of stairs and lifts. It was also

noted that the response time and coping time during an emergency was up to two

thirds of the total time to evacuate a building. With the use of stairs and lifts, the

building can be evacuated promptly with significantly reduced evacuation time.

However, this requires extensive research work in view of the uncertainty in fires,

life threatening conditions arising from fire effluents and effects on human

physiological and psychological behavioural response.

This research project began by identifying issues relating to the lift safety and was

aimed at determining the feasibility of using lifts (with suitable design modifications)

as a safe and efficient evacuation facility in apartment buildings during fires and

other emergencies. The research can be extended to other types of buildings.

3

1.2 Problem Statement

Australian building fire statistics show that there were 12.9% apartment fires

resulting in 12.5% fatal injuries and 18% nonfatal injuries of total residential fires

during the year 1993-94 (King, 1997). A total of 1547 fires were reported in

apartment buildings causing 8 fatalities and 133 nonfatal injuries. A recent study

showed that 32.1% of the residential fire deaths consisted of aged less than 15 years

and 25.9% consisted of aged 60 years or above during the period of 1997 to 2003

(Miller, 2005). Thirteen percent of the population living in high-rise apartment

buildings in Australia are aged 65 years or above (ABS, 2004). This pattern is also

reflected in other international studies. A study in Japan found that aged (65 years

and above) accounted for 47.8% of residential fire deaths (Sekizawa, 1991). The risk

for disabled is five times higher than the average population. In New Zealand, 26.5%

of the residential fire deaths consisted of aged persons during the period of 1991 to

1996 (Duncanson et al., 2000). Twelve percent of the population are aged in New

Zealand (Dunstan and Thomson, 2006). These data show that children (less than 15

years) and aged persons (more than 60 years) are more prone to risks in residential

fires than those from other age groups.

Figure 1-1 – Fire and Rescue from Pallister Plaissance Apartments, USA

(Fire caused 3 causalities in a 12 storey building fire (188 units); occupants were

rescued from windows due to heavy smoke in evacuation routes; April 2000, Source:

The Detroit News)

The aged and disabled persons are slower to evacuate through narrow stairs with an

average walking speed of 0.43 m/s (Proulx, 1995). They might adversely affect the

evacuation of other occupants. Due to their slow movement, the problem of using

4

stairs, such as bottleneck, queuing and stampede during emergency evacuation can

be encountered. Uncertainty and anxiety can be expected amongst the aged and

disabled occupants during building evacuation (Proulx, 1995). In one of the buildings

studied by Proulx (1995), many elderly and mobility impaired residents were

unsettled by the fire alarm, not knowing what to do. A woman in a wheelchair

entered the main staircase blocking the way for descending occupants, threatening

that she would go down the stairs with her wheelchair. Residents took her back to

the corridor and stayed with her until the arrival of fire fighters.

Subsequent to the WTC bomb explosion on February 26, 1993, Juillet (1993)

conducted an interview of 27 disabled and impaired persons in one of the towers.

The evacuation time with the assistance of other occupants or emergency personnel

was reported to range from 40 minutes to over 9 hours with an average of 3.34 hours.

Some of the permanently disabled residents were not able to evacuate the building

without aid. Pauls (1977) determined that 3% of occupants in high-rise buildings in

Canada were unable to use stairs due to their permanent or temporary immobility.

Evacuation time increases with the number of stair flights. Long evacuation times

might endanger the life of evacuees due to tiredness, dizziness, slipping on surfaces

or becoming less capable physically. Evacuees normally experience fatigue after

about five minutes of travelling downstairs (So and Yu, 2003). Research has also

shown that the evacuees will begin to suffer fatigue when they have travelled about

18 storeys (So and Yu, 2003).

Building occupants often add complexity and danger to their evacuation process by

remaining in the building until the danger becomes severe. Also, prior experiences

with false alarms result in occupants attempting to validate the alarm before starting

the evacuation (Allen, 1995). Precious time is lost during actual fire emergencies.

Sometimes hazardous conditions arise quickly and in such circumstances occupants

may make irrational decisions when searching for the escape route (Barnett et al.,

1992).

Fire fighters also need a safe, fast and reliable mode of transport system because they

are not expected to climb stairs to much higher levels with heavy gear. Fire fighters’

5

movement can be delayed due to evacuees’ using stairs (Kuligowski, 2003).

Presently fire fighters’ lifts are permitted to be used up to one level below the fire-

affected floor depending upon the safety of lift evacuation system.

There is an increasing trend of high-rise apartment building construction in Australia

(ABS, 2005). The 323 m high Q1 tower with 78 storeys was built in Australia

during 2005 and is the tallest residential tower in the world (Emporis, 2005).

Another such tower in Australia is the 297 m high Eureka tower with 88 storeys

(Emporis, 2005). Although stairs are the main evacuation facility in high-rise

buildings, the use of lifts is to be considered as an alternative evacuation facility for

the aged, motion impaired and physically weak people in this study.

1.3 Research Statement (Hypothesis)

Occupants in high-rise apartment buildings need safe and efficient evacuation

facilities during fire emergencies. Currently, stairs are the only evacuation facility

permitted by building regulations for high-rise buildings because lifts are not

considered safe in fire emergencies. The risks associated with the use of lifts are

manageable. The use of lifts for emergency evacuation is feasible. If lifts are found

to be safe for emergency evacuation, occupants including the aged and disabled will

have an alternative evacuation system and can safely evacuate the buildings. This

will reduce the injuries considerably.

1.4 Objectives of Research

This research explores the feasibility of using lifts as a safe alternative evacuation

facility in apartment buildings. The main objective of this study is:

To study if lifts (elevators) provide an acceptable means for evacuation in apartment

buildings greater than 25 m in effective height during fire and other emergencies.

The specific objectives of this research are:

6

1. To identify the risks associated with the use of lifts for emergency evacuation and

develop an inter-relationship among the risks.

2. To establish a research strategy for an acceptable level of risk and consider

suitable design options and evacuation strategies.

3. To investigate a suitable risk assessment approach for lift evacuation system.

4. To gain a better understanding of residents’ preferred access and egress routes

and expert’s opinion on the use of lifts.

5. To develop a model for the building evacuation under uncertainties associated

with human social, behavioural and physical movement (with a priori heuristics

of the lift domain) and determine the probable time for safe evacuation.

6. To develop a model for the fire and toxic hazards under variable conditions and

determine the probable time when evacuees are predicted to become

incapacitated during exposure of fire effluents in evacuation routes.

7. To determine the reliability of lift operational mechanism and the feasibility of

reliability improvement for design options.

8. To assess the risks in building evacuation systems for a comparative analysis.

9. To study the feasibility of using lifts as an alternative evacuation facility and

propose redundancy measures for the safe and efficient lift evacuation system.

1.5 Scope

The scope of this research is to determine the feasibility of lifts as an alternative

evacuation facility in apartment buildings greater than 25 m (high-rise) in effective

height. This system could also be applied to non-fire emergency evacuation (like

bomb threat, gas leakage or similar calamities). The existing and modified lift

designs are to be considered in determining the safe lift evacuation system.

The scope of this research does not include buildings less than 25 m (low-rise) since

the benefits of lift evacuation will be far outweighed due to cost considerations and

occupants will feel comfortable going down the stairs rather than waiting for a lift

(Smith, 2003).

7

1.6 Research Methodology

The risk in lift evacuation system involves complexity, represented by multiple

attributes and requiring diverse sources of evidence to demonstrate its achievement.

The research methodology used in this project involves the following components:

• To identify all the significant risks in the lift evacuation system and develop a

relationship among the risks.

• Rank all the risks (risk priorities) in terms of likelihood of occurrence and

expected impact on the building evacuees.

• Establish a research strategy for an acceptable level of risk.

• Identify risk control design options and evacuation strategies for evaluating

risks.

• Quantify consequences with the models and techniques.

• Conduct risk assessment with a suitable method/ technique.

• Select appropriate risk control design options.

A relationship is developed among the risks related to the use of lifts for emergency

evacuation. The relationship identifies the key issues to be addressed. The issues of

human behavioural response, fire hazards and lift operational mechanism give rise to

three risks i.e. decision uncertainty, panic and injuries (nonfatal and fatal). The risks

of “decision uncertainty, panic and injuries” may be interrelated, which may

ultimately lead to psychological or physiological impact. These related risks form a

complex process for risk assessment. The research strategy involves risk

management by reducing the level of consequences to an acceptable level. The risk

consequences are reduced by considering design options and evacuation strategies.

The risks are quantified from building evacuation simulation models, fire hazard

models and probabilistic risk models.

Stochastic model of building evacuation, fire hazard model of life threatening

condition and probabilistic analysis of reliability are developed for quantifying the

parametric values of risks. The technical parameters are multi-dimensional. The

direct measurement of risk using the traditional approach (i.e. Risk = Probability of

8

Occurrence × Severity of Consequences) involves complexity. An indirect evaluation

procedure based on Multi-Objectives Decision Analysis (MODA) approach is used.

This risk decision analysis approach constitutes the process of analysing and scoring

the parameters. The parameters are given weights based on data generated from the

statistics and simple analytical techniques are used. The priorities of conflicting key

issues (risks) are assigned with the help of Analytical Hierarchy Process (AHP). The

results of the analysis are theoretically sound and justified for making suitable

decisions. This research methodology is elaborative and requires identification of

various key issues from the literature review. Therefore, full details of the research

methodology are presented in a separate chapter (see Chapter 3).

1.7 Thesis Contents

This thesis consists of nine chapters followed by references and appendices.

Followed by the Introduction chapter, the contents of the thesis are:

Chapter 2 reviews the literature on the current trends in fire emergency evacuation

systems for high-rise buildings and human behaviour in case of fire, tenability limits

of fire and smoke, risk assessment techniques and related models.

Chapter 3 presents the details of the research methodology used in this project.

Details of the research methodology based on Multi-Objectives Decision Analysis

(with Analytical Hierarchy Process) on a comparative basis are given. Inter-

relationships of identified risks are presented. Design options and evacuation

strategies are identified. Research strategy based on reducing risk consequences to an

acceptable level is described. (Objectives 1, 2 and 3)

Chapter 4 presents the stochastic models for building evacuation systems (lifts and

stairs). The stochastic models are developed using SIMAN ARENA software. Lift

time periods, stair time periods and existence of queuing occupants are derived as

output variables for risk assessment. The results of a survey and interview are also

presented. (Objectives 4 and 5)

Chapter 5 analyses the fire and smoke hazards under variable conditions. Tenability

limits relating to temperature, smoke, toxic gases and visibility are considered in

9

modelling. The concepts of ‘fractional effective dose’ and ‘safety index’ are

presented for evaluating evacuees’ safety. (Objective 6)

Chapter 6 analyses the reliability of lift operational mechanisms. Risks associated

with water spread, malfunctioning of lifts and electric power failure are modelled and

analysed using probabilistic techniques, i.e. complex event tree and fault tree

analyses. (Objective 7)

Chapter 7 provides an integrated model for risk assessment based on the Multi-

Objectives Decision Analysis method. The model is proposed to comprehensively

evaluate all the risks associated with the lift system for the design options and

evacuation strategies. (Objective 8)

Chapter 8 demonstrates the feasibility of designs for lift evacuation systems based on

conceptual design options and evacuation strategic planning. Redundancy measures

are proposed. (Objective 9)

Chapter 9 presents the summary, research findings and scope for future work.

10

2. LITERATURE REVIEW

Over the last decade or so there has been a proliferation of interest on the topic of lift

evacuation system. Lifts used in the early of 20th century lacked many fire safety

features. For example, many lifts were located in open shafts, which acted as a

chimney during fires (Martin, 2003). Some lifts were not provided with fire resistive

doors and car cabins were exposed to risk areas. Lift doors were not fire rated and

they could transmit heat and smoke to the lift shafts and to other floors. These lifts

were considered unsafe for fire evacuation from the viewpoint of operational safety.

Major concerns of such lifts were loss of power supply, smoke infiltration, lifts

passing through danger zones and lift car being exposed to heat radiation, lift

software failure, electric short circuit of lift re-call button, damage by water run off

from fire fighting operation, occupant panic and lift overloading (Klote, 1982).

Whilst most lifts adhere with the national and international safety norms, minor

modifications in lift design can place lifts as the next available option of safe

evacuation facility in emergencies. Research and development work are already

under way in the US. Evidences from literature and preliminary research work

demonstrated that design of a lift evacuation system for a small number of people is

feasible (Klote et al., 1995). A lift evacuation system would be most beneficial for

the disabled in office buildings and for all residents in luxury apartment buildings

due to low occupancy load (Klote et al., 1995). However, the use of lifts for large

number of people was not thought practical at the time, due to the required

complexity of the system (Klote et al., 1995). Pauls et al. (1991) suggested that

evacuation via lifts should only be an option for occupants who can not use the stairs.

A literature survey was conducted to investigate the risks associated with the use of

lifts and need for the lifts as an alternative emergency evacuation facility in

buildings. The literatures were obtained from various search engines and the

National Institute of Standards and Technology (NIST) website. The use of lifts for

emergency evacuation is a multifaceted topic and therefore the results of the

literature review are organised into different categories as given next.

11

2.1 Current Regulations in Relation to the Use of Lifts for Evacuation

Lifts can be broadly classified as unprotected (normal) lift and protected emergency

(fire fighter) lift. Internationally recognized ASME A17.1 (2000) does not

recommend the use of normal lifts in fire emergencies and it also does not elaborate

precautionary measures against fire. However, the US NFPA 101 (2000) Life Safety

Code includes the provision of a normal lift as a secondary means of evacuation

facility for air traffic control towers only. The provision is permitted due to the small

footprint of the building, where the construction of two remote stairs is not possible

and moreover the building is not accessible to the general public. An example of a

structure that uses lifts as the secondary means of evacuation is the Stratosphere

Tower in Las Vegas, USA.

In a survey conducted by the International Organization of Standardization (ISO/TR

16765, 2003), it was determined that there is a specific requirement for fire fighter

lifts, in at least 12 countries including the US, UK, Japan and Australia. The fire

fighter lifts are recommended for use by fire fighters and not by the general public

during fire emergencies. This is particularly important in high-rise buildings, where

the carriage of heavy fire fighting gear takes time and uses valuable resources

(Degenkolb, 1991). An example of this was the 62 storey First Interstate Bank fire in

1988 in Los Angles, where lifts were not used, and a fire on the 12th floor required

100 fire service men to carry equipment up the stairs (Degenkolb, 1991) and took

more than 3 ½ hours to control the fire. The scenario might have been worse if the

fire had occurred on the upper levels.

The Building Code of Australia (ABCB, 2005) mandates the installation of

emergency lifts in all the buildings which have an effective height of more than 25m

and buildings in which patient care areas are located at a level that does not have

direct egress to a road or open space. The emergency lifts contain the provisions for

fire brigade control and facilities for people with disabilities.

The US Building Code mandates the installation of fire fighter lifts for fire fighting

and rescue purposes. ASME A17.1 (Safety Code for Elevators and Escalators)

recommends lift operation in two phases during fires (ASME A17.1, 2000). In

12

Phase-1, smoke detector senses the fire and sends a signal to the lift control panel.

All the lifts move to the ground floor and halt, resulting in no lift operation by the

public. In Phase-2, the fire brigade arrives at the building and manually overrides the

lift operation for rescue and fire fighting operations. Fire brigade personnel can use

the fire fighter lifts for evacuating the aged and disabled persons and use them for

access and egress for fire fighting purposes.

The UK Building Code (NBS, 2000) requires fire fighting shaft for buildings, which

have occupied space at more than 18 m above and/or 10 m below the fire brigade access

level. Firefighting shafts incorporate a firefighting lift that opens into the lobby (see

Figure 2-1). The lift has a back-up electrical supply and car control overrides. The

primary function of the lift is to transport firefighting personnel and their equipment

to the scene of a fire with the minimum amount of time and effort. They enable

firefighting operations to start quickly and in comparative safety by providing a safe

route from the point of entry to the floor where the fire has occurred. It may also be

used to help evacuate less mobile people in the event of fire, provided the evacuation

is supervised and managed.

Figure 2-1 – Fire Shaft and Stairs Protected with Lift Lobby (NBS, 2000)

The CEN (2003) allows the use of lifts one level one level below the fire affected floor

(see Figure 2-2). In such circumstances, fire fighters must be aware of the location of

fire affected area.

13

Figure 2-2 – Fire Fighter Lift is utilized up to one level below the Fire-Affected

Floor (CEN, 2003)

2.2 Recent Developments in the Subject Area

During fires, lifts are taken out of service and people are advised not to use lifts.

Deliberations over the years resulted in varying philosophies toward the use of lifts

for emergency evacuation. The American Society of Mechanical Engineers (ASME

International) has organised several symposiums, where lift industries responded to

the inquiries of using lifts for emergency evacuations. Based primarily on influences

from the fire fighting communities, the emphasis has always been placed on the use

of lifts for fire fighting only, and not for general public use (Koshak, 2003).

A workshop on the “Use of Elevators in Fires and Other Emergencies” was held in

March 2004 in Atlanta, Georgia (NIST Special Publication 983, 2003). This

workshop was co-sponsored by ASME International, NIST, International Code

Council (ICC), National Fire Protection Association (NFPA), US Access Board, and

the International Association of Fire Fighters (IAFF). The goal of the workshop was

to develop concrete proposals after considering ‘pros and cons’ for the use of lifts

during fire emergencies. The ‘pros’ include the advantages of lift systems, whereas

‘cons’ are the risks associated with their use during fire emergencies.

14

The consensus of workshop attendees was that the lift operation should work only

until Phase-1 goes into effect. However, the lift operation is not feasible to control

prior to Phase-1. The workshop attendees were of the opinion that the building codes

should have scope for this operation. They also made recommendations to the

ASME A17 Emergency Operation Subcommittee for further research addressing the

technical issues and develop performance requirements for lift evacuation system

during fire emergencies.

NIST funded a research project on “Analysis of the life safety consequences of

smoke migration through elevator shafts” conducted by Klote (2003). The study was

confined to the smoke movement in office buildings with the current infrastructure of

lift system and was conducted with the help of fire and smoke simulation modelling

with CFAST (Peakcock et al., 2004) and CONTAMW (Walton, 1993). The results

showed that unsafe conditions arise on upper levels through lift shafts during a fully

developed fire. However, sprinkler controlled fire reduced the unsafe conditions in

the buildings. The occupants’ evacuation aspects were not considered in this study.

The author also stated that future research is needed to evaluate the extent to which

compartmentation failure would impact smoke flow through lift shafts (Klote, 2003).

He also stated that the approach of the computer models (CFAST and CONTAMW)

was cumbersome and yielded questionable results for scenarios involving reverse

stack effect.

The City University of Hong Kong has also taken initiatives to explore the possibility

of using lifts in super high-rise buildings. In an article published in the Fire

Prevention Journal, So and Yu (2003) stated that all the problems could be solved

without too much difficulty by increasing the capital cost of the system but the

complex psychological reaction of the evacuees could be a major obstacle. They

argued that lifts could be used for emergency escape provided evacuation must be as

quick as possible, residents must feel safe to wait for lift services and at every stop a

lift car must have space to accommodate waiting passengers. However, it would be

difficult in practice to satisfy all the stated premises simultaneously.

A performance-based fire engineering approach was used in the design of the 88-

storey Eureka Tower, Melbourne (Aloi and Rogers, 2002). A lift evacuation strategy

15

was used in the building design for emergency exit. The lift arrangement was

stacked into vertical evacuation zones. This led to an additional evacuation facility

for the occupants, who can evacuate via stairs within the fire-affected zone until they

reach the next lift transfer level (on levels 24 and 52). At transfer floors, occupants

may use express lifts to the ground floor. Occupants with disabilities are assisted by

fire fighters in their dedicated lifts within the zone of fire-affected floor. (In tall

buildings, lift zoning is made for a group of floors for providing efficient lift service

to the occupants). This gave a scope for further research for extending the lifts for

other occupants within zones of fire-affected floors.

2.3 Statistics

A series of significant fires over the years has demonstrated the danger to the

occupants in high-rise apartment buildings. In the United States, over 80% of all fire

deaths occur in residential occupancy and about 20% of these fatalities occur in

apartment buildings. Aged people are mainly at threat in high-rise apartment

buildings. Sekizawa (1991) studied the death pattern of residential fires and found

that people 65 years and older account for 47.8% fire deaths. The risk of death in a

residential fire for disabled residents is 5 times higher than the average population.

The risk for a person greater than 65 years of age and bedridden is 41 times higher

than the average population. According to Hall (2001), apartment occupancies have

experienced the highest frequency of fires, deaths and injuries in high-rise buildings.

The statistics for fires by the type of high-rise occupancy in the US during the year

1998 are shown in Table 2-1.

Table 2-1: Summary of High-Rise Building Fires in the US for 1998 (Hall, 2001)

16

Apartment buildings pose the biggest fire problem in the US and therefore the

occupants of these buildings are at the highest risk. Due to the presence of fire

fighting systems, the fire itself can be limited to the room of origin, however, smoke

spreads far beyond and presents risk to the occupants. Researchers found that 14%

of fires in apartment buildings resulted in smoke propagation beyond the origin of

the fire-affected floor during the year 1994-98 (see Table 2-2).

Table 2-2: Extent of Fire and Smoke Damage in Apartment Buildings for 1994-98

(Hall, 2001)

Brennan’s (1999) analysis of US NFIRS fire statistics over a period of 10 years

estimated that the percentage of fatalities outside the room of origin in evacuation

routes was approximately 15%. Occupants were reported to have died due to smoke

in stairs and lifts. The typical area of origin of fires in high-rise apartment buildings

is shown in Table 2-3. Most fires occurred in the kitchen, followed by bedrooms.

Table 2-3: Locations of Fires in Apartment Buildings for 1994-98 (NFPA, 1999)

Limited data available on Australian building fires shows that fire in residential

properties accounted for 62.4% of all the structural fires during the year 1993-94.

Among residential fires, 12.9% were apartment/ flat fires resulting in 12.5% of total

fatalities and 18% of total nonfatal injuries as shown in Table 2-4.

17

Table 2-4: Australian Fires, Fatalities and Nonfatal Injuries for 1993-94 (King, 1997)

The details of fire related fatalities in New Zealand for the period 1997-2002 are

shown in Table 2-5. Average annual residential fire fatalities are 21.8.

Table 2-5: New Zealand Fire Fatalities for 1997-2002 (Miller, 2005)

The causes of deaths for 115 victims in New Zealand for the period 1997-2002 are

shown in Table 2-6 (Miller, 2005). The cause of death relates directly to three

general areas of fatal effects – consequences of exposure to fire (burns, thermal

injuries to airways and incineration), inhalation of toxic products of combustion

(smoke, CO2, CO, and other poisonous gases, hypoxia and asphyxia) and shock from

injuries that precipitate death from pre-existing health conditions (cardiac failure and

respiratory diseases).

Table 2-6: Causes of Death in New Zealand for 1997-2002 (Miller, 2005)

18

It can be noted that the death can be attributed to more than one cause – 79 (60.8%)

died from a single cause, 47 (36.2%) from two causes and 4 (3.1%) from three

causes. Miller (2005) also indicated that 28.5% were found dead while attempting to

evacuate or escape and later died of their injuries. Aged and disabled persons have a

‘high risk’ level than others. The causes of deaths due to pre-existing health

conditions will be more in the ‘high risk’ group. The data indicated that death from

pre-existing health condition was 5.95% (see Figure 2-3).

Figure 2-3 – Factors Causing Death during Residential Fires (Miller, 2005)

In Australia, the number of people living in high-rise apartments rose from

approximately 129,000 in 1981 to around 334,000 in 2001, representing an increase

of roughly double the number of people living in private dwelling units (ABS, 2004).

The socio-demographic profile of people living in high-rise units shows that there is

a decrease in aged population (65 years or above) from 17% to 13% (see Table 2-7).

Table 2-7: People living in High-Rise Apartment Buildings, Australia (ABS, 2004)

The number of fires with the increase of people living in high-rise apartment

buildings illustrates the necessity of reviewing the evacuation strategies and

procedures for prompt exit.

19

2.4 Human Behaviour during Fire

Proulx (2003) mentioned after several studies of evacuation of tall buildings that

normal patterns of behaviour and movement route choices tend to persist during

emergency situations. She further stated that occupants often ignore the fire cues or

spend time investigating, seeking information about the nature and seriousness of the

situation, which delay the evacuation time. With the ambiguous information and

short time for decision making, people are likely to choose their most familiar way

out of the building. Visitors often use familiar entry and exit routes in emergency

situations and they may not be familiar with the protected evacuation routes.

Sime’s (1983) study provided strong evidence of human behaviour as affiliative

behaviour during building evacuation. The affiliative model exhibits occupants’

strong tendency toward familiar people and familiar places (such as usual entrance

route). Individuals respond quickly to ambiguous cues, whereas intact individuals in

groups did not begin to evacuate until there was a clear sign of the fire threat.

In a 20-storey apartment building (Japan), a major fire occurred on October 28, 1996,

which engulfed from 9th floor to 20

th floor within 30 minutes. Subsequently, a survey

was conducted in the building, which showed that 47% residents used lifts, 42%

residents used stairs and 7% residents used both. Not even a single resident used

stair on 18th to 20th floors (Sekizawa et al., 1996). In the Forest Lane fire, 40% of the

respondents said that they used lifts (Proulx, 1995). This included people who were

assisted by rescue personnel and occupants who were not successful in escape

through stairs. Occupants may prefer to use lifts, instead of stairs, for evacuation due

to the physical exertion of walking down several flights of stairs (Klote et al., 1993a).

The principal variables influencing an occupant’s decision to move through smoke

tends to be recollecting the exit location and ability to estimate the travel distance

required; secondary variables are the perception of the severity of the smoke and heat

(Bryan, 1983). Occupants try to evacuate the building through evacuation facilities

if the smoke and hot gases are within the tenability limits. Due to poor visibility,

occupants’ walking speed also reduces to a greater extent. Jin and Yamada (1989)

also reported that occupant mental capability reduces with the increase of smoke

20

density and increased radiant heat exposure. In such circumstances occupants

sometimes take wrong decisions about their safe evacuation route.

Proulx (1993) developed a stress model to demonstrate the stress levels induced in

people while making a decision in fire conditions (see Figure 2-4). The stress model

starts with the perception of ambiguous information, which is interpreted in the

processing system resulting in denial of information and non-reactive response.

Occupants tend to ignore the information with the frequency of false fire alarms. A

state of uncertainty prevails with the repeated ambiguous information that induces a

feeling of stress. Overloading of information leads to fear during the emergency

situation, inflicting an increased level of stress and concern of safety. Irrelevant

information induces high levels of stress causing worry about self concern for own

performance in overcoming the emergency situation. Irrelevant information further

manifests a state of confusion since an individual puts in more effort to resolve the

problem and that may result in fatigue and inefficiency.

Figure 2-4 – Stress Model (Proulx, 1993)

21

Schultz (1968) has defined panic type of behaviour a fear-induced behaviour which

is non-rational, non-adaptive, and non-social, which serves to reduce the escape

possibilities of the group as a whole. The perception of fear can further induce in to

a panic behaviour. An example of the way in which the concept of panic is attributed

to people, on the basis of the outcome of a large scale fire tragedy occurred on 28

May, 1977, is provided by a comparison of the extensive report on the Beverly Hills

Supper Club fire, USA. Report concluded that ‘Panic is not considered a major

contributing factor to the large loss of life, but such behaviour probably did occur

when people knew they could not escape’ (Canter, 1990). While the evidence for

panic occurring after people knew they could not escape is inconclusive, the fact is

clear that there was no panic while there was reasonable access to the exits (see

Figure 2-5).

Figure 2-5 – Fear induced Panic Behaviour

There has been an argument by another school of researchers that panic is an event of

rarity during emergency evacuations (Bryan, 2002). Researches have shown that

human behavioural response is consistent during fires and people take rational

decisions. Ramachandran (1991) also found people generally act rationally and

appropriately and they do not panic, which is due to the fact that information is

available to people regarding the existence, size and location of the fire. If the

evacuation routes are not available during the immediate threat of fire, the issue of

irrational behavioural may arise. However, literature survey (Canter, 1990) remains

inconclusive about the perception of issues of human behavioural response during

the non-availability of evacuation route (such as stairs or lifts). Human behavioural

response may vary during lift waiting time and it is likely that a stage of ‘decision

Availability of information

Non-availability of information

such as evacuation

routes

Fear induced in limited time

(or danger)

Non-panic

behaviour

Panic behaviour

...inconclusive?

22

uncertainty’ may prevail amongst them. When building occupants are subjected to

perceived life threatening conditions and there is a hope of survival, there may be an

urge of doing something, which may lead to panic. Therefore, it is possible that the

phenomena of ‘decision uncertainty’ and ‘panic’ could occur, no matter how rare

they are.

Whilst occupants waiting for a lift, concern for their life safety may also arrive,

although, primary means of evacuation such as stairs are available in the building for

evacuation. Moreover if occupants are waiting for lifts and lift arrives at a later stage

with a full occupant load from upper levels, the waiting population may adopt

competitive behaviour. This may cause the lift car to stop and remain at the floor

(Klote et al., 1993b). Competitive behaviour may depend on several factors that may

include lift waiting period, number of evacuees in queue, building features, number

of lifts, fire protection system and level of the fire-affected floor. Therefore the

aspects of human behavioural response such as ‘decision uncertainty’ and ‘panic’

require an analysis in the research.

2.5 Fire and Smoke Hazards

Vertical fire spread in a building could occur through vertical shafts such as garbage

chute, electrical, communication or/ and plumbing shafts as these shafts are normally

located in public corridor and near to residential unit (risk area). Vertical shafts like

stairs and lifts in high-rise buildings are usually required to be isolated with fire rated

doors (ABCB, 2005) and therefore less prone to fire spread but can be a major path

of smoke spread in buildings. US statistics also demonstrate that smoke spread

through vertical shafts accounts for about 95% of the upward movement of smoke in

typical high-rise buildings (Tamura, 1994). Sixty five percent of the vertical

migration of smoke in buildings occurs through the lift doors and shafts whereas

other building system combined together contribute to the remaining 35% (Tamura

and Shaw, 1976). Seventy five percent of the reported incidents show that the smoke

migrated in the buildings where there is no lift lobby (unprotected lift). Smoke also

migrated in 25% of the reported incidents where there was a lift lobby (protected

lift). These statistics were complied by Smoke Guard Corp (USA) using raw data

23

from the US Fire Administration. Other vertical shafts can also act as a passageway

for smoke transfer.

The NFPA data for 1993-97 reveals the location of the victims, which indicates 74%

were intimate with the fire, 20% were on the same floor and 6% were in other

locations (Proulx, 2000). Brennan’s (1999) analysis of US National Fire Incident

Reporting System (NFIRS) fire statistics over a period of 10 years (1983-1993

except 1986) found the number of victims outside the room of origin was

approximately 308 with a maximum of 478 (48 per year). The total number of

victims in apartments was 3,126. Therefore it is estimated that 15% of the victims are

due to occupants attempting to escape. An example of smoke spread in the building

is the MGM Grand (85 fatalities), where all the 61 victims succumbed to smoke

inhalation and asphyxiation. Of the 61 victims, 25 were found in rooms, 22 in

corridors, 9 in stairways and 5 in lifts (NFPA, 1982). Incidents were also reported

where 60% of the hotel guests had moved in to the smoke filled environment by a

distance of more than 21 m. Half of those guests moving through smoke estimated

that visibility was only 1.2 m or less. Two thirds reported turning back when

visibility was 1.5 m (Pauls, 1995). Pauls (1995) also reported that the British

population turned back, when the visibility distance decreased (see Table 2-8). The

visibility distance has been rounded up after converting from British units to SI units.

Table 2-8: Effects of Visibility distance on Turning Back during Smoke in Corridors

(Pauls, 1995)

Smoke can travel at 0.5 m/s - 2.5 m/s under fire conditions. While escaping the fire

affected unit, residents often leave their SOU door in open position (Willey, 1973).

This allows a significant amount of smoke to enter the public escape path. Residents

24

have lost their lives as a result of doors being in the open position. In the Baptist

Towers Home for the Senior Citizens fire (Willey, 1973), the door of a fire-affected

unit was left open, which resulted in 10 causalities in the building. Similarly in the

Rockefeller Park Towers Fire (Bell, 1983), the door was left open resulting in five

causalities. The estimated reliability of passive protection is 90% for construction

with openings (with self-closers like unit doors) under pre-flashover and flashover

conditions (FCRC, 1996). In reality, the door closure mechanism may also fail due to

which the fire door may be in open position. Therefore, the estimated reliability of

door would be much lower value.

Lift shafts are considered to be one of the major paths of smoke migration to other

floors. Although lift cars are fire rated, but the gaps provided to allow trouble free

operation of the doors may result in large quantities of smoke leakage. ASME A17.1

(2000) and AS 1735.1 (2003) permit a maximum gap of 6.5 mm between lift landing

door and frame, and the leakage area calculated is in the range of 0.045 to 0.065 m2

per door. The leakage area of the lift doors is the primary factor in causing smoke to

migrate to upper floors in a building. Construction openings also contribute to the

spread of smoke to upper floors. Smoke movement is further influenced by wind

speed, stack effect and piston effect. All these factors are related to the presence of

leakage area, thus smoke can move considerable distance in buildings.

The action of wind is an important feature in the movement of smoke through lift

shafts. The wind speed is a function of height above the ground at a time, being

nearly zero at ground level and gradually increasing with height. Hence, a high-rise

building will have the major volume of air to follow its path, which causes positive

and negative wind pressures on either side of the building. Window glass breakage

may aggravate the scenario. Roytman (1969) noted that a room gas temperature of

around 300°C is needed for glass breakage to occur.

The stack effect occurs whenever there is any temperature differential between

exterior and interior of a building. Usually temperature difference exists between

interior and exterior atmosphere during fires and stack effect plays a vital role in the

25

smoke movement in lift shafts. The stack effect is significant in high-rise buildings

although it occurs in small buildings too.

Lift movement causes piston effect, which could increase the smoke spread instantly

in the lift shaft. Due to piston effect, smoke may be pulled into and pushed out of the

shaft. Experiments were conducted in a 15 storey hotel in Mississauga, Ontario to

investigate the piston effect and evaluate the model (Klote and Tamura, 1986). The

maximum pressure differential between floor level and lift shaft was measured to be

16 Pa at floor level of the top floor, which gradually decreased as the lift car

approached the ground floor. This value indicated a flow from the building interior

through the lift lobby and into the lift shaft. Analysis of the experimental data

together with modelling results yielded the conclusion that lift piston effect was of

significance only for single car shaft (see Figure 2-6) and could be ignored in the

case of a multiple car shaft due to open peripheral space.

Figure 2-6 – Piston Effect: Lift acts as a Piston for Smoke Movement

Lift cars may be exposed to fire. Lift doors may be in the risk area and may be

damaged in the event of a fire. The doors can warp and do not open and close freely

resulting in loss of the lift for evacuation from fire-affected floors. If the lift lobbies

are not enclosed, smoke and hot gases can flow into the lobby and travel in the lift

shafts (see Figure 2-7).

26

Figure 2-7 – Smoke Infiltration and Heat Exposure to Lift Landing Door

In one of the tests conducted by Bennetts et al. (1999), a fire was ignited in front of

the lift shaft to observe the temperature inside the lift shaft in a two storey building.

Peak heat release rate was approximately 9 MW in 18 minutes after the ignition. The

maximum temperature of the door reached about 1,000°C. The temperature inside

the lift shaft reported to be not more than 160°C. The test determined the maximum

temperature inside the lift shaft for providing the fire rating to the steel components.

Other parameters such as smoke leakage and spread of toxic gases were not

measured. The condition of lift door after the fire test is shown in Figure 2-8 whereas

the temperature inside the lift shaft for air and rear wall is shown in Figure 2-9.

Figure 2-8 – Lift Door after the Fire Test (Bennetts et al., 1999)

27

Figure 2-9 – Temperature inside the Lift Shaft (Bennetts et al., 1999)

Steel distortion at elevated temperature could increase the door gap and may increase

the flow of hot smoke. Tamura and Shaw (1976) measured the air leakage rate

through lift and stair doors and found that at a pressure differential of 75 Pa the air

leakage through lift door was determined to vary approximately linearly with the

width of the crack between the door and doorframe. For a crack width of 2.0 mm, the

air leak rate per door was measured as 0.1 m3/s. For a crack width of 7.0 mm, it was

measured as 0.45 m3/s.

The Building Code of Australia (ABCB, 2005) prescriptive requirements specify

pressurisation for the fire isolated exits in high-rise apartment buildings. The most

common forms of smoke control in apartment buildings are the pressurisation of stair

shafts (this provision is not for lift shafts). A study of smoke control reliability by

Zhao (1998) through a fault tree analysis found that zoned smoke control system has

a reliability between 52% and 62% for buildings between 5 and 20 storeys and stair

pressurisation system has a reliability of about 90%. Smoke may also be controlled

with public corridor pressurisation. In a fire that occurred in Carlyle Apartment

(Taylor, 1975), the door from fire unit to the corridor was burnt. Pressurisation

system was so effective, evacuating residents were able to walk past the apartment

and observed the burning inside the unit.

During the normal operation, the temperature generated inside the lift machine room

is governed by the lift code AS 1735.1 (2003). This code requires that the

temperature in the lift machine room (LMR) should not exceed 43°C (although

28

commercial chips are rated to 70°C). During a fire occurrence, the lift shaft may

carry hot smoke to LMR though floor openings and causes rise in temperature of

electronic components. The higher temperature may cause lift software failure,

which may inadvertently bring lift car to the fire-affected floor. Hence, the provision

of venting in lift shafts is generally incorporated in the codes. Many codes state that

holes in the machine room floor are only permitted for the passage of ropes, cables or

other moving lift equipment and are limited so as to provide no greater than 51 mm

clearance on all sides. Some codes permit venting by means of floor grates into the

machine room with mechanical ventilation to the outside. ASME A17.1 (2000)

recommends the provision for protecting the LMR; cable slots and other openings

between the LMR and lift shaft are to be sleeved from the machine room floor to a

point less than 30 cm below the lift shaft vent (see Figure 2-10).

Figure 2-10 – LMR Protected with Sleeves (smoke dissipates through vent)

It is noteworthy that ASME A17.1 (2000) provides adequate resistance to smoke

spread to LMR, but it does not provide adequate smoke protection for main lift

shafts. Without protecting the lift shafts for safe use of lifts, the necessity of LMR

protection is of little use. If lifts remain functional during hazardous conditions in the

29

lift shafts, in such circumstances if evacuees use lifts, they may be exposed to smoke,

heat and toxic products.

The standards require the provision of protected lift lobbies for emergency lifts only.

The emergency lifts are permitted to be used by fire fighters only, but not by general

public. This aspect needs further research and analysis for providing lifts as an

emergency exit for general public. Influences of wind and stack effect on smoke

spread through lift shafts need to be addressed as random variables in the risk

assessment of lift systems.

2.6 Toxicity of Fire Effluents

Toxicity of fire effluents is expressed in terms of time-additive values. Time-additive

values account for the effect of exposure to a particular gas (or gases) over a period

of time rather than an instantaneous exposure. Toxicity of fire effluents is measured

in terms of the fractional effective dose (FED). The FED for a constant concentration

(C) of a toxicant product is the dose received up to exposure time (t) divided by the

dose required to cause incapacitation or death (Ct). The FED are calculated by the

(1) mass loss models; or (2) toxic gas models.

2.6.1 Mass Loss Models

Mass loss rate is determined either by direct measurement of material in the fire test

or by mathematical modelling. The tests, operated under condition relevant to those

in the fire, supply the lethal mass loss exposure dose expressed in g min/m3. The

FED is calculated for each small time interval. Continuous summation of the FED is

carried out to calculate the total accumulated exposure dose of a toxic species. If

several materials are involved in a fire, the FEDs of each material are summed. A

number of methods for applying this approach have been developed and two of them

are given below:

• Purser mass loss FED model (Purser, 2002) is based on mass loss burning

rate (kg/min) and dispersal volume (kg/m3). The FED for toxicity by mass

30

loss is the summation of the exposure dose of toxic gases, based on an

average smoke toxicity lethal concentration time product of 300 g min/m3.

• Hazard I model (Peacock, et al., 1991) assumes that the smoke toxicity of the

vast majority of combustible materials is virtually the same. The FED for

toxicity by mass loss is the summation of the exposure dose of toxic gases,

based on an average smoke toxicity lethal concentration time product of 900

g min/m3. ISO TS 13571 states that this value is valid for well-ventilated pre-

flashover fires and that half of that value is valid for vitiated post-flashover

fires.

However, these two models do not distinguish between the different effects of

individual fire gases and derives an estimate of toxic potency from the overall lethal

effects of a toxic effluent mixture.

2.6.2 Toxic Gas Models

This method is based on the composition of combustion products and its toxic effects

as a function of time. Generally, a small number of combustion products is

considered for calculating the FED for each toxicant product. A number of methods

for applying this approach has been developed:

• Hartzell toxic gas FED model (Hartzell, 2001) is based on the concept of

upper and lower limits of exposure on occupants to toxic fire gases.

Exposure in excess of an upper limit would be expected to result in serious

harm to a significant number of occupants, while exposure below a lower

limit should ensure that essentially all occupants would be safe from harmful

effects. FED is described as a cumulative effect of exposure to asphyxiant

(or narcotic) gases and is expressed as (Hartzell, 2001):

( )t

Ct

CFED i

n

i

t

t

∆= ∑∑= 11

2

1

2. 1

31

where Ci is the concentration of the asphyxiant gas i in ppm and (Ct)i is the

specific exposure dose in ppm-min required to produce incapacitation.

Equation 2.1 can be written in terms of CO and HCN as:

[ ] [ ]∑∑

∆+

∆=

2

1

2

1)(

t

t HCN

t

t CO Ct

tHCN

Ct

tCOFED 2. 2

Fractional Effective Concentration “FEC” approach is used to calculate the

risk associated with irritant gases (Hartzell, 2001). A total FEC for effects

due to irritant gases, being cumulative, is shown in the following equation

(Hartzell, 2001):

[ ] [ ] [ ] [ ] [ ] [ ]i

i

NOSOHFHBrHCl IC

Irritant

IC

NO

IC

SO

IC

HF

IC

HBr

IC

HClFEC ++++++= ..........

22

22 2. 3

It is noteworthy that the setting of safe exposure criteria at higher FED values

does not provide much additional evacuation time in a rapidly growing fire

scenario. An FED value of 1.0, at which point many occupants are likely to

be overcome by smoke, is reached within only about 2.4 min, following the

0.1 FED safe criterion (Hartzell et al., 1985).

• N-gas model (Levin et al., 1987) addresses the lethal interactions in rats of

four gases (CO, CO2, HCN and low O2). It does not allow for the integration

of changing concentrations with time. It is used largely for 30 minutes

exposures to a constant concentration. It is useful to determine the extent of

which, rats lethality can be explained in terms of the four common gases.

• Human incapacitation model (Purser, 2002) is applied to actual physiological

uptake function and to the effects of major toxic fire gases. It is designed to

predict toxic hazard in terms of exposure dose and time to incapacitation for

humans in fires. This model is intended for use in the current research.

32

2.6.3 Human Incapacitation Model

Purser (2002) addresses tenability limits for smoke toxicity by asphyxiant and heat or

other thermal effects. In his approach, asphyxiant was separated from irritants.

Asphyxiant is addressed by summation of the exposure dose of the individual toxic

gases, based on their individual concentration at each time period. The safe escape

criterion based on the asphyxiant toxicants CO and HCN would most appropriately

be one-tenth of the dose known from experiments to incapacitate non-human

primates. The hypoxia low concentration of oxygen (< 10%) and high level of CO2

(>5%) exacerbate the effect of asphyxiant. CO2 increases the rate of uptake CO and

HCN (hyper-ventilation). People suffering from cardiac failures and respiratory

diseases exhibit greater sensitivity to the effect of asphyxiant (Purser, 2002).

Asphyxiant hydrogen cyanide (HCN) is important if the burning material contains

nitrogen. However, the ultimate effects of the both the asphyxiant are similar, the

pattern of toxicity during the early stages is different (Purser, 2002). The onset of CO

intoxication is slow and insidious, HCN intoxication is rapid and dramatic (Purser,

2002). However, HCN is not routinely measured as a part of post-mortem process.

The effects of irritant gases are determined from the mass loss of material divided by

the volume of air into which the material is dispersed. Irritant fire products cause

painful effects to the eyes and upper respiratory tract and sometimes to lungs.

Irritants are generally not regarded as presenting an initial threat to escape. The

effects of low concentrations of irritants can best be considered as adding to the

obscuration effects (visibility) of smoke by producing mild eye and upper respiratory

tract irritation (Purser, 2002).

Purser (2002) uses the tenability criteria for radiant heat and for convective heat.

Tenable conditions within the building are assumed to be maintained, provided the

hot smoke layer remains at 2.1 m above the floor. Radiant heat flux greater than 2.5

kW/m2 signals the onset of hazardous conditions for occupants. An average hot layer

temperature of 200°C for shallow layers is used to determine criteria for signal that

the heat flux value of 2.5 kW/m2 has exceeded. A hot layer height less than 1.8 m

above the finished floor level of temperature 60°C is the onset of unsafe condition

33

(FCRC, 1996 and Purser, 2002). Radiant heat and convective heat due to temperature

are dose related and FEDs are determined.

2.6.4 Visibility

Lack of visibility does not have a physiological effect. However, even a small

quantity of smoke reduces visibility, and the reduced visibility can lead occupants

failing to find the way out. Thus, occupants may be trapped and suffer the effects of

other fire effluents. Exit sign of light emitting type or light-reflecting type plays a

vital role in visibility. The empirical relations have been established to determine the

visibility distance (V) for exit signs (Jin, 2002):

)()10~5(m

CV

s

= for a light-emitting sign 2. 4

)()4~2(m

CV

s

= for a light-reflecting sign 2. 5

where Cs is extinction coefficient (1/m). The visibility of objects such as walls or

long corridor varies depending on the interior of its contrast condition; however,

minimum value for light reflecting signs may be applicable. The extinction

coefficient should not be higher than 0.5 m-1 for a visibility distance of 4 m.

It is assumed that smoke visibility and heat have no effect on FED for asphyxiant.

Although some effects are likely, no quantitative information is available. Beyler

(2004) concluded that visibility criterion is reached before carbon monoxide hazards

arise in most egress situations and soot & CO yields are well correlated.

2.7 Lift Operational Systems

2.7.1 Overview

Two types of lifts are normally used viz. hydraulic and electric. The hydraulic lift

pushes the car up on a shaft filled with compressed oil whereas the electric motor is

used for lifting car up or down in electric type lift. Hydraulic lifts are slower than

34

electric lifts, having a speed less than of 0.5 m/s, whereas electric lifts operate at

faster than 1 m/s (or in excess of 5 m/s) (Straskosch, 1998). Hence, hydraulic lifts are

suitable for low-rise buildings while traction type electric lifts are suitable for mid or

high-rise buildings.

The traction type lifts suspend the car by cables from the pulley system, using

counterweights to minimize energy expenditure. The counter weights are normally

equal to 50% of full load car capacity. The traction pulley may be attached to

driving motor directly or through gear according to which the lift is called gearless or

geared type. Gearless design is employed in high speed systems (2 m/s and above)

whereas gear type lift is used in slow speed systems (0.125 m/s to 2.3 m/s)

(Straskosch, 1998).

To control the weight capacity, load sensors located under the floor can weigh each

car after stops and when the car has reached its maximum capacity, the controller

will often ignore incoming calls until enough passengers leave. The capacity limits

are intended more for passenger comfort and the adjustment of counterweights. Once

the controller senses an unplanned release from the main cables, a set of emergency

brakes is immediately deployed. The car will not move from that location by itself.

Even if the brakes did fail mechanically, the counterweights would assist to slow the

car's descent (ASME A 17.1, 2000).

Building codes control the maximum number of lifts permitted in a single shaft in

order to limit the potential of a fire disabling all lift services in a structure. Most

building codes permit three or fewer lifts in the same shaft enclosure. When there

are four lifts, they must be in at least two separate shaft enclosures. When there are

more than four lifts, not more than four can be in the same shaft enclosure (AS

1735.1, 2003).

2.7.2 Lift Dispatch Control

The problem of down peak traffic has been addressed by efficient lift dispatching

strategies, which decide where to go and which should be served first. In practice, lift

dispatchers are designed heuristically and evaluated on simulated buildings.

35

Passenger arrivals are modelled as discrete, stochastic events, with arrival rates

varying frequently over the course of a simulated day. It was possible to simulate the

evacuation by lifts using control with intelligence for the evacuation. However, it

will be a major challenge to develop a dispatch control policy for evacuating the

occupants systematically and intelligently above the fire-affected floor. The lift

group controller implements dispatch rules that decide where the cars should go and

stop. The currently used modern approach of genetic algorithms based lift group

control system is utilized to multi-objective optimization in a dynamically changing

process control environment (Tyni and Ylinen, 2006). The chromosome is built up

by taking the landing calls one by one and inserting them into the chromosome as

‘‘call genes’’. This coding approach offers a natural way to meet the requirement that

each call should be responded to only once by one of the lifts. The genetic algorithm

allows a better performance attending to the system waiting times than the traditional

duplex algorithms. The system can be effective during the heavy traffic conditions,

which can invariably be seen during fire emergencies. The latest technological

advancements such as neural networks, fuzzy logic and genetic algorithms provide

better performance (Cortes et at., 2003) and can be useful for lift emergency

evacuation.

2.7.3 Concerns of Lift Operational Mechanism

If lifts are considered for building evacuation, they must be reliable for safe

operation. But the most common problem is the temporary electric power failure.

Statistics reveals that power failure rates in urban Australian locations are

approximately three outages of 10 minutes duration per annum, i.e., a failure rate of

5.7 × 10-5 per annum (Lacey, 2000). (The failure rate is the duration of power failure

per annum). However, power failure may occur during fires since fire fighters cut

the power supply for fire fighting operations in order to avoid possible electrocution.

Power failure may occur due to fire in the electrical system or water damage in the

lift system. The water damage occurs due to fire fighting operation and water based

fire fighting installation. The risks relating to lift operational mechanism need to be

evaluated during fire emergencies.

36

2.7.4 Lift Protection Measures

The lifts are provided with protection measures to avoid electric supply failure.

Emergency power supply of the same voltage characteristics of normal power via the

normal feeder to run the lift system is required (ASME A 17.1, 2000). Transfer

switch is provided with an adjustable time delay of approximately 20-60 seconds for

pre-transfer signal in either direction. The electric supply for the lifts is provided on

separate circuit from the main switch rooms and is taken through armoured cable

separately through respective lift shafts in fire protected route. Automatic Rescue

Device (ARD) meant for the purpose of bringing the car to the nearest landing door

up or down (depending upon the load condition), are also available. ARD normally

activates rescue operation within 10 seconds of normal power supply failure and

operates with the help of battery back up. For a large number of lifts, there is a

winch in the lift machine room at the top of the building that the fire service use to

lower the stopped car to the next floor down. Even with the gearing involved, this is

a very strenuous manual activity and the fire fighters have to take turns at it (HMSO,

1993).

Protected lift lobbies are required according to ASME A17.1 Safety Code to avoid

water damage. Lifts are designed so that water entering hall fixtures will not shut

down the lift when on Fire Phase II operation, which is generally resolved by

providing water resistant components and preventing water entering the lift shaft.

ASME A17.1 also requires the shutdown of power to the lift upon or prior to the

application of water in lift machine rooms or hoist ways (though fire in lift machine

room or lift shaft shut may not require complete evacuation of building). This

shutdown can be accomplished by a detection system with sufficient sensitivity that

operates prior to the activation of the sprinklers (NFPA 72, 2002).

2.7.5 Sprinklers in Residential Buildings

Based on numerous studies, the reliability of fire sprinkler systems has been

documented. The fire sprinkler system reliability ranges from 81.3% to 100%

representing a significant range of performance. The following table summarizes the

37

performance of automatic sprinklers in residential occupancy group during the period

1886 to 1986 (Marryatt, 1988).

Table 2-9: Performance of Automatic Sprinkler System in Residential Buildings

(Marryatt, 1988)

Thirty three fires were reported in apartment type buildings that are sprinkler

protected. Marryatt reports that 100% of fires that occurred in single-family unit

dwellings were controlled by sprinklers. Marryatt reports an average of 1.22

sprinklers in operation for this type of occupancy with no fatality reported. Whilst

the scope of the research does not solely relate to ‘no fatality’ scenario, it is

considered that there would be nonfatal and fatal injuries in apartment buildings due

to the presence of smoke and asphyxiant gases. It must be noted that the higher

reliability of fire sprinklers reported by Marryatt of 100% reflect fire sprinkler

systems where inspections, testing and maintenance were exceeded normal

expectations and applicable generally to installations in Australia and New Zealand.

The study by Bukowski et al. (1999) found the reliability of sprinklers to be 96.6%

for residential occupancy and 94.6% for overall occupancy. The NFPA statistics

(Rohr, 2001) for ten years reporting period from 1989 to 1998 indicates the

operational reliability of automatic sprinkler systems for apartment buildings is

87.6%. Bukowski et al. (1999) and NFPA findings demonstrate the sprinkler

performance on “real” fires that have occurred inside apartment type buildings due to

which there may be nonfatal and fatal injuries in apartment building.

38

2.7.6 Evaluation of Lift Evacuation Time

The performance of lift systems depends on average waiting time and handling

capacity. Handling capacity is the number of people served in a given period during

a round trip time (RTT) for a lift. Handling capacity is calculated for traffic up-peak

in five minutes and represented as HC. RTT is the average time taken for a single

lift to complete its cycle, i.e. receiving passengers on the ground floor, releasing

passengers on the upper floors and returning to the ground floor. RTT is used as the

basis for estimating average waiting time. The evacuation time consists of the sum

of all the RTT divided by the number of lifts plus the time needed to start up the lift

evacuation and the travel time from the lift lobby to the outside (or to another safe

location). Accounting for inefficiencies of lift operation, this evacuation time can be

expressed as:

t J

) + ( + t + t = t j r,

m

=j

oae ∑1

1 η 2.6

where

tr,j is the time for round trip j

m is the number of round trips

J is the number of elevators

η is the trip inefficiency (typically 0.1 for rounding off of probable stops, door

operating time, door starting and stopping time and the unpredictability of people)

ta is elevator evacuation start up time

to is the travel time from the elevator lobby to the outside or to another safe location.

The round trip time tr is can be written as:

t + t = t sTr 2 2.7

where

ts is the standing time

tT is the travel time for one way of the round trip

The standing time is the sum of the time to open and close the lift doors twice, the

time for people to enter the lift and the time for people to leave the lift. The trip

inefficiency accounts for trips to empty floors and trips to pick only a few stragglers.

39

ELVAC (Klote et al., 1991) model is used to determine the lift evacuation time (see

Section 2.10). Traffic condition varies with time in apartment buildings too. Down-

peak traffic substantially increases in the morning when people are going to work

and up-peak traffic can be noticed during evening hours. Down peak traffic would be

similar to a fire condition, if lifts are used for evacuation. The down-peak traffic is

1.5-1.8 times more efficient since the occupants will be using the same lift.

Generally acceptable waiting time for the apartment buildings is 50 to 80 seconds

with 5% to 8% of population handled during a 5 minute period.

2.8 Stair Evacuation System

2.8.1 Overview

Evacuation time depends on the availability of number of stairs in the buildings. The

number and width of stairs depend on the floor population and travel distance

requirement and mainly governed by the building codes. Normally two fire-isolated

stairs are the minimum requirement for a high-rise building. The BCA 2005 DTS

provisions state that the fire-isolated stairs are required to be protected with an open

access balcony, a smoke lobby or to be pressurised. The stair evacuation facility can

be protected with dual system i.e. double compartmentation (with smoke lobby) or

compartmentation and pressurisation.

2.8.2 Evaluation of Stair Travelling Time

The travel time for building evacuation is calculated for the evacuees in the building

since it depends on alarm, sign of danger and time of day. The computer models for

stair evacuation give optimistic results since they assume that stairs do not become

overcrowded. When the occupant density is more than 3.5 persons per square meter,

flow is very congested and slow (Buchanan, 2001). Melinek and Booth (1975)

estimated the evacuation scenarios in two categories. In one case, stairs are

congested and there is virtually no flow. In another case, occupants walk freely. The

stair travelling time is determined using the following equations (maximum value).

40

s

s

tWF

nNt +=1 2.8

s

s

n ntWF

Nt += 2.9

where

t1 is the egress time (congestion)

tn is the egress time (free walk)

n is the number of floors

N is the number of people per floor and exit

Fs is the nominal occupant flow on stairs (persons/meter/second)

W is the effective width of the stairs

ts is the walking time between adjacent floors

Melinek and Booth (1975) recommenfed a typical value of 16 seconds for ts and 1.1

persons/m/s for Fs. However, ts is greatly increased in the presence of elderly

persons. Their walking speed in the stairs is generally 0.5-0.8 m/s. Therefore ts is 30

seconds for a combined population of aged and young occupants. The value of Fs is

0.5 persons/m/s. The expressions indicated above are used for verification purposes.

The difference between the use of stairs and lifts for occupant decent ranged between

15 to 30 minutes. This is a large difference in time lost to travel by stairs, especially

when a fire can grow significantly in a matter of minutes. In 15 minutes, the

environment can be less toxic for the occupants, smaller fires and the property less

damaged. Nelson and MacLennan (1995) indicate that, for stair evacuation, actual

evacuation time can be two or even three times as long. The stairs is the fastest way

of evacuation for low-rise and mid-rise buildings. But the evacuation time by lifts

becomes more favourable for building with more than 100 persons per floor and with

more than 50 floors (Siikonen and Hakonen, 2003).

41

2.9 Risk Assessment Methods

Larsson (2000) classified fire risk analysis into three categories i.e. regulations and

checklists, ranking methods and probabilistic (or quantitative) methods. Regulations

and checklists involve limited risk assessment techniques. A review of ranking and

probabilistic methods is given below:

1. Ranking methods have been developed with the purpose of simplifying the risk

assessment process. Ranking method involves identification of every single

factor that affects the level of safety and data gathering. The importance of

each factor has to be decided by assigning a value (or called weight). This

value is based on the knowledge and the experience of experts. Assigned

values are operated by combination of arithmetic functions to achieve a single

value. The value is a measure of the level of risk and it is possible to compare

this to other similar designs and to a stipulated minimum value. The following

ranking methods are commonly used:

• The Multi-Objectives Decision Analysis (MODA) provides a systematic

approach that addresses several factors that complicate the selection

decision in any simulation design situation. The MODA is also called

Multi-Attributes Decision Analysis. There is not much work done on the

MODA approach in fire safety risk engineering. However, this approach is

used in economics, nuclear energy and resources, policy analysis, scientific

research management, industrial management, manpower planning and

medical diagnosis and defence. The applications in a variety of areas

demonstrate that decision analysis continues to be a widely used approach

for a variety of strategic and tactical decisions (Keefer et al., 2002). The

MODA approach is used in this research project (see Chapter 7).

• The Analytical Hierarchy Process (AHP) operates by using pair wise

comparison judgment to consider factors which are not effectively

quantified. This process is similar to MODA. Factors subject to

uncertainty, ill-defined parameters, conflicting objectives and inexactness

in measurement may be considered with the judgmental process (Saaty,

42

1980). However, the judgment can be highly variable and difficult to work

with. One may study the consistency of judgment and its validity under

certain contexts. This approach is used in a variety of areas, as referred in

MODA. The AHP is used in combination with MODA in this research

project (see Chapters 3 and 7).

• The AHP evolved to an application of Fire Risk Index Method (Watts,

2002). The University of Edinburgh developed this approach initially for a

study to improve the evaluation of fire safety in the U.K. hospitals through

a systematic method of appraisal (Watts, 2002). Risk Index can be used to

describe the impact on the building structure. The Delphi Panel awarded

weights to various related attributes (the Delphi Panel, consists of a group

of experts, who never meet physically and all communication is through a

group controller). The approach yields an effectiveness of a specific design

solution for a given objective and provides a rational basis for possible

design improvements. Earlier reports have described the development of

the Fire Risk Index method and demonstrated that the Fire Risk Index

method can be a very useful tool (Larsson, 2000, Karlsson, 2000, Hultquist

and Karlsson, 2000). However, the Fire Risk Index analysis may not be

appropriate, where greater sophistication is required (Watts, 2002).

2. Probabilistic methods provide quantitative values, typically produced by

methods that can be traced back through explicit assumptions, data and

mathematical relationships to the underlying risk distribution.

• An event tree is a graphical logic model that identifies and quantifies

possible outcomes following an initiating event. The tree structure is

organized by temporal sequence. Probabilities can be calculated and

consequences are assigned to the end states along the tree. Each path

through the event tree defines a scenario. Various outcomes for state of

functioning/ non-functioning of event/ system can be shown. Complex

process can be analysed by modifying event tree into a hybrid combination

of parallel and series system (a complex parallel and series system). Basic

event tree approach is used in fire safety engineering under varying

43

conditions of fire detection, sprinkler operation, door opening and closing

and fire brigade intervention for determining the probability of fire

extinguishment. This approach is not suitable for modeling the building

evacuation under the dynamic scenario (temporal fire and smoke spread).

• Fault-tree provides a simple graphical model based on circuit diagram that

can be used to analyse potential errors in a design. Fault tree is constructed

from events and gates. Fault tree begins with basic events, which represent

the underlying failures that lead to an accident, to top event (outcome).

Numerical probabilities of occurrence are entered and propagated through

the tree to evaluate probability of the foreseeable, undesirable event. This

approach is widely used in electrical engineering for risk assessments in

electrical power stations and provides credible results.

• Safety Index Method involves a complex way of evaluating the level of

fire risk. The advantage is the precision of the results. The risk orientated

analysis started with the selection of potential fire hazards, which could

endanger occupants inside the building. The safety index is described in

terms of escape time margin. The disadvantage with the safety index β

method is that it only considers the escape time of the last person reaching

the safe area for describing the probability of failure of evacuation route

(Frantzich, 1997a). The probability is only addressing the fact that persons

are unable to evacuate safely, i.e. the time margin is less than zero

(negative safety index). No information is available of the probability that

exactly one or two or more are unable to escape (Frantzich, 1997a).

However, this method is effective for a comparative design analysis.

• Monte Carlo technique is used for the risk assessment having stochastic or

probabilistic basis. The Monte Carlo analysis can provide the distributions

of the output variables and their sensitivities to the input variables. Typical

outputs are, for example, the times of component failure, fire detection and

flashover (Hostikka and Rahkonen, 2003).

44

Table 2-10 gives an overview of the common risk assessment methods listed above

and their advantages, disadvantages and ability to meet the research requirement.

Table 2-10: Risk Assessment Methods

Method Advantages Disadvantages Meets the Research

Requirement

Multi-Objectives Decision Analysis

Comparable; Used in multi-element system; resolve conflicting issues

Lack of statistical resources

Yes; Used for addressing the risks

Analytical Hierarchy Process

Comparable; Used in multi-element system; resolve conflicting issues

Hierarchical relationship; Lack of statistical resources

Partially used in-conjunction with MODA

Risk Index Method

Comparable; Used in multi-element system

Workforce requirement (Delphi Panel)

No

Event Tree Analysis

Easy to demonstrate state of functioning/ non-functioning

Not suitable for modelling dynamic scenarios

Yes; Modified form of complex parallel and series system is used.

Fault Tree Analysis

Used in multi-element system/ process; Identifies all possible causes of a specified undesired event

Lack of statistical resources; Not suitable for modelling dynamic scenarios

Yes; Used for reliability study

Safety Index Method

Comparable; Provide precise results

Negative safety index as it considers the entire population for safe margin

Yes; Used for addressing the unsafe conditions in buildings

Monte Carlo Simulation

Comparable; Provide precise results

Time consuming; Lack of statistical resources

Yes; Used for addressing the various probability distribution functions

2.10 Application of Computer Models in the Research Study

Computer models are mainly identified as zone model, field model and egress model.

Field and egress models are used in this research study. A brief is given below.

The zone model ‘CFAST’ (Peakcock et al., 2004) divides the compartment into two

zones and solves the conservation equations within the individual zones, whereas a

field model ‘Fire Dynamic Simulator’ FDS (McGrattan et al., 2004) divides the

compartment into a large number of volumes and solves the conservation equations

within the individual volume and therefore is more complex. The zone model

45

‘CFAST’ is used for peak temperature and smoke transport calculations and it

assumes that the compositions of layers are uniform, thus, the temperature and other

properties are the same throughout each layer (Peakcock et al., 2004). This

assumption is less valid for very large spaces or long narrow spaces such as stairs

and lift shafts.

A computer simulation developed by the NIST called CONTAMW (Walton, 1993)

was used to analyse the effect of compartmentalization strategy in buildings. The

CONTAMW (a multi-zone, multiple floor airflow network analyser) is able to model

wind and stack effects and can also predict smoke movement. The CONTAMW

simulations used in combination with zone model (CFAST) showed that smoke

movement was significantly reduced by the compartmentalization strategy (Klote,

2003). Klote (2003) stated that future research is needed to evaluate the extent to

which compartmentation failure would impact smoke flow through lift shafts.

However both softwares had limitations in considering non-deterministic variables

with the flow of fire, smoke and hot gases in long shafts. With the intent of further

extending the research, field model ‘FDS’ is used in the current study.

ELVAC (Klote et al., 1991) presents the analysis of people movement by lifts in the

buildings, which incorporates more details about lift motion and lift loading and

unloading. The model is developed based on mathematical derivations described in

lift engineering (see Section 2.7.5). However, ELVAC model does not incorporate

horizontal components (deterministic and non-deterministic movement) on upper

levels. Software Building Traffic Simulator (BTS), which is an advanced version of

Advanced Lift Traffic Simulation (ALTS) (Siikonen, 1989), is also used in lift

industries for simulating passenger traffic in the buildings (Hakonen, 2003). BTS is

designed with the purpose to analyse the performance of a lift system, to demonstrate

lift systems for customers and to test lift group control software, but this software has

also the same limitations as in ELVAC.

EVACNET (Kisko et al., 1998) can model building evacuations using stairs and lifts.

The program accepts a network description of a building and information on its

initial contents at the beginning of the evacuation. From this information, EVACNET

produces results that describe an optimal evacuation of the building. Each evacuation

46

is optimal in the sense that it minimizes the time to evacuate the building. Model

incorporates several service levels for walkway and stairway. The level of service can

be assumed to be 50-95% of maximum capacity in the stairs, where speeds are

restricted, passing is virtually impossible and reverse flows are severely restricted.

Both the models (ELVAC and EVACNET) used in evacuation modelling are

deterministic models and these models can not predict the realistic situations such as

occupants’ pre-movement and movement activities. The problem is resolved by

using industrial application SIMAN ARENA (Rockwell, 2000) discrete event

simulation model. Discrete event simulation is a type of simulation where occupants’

pre-movement and movement activities can be incorporated by probability

distributions. This model is used in determining the lift waiting time, lift evacuation

time and the number of occupants in queue. Simulation models for lifts and stairs

have been prepared so that the issues relating to human behavioural response (such

as uncertainty and panic) and life threatening conditions can be adjudged and

compared. However, this software does not simulate fire.

For risk assessment, @RISK package (Palisade Corp, 1996) is used based on Monte

Carlo technique, where all the parametric probability distributions are analyzed for

determining the risks. Uncertain input values in spreadsheet are specified as

probability distributions. An input value is a value in a spreadsheet cell or formula

which is used to generate results in spreadsheets. A probability distribution describes

the range of possible values for the input is substituted for its original single fixed

value. Available graphs include probability distributions of possible output variable

values and cumulative probability curves. By dragging the delimiters displayed on a

histogram or cumulative graph, target probabilities can be calculated.

2.11 Discussion and Summary

Stairs are the only evacuation facility acceptable by prescriptive building regulations

for high-rise buildings as lifts are not considered safe during fire emergencies. The

literature review reveals the following points, which warrant the use of lifts as an

emergency evacuation facility in high-rise apartment buildings:

47

• Due to limited physical capability, aged and disabled persons are slow to

walk in stairs. The aged and disabled persons may require lifts for prompt

and safe evacuation from the building during emergency situations.

• People find enormous difficulties in evacuating high-rise buildings. Long

evacuation time may sometimes endanger the life of evacuees or cause

injury due to tiredness, dizziness, slipping on surfaces or becoming less

capable physically.

• Although stair design is compliant with building regulations, the

travelling speeds as a whole may be affected by bottlenecks, turns, and

obstacles in the stairs, all of which can cause crowding at certain points

and hinder timely evacuation of high-rise buildings.

Literature review reveals that people generally act rationally and appropriately

(Proulx, 2003). Normal patterns of behaviour and movement route choices tend to

persist during emergency situations. However, under certain circumstances, evacuees

could become impatient and overcrowd the lifts, which can cause the car to stop

functioning and remain at the floor. Considering that panic may be a rare event and

the outcome of the study of this phenomenon is inclusive, it is assumed that evacuees

may become panic of the dangerous conditions.

Literature review identifies that lifts are exposed to several risks during building fires

that require further works for bringing lifts to a required safety standard (Klote,

1982). Due to a gap between lift door and lift landing frame, smoke can move to lift

shafts and upper levels in buildings. Sixty five percent of the vertical migration of

smoke occurs through the lift doors and shafts, which is further influenced by stack

and wind effects. The lift operational mechanism is also not yet fully reliable. To

provide a safe alternative evacuation system, the risks relating to lift operational

mechanism must be addressed for improving the reliability of lift systems. Literature

review also revealed that a lift evacuation system for a small number of people is

feasible (Klote et al., 1995 and Paul et al., 1991). However, these preliminary

research works have not been quantified.

The issues relating to human response, fire hazards and lift operational mechanism

need to be addressed in an integrated manner to determine the feasibility of lift

48

evacuation systems. The safety of evacuees is to be considered under variable

conditions of fires and human behaviour. If all the fire safety issues relating to lifts

are resolved appropriately, lifts can be considered as an option for emergency

evacuation.

Literature review also reveals that a combined use of stairs and lifts can reduce the

evacuation time considerably during emergencies in mega high-rise buildings (Klote,

et al., 1993a, Siikonen and Hakonen, 2003). Greater evacuation efficiency occurs as

the height of the building increases. However, this aspect does not include the issues

related to the safe use of lifts during fire emergencies. The current research is

performed to address the need to use lifts for evacuation purposes and to answer the

following questions:

• Can protected lifts provide adequate safety for general public during fire

emergencies?

• Is any additional provision required for providing lifts as an emergency

evacuation facility? What can be the acceptable level of risk for lift

systems?

• What is the suitable risk assessment approach for lift evacuation systems?

This research focuses on the above points for exploring the safe use of lifts on a

comparative basis, with stairs, after establishing the inter-relationships among the

potential risks.

49

3. RESEARCH METHODOLOGY

Risk assessment remains a challenging task, especially in view of the uncertainties

related to extreme events exceeding the safety limit criterion of fire safety measures.

Detailed spatial information on risks is extremely important in determining the use of

lifts as an alternative evacuation facility in the buildings. This chapter provides an

overview of research methods, indicating how the subsequent chapters fit together

and demonstrates how an integrated risk assessment method connects to other

evaluative activities. The objectives of this chapter are:

1. To identify the risks associated with the use of lifts for emergency

evacuation and develop an inter-relationship among the risks.

2. To establish a research strategy for an acceptable level of risk and consider

suitable design options and evacuation strategies.

3. To investigate a suitable risk assessment approach for lift evacuation

system.

Risk is defined as the probability of a specific undesirable event occurring in specific

circumstances arising from the realisation of a special hazard (Magnusson, 1996).

Risk is expressed as a function of the probability of an event occurrence and the

consequences of that event occurrence, which can be represented as:

Risk = Probability of Occurrence × Severity of Consequence

Risk increases as a function of the probability of occurrence and the severity of the

consequences. No activity is risk free and therefore a research strategy is adopted to

reduce the probability of occurrence or severity of the consequences or both to

achieve an acceptable level of risk (research methodology for evaluating the risks is

different from this research strategy; research methodology involves selecting

suitable methods whereas research strategy involves appropriate planning).

50

The research methodology involves the following components:

• To identify all the significant risks in the lift evacuation system and develop a

relationship among the risks.

• Rank all the risks (or risk priorities) in terms of likelihood of occurrence and

expected impact upon the building evacuees.

• Establish a research strategy for an acceptable level of risk.

• Identify risk control design options and evacuation strategies for evaluating

risks.

• Quantify consequences with the models and techniques (for example,

stochastic evacuation models, fire hazard models and probabilistic analysis

techniques).

• Conduct risk assessment with a suitable method (Multi-Objectives Decision

Analysis – MODA method is used).

• Select appropriate risk control design option.

The methods and results from this research form the base for comprehensive risk

analysis and risk management strategies. Therefore, an appropriate scale of risks is a

fundamental precondition for a reliable risk assessment. These risks need to be

prioritised and presented in a straightforward, readily understandable format that

shows both the risks and how they are being managed.

3.1.1 Risk Identification

Building regulations worldwide do not permit the use of lifts as a safe mode of

vertical transport system for building occupants during fire emergencies. The

traditional approach of not using lifts as an evacuation facility during fires is mainly

due to the following controversial and unresolved issues (Klote, 1982):

Issues related to human behavioural response

1. Irrational human psychological behaviour

� People may be ‘uncertain’ (or doubtful) of their decision of choosing

lifts for evacuation during lift waiting period; and

51

� People may ‘panic’ in the lift lobby (or in the lift cabin) under certain

circumstances (due to 5, 6, 7 and 8 shown next).

Issues related to fire hazards

2. Smoke and toxic gases spread to lift lobby and vertical lift shaft (vertical

shaft as a major path for smoke and toxic product spread); and

3. Lift passing through fire-affected floor and possible exposure to high

temperature; and

4. Influence of wind speed and stack effect on smoke spread.

� Because of 2, 3 and 4, people may be exposed to ‘life threatening

conditions’ in the lift lobby (or in the lift cabin).

Issues related to lift operational mechanisms

5. Malfunctioning of lift equipment may inadvertently cause the lift to go to the

fire-affected floor; and

6. Effect of water from fire fighting or sprinkler operation on electrical systems;

and

7. Loss of electric power supply during fire; and

8. Lift unavailability due to maintenance.

� Because of 5, 6, 7 and 8, people may ‘panic’ in the lift lobby (or in the

lift cabin).

Figure 3-1 shows the risk cause-effect relationship, developed among the above

issues related to the use of lifts during fire emergencies. The risk issues broadly fall

under three categories i.e. human behavioural response, fire hazards and lift

operational mechanism. These issues can lead to the risks of decision uncertainty (a

psychological impact), panic (mainly psychological impact) and injuries (nonfatal

and fatal – mainly physiological impact).

52

Figure 3-1 – Risks involved in the Lift Evacuation System

53

3.1.2 Analytical Hierarchical Process for Risk Priorities

The Analytical Hierarchical Process (AHP), developed by Saaty (1980), is one of the

more extensively used approaches in multi-objectives decision making methods. The

AHP has been applied to a wide variety of decisions and human judgment processes

(Lee et al., 2001). The AHP involves three basic steps:

1. hierarchical structure; and

2. comparative judgments or defining and executing data to obtain pair-wise

comparison data on elements of the hierarchical structure; and

3. synthesis of priorities or constructing an overall priority rating.

A hierarchy is an abstraction of the structure of the system for studying the

functional interactions of its components and their impact on the entire system. The

abstraction can take several related forms, all of which essentially descend from an

overall objective. Problems under complex conditions are analyzed into a hierarchy

structure. The hierarchical structure is prepared based on previous studies and

empirical experiences. Once a hierarchy has been developed, a pair wise comparison

is needed to determine the relative importance of the elements (or entities) in each

hierarchical level. For a pair wise comparison, the analysis involves

1. developing a comparison matrix at each level of the hierarchy starting from

the second level and working down; and

2. computing the relative weights for each element of the hierarchy; and

3. estimating the consistency ratio to check the consistency of the judgment.

The pair wise comparison at a given level can be reduced to a number of square

matrices A = [aij]nn as:

nnnn

n

n

aaa

aaa

aaa

.........

.

.

........

........

21

22221

11211

54

The matrix has reciprocal property as:

ji

ija

a1

=

In AHP, Saaty (1980) recommended numerical values 1, 3, 5, 7 and 9 for making

subjective pair-wise comparisons (see Table 3-1). Intermediate values between two

adjacent judgments can also be assigned. The increasing numerical values indicate

increasing importance.

Table 3-1: The 9-Point Scale (Saaty, 1980)

The vector weights, w = [w1,w2, . . . ,wn], is computed on the basis of Saaty’s

eigenvector procedure in the following two steps:

a. The pair-wise comparison matrix, A = [aij]nn, is normalized by the

following equation:

∑=

=n

ji

ij

ijij

a

aa

1,

*

3.1

b. The weights are computed by the following equation:

n

a

w

n

ji

ij

i

∑== 1,

*

3.2

55

Then, the weight of all the elements is

11

=∑=

n

i

iw 3.3

If λ is a number and w is a non-zero vector, then w is called eigenvector of A and λ is

the associated eigenvalue. The following equation gives the relation between the

vector weights w and the pair-wise comparison matrix A:

wAw λ= 3.4

This equation can be written as:

( ) 0=− wIA λ 3.5

where I is the identity matrix (with elements on the main diagonal set to 1). Then the

determinant equation is:

0)det( =− IA λ 3.6

The maximum eigenvalue λmax is an important validating parameter in AHP (Saaty,

1980):

n≥maxλ 3.7

The λmax is used as a reference index to screen information by calculating the

consistency ratio (CR) of the estimated vector. When the matrix is perfectly

consistent, λmax equals to n. If the matrix is not perfectly consistent, λmax, is greater

than n. The larger the λmax, the greater is the degree of inconsistency. To calculate

the CR, the consistency index (CI) for each matrix of order n can be obtained from

equation:

56

1

max

−

−=n

nCI

λ 3.8

The CR can be calculated using the following equation:

RI

CICR = 3.9

where RI is the random inconsistency index. The random inconsistency indices for

the matrices of the order of 1 to 10 are given in Table 3-2. If CR < 0.1, the

comparisons are consistent and if CR > 0.1, the comparisons are of inconsistence

judgment.

Table 3-2: Random Inconsistency Indices (Source: Saaty, 1980)

The risks are given priorities from the AHP and a relationship is developed on

numerical scale as given below:

Hierarchical Risk Levels: A triangle-shaped diagram is used to indicate the degree

of hazards associated with risks (see Figure 3-2). The degree of hazard is utilised to

indicate hazard rating. The diagram identifies three colour-coded categories of

hazard for each risk. The risk of decision uncertainty is shown with green, panic

with blue and injuries (fatal and/ or nonfatal) with red colour. The triangle indicates

convergence from low level to high level as the severity of risk consequences

increases to the maximum at the top. However, the area of triangle indicates

divergence from high level to low level as the probability of risk occurrence is the

maximum at the bottom.

57

Figure 3-2 – Risks at Three Hierarchical Levels

Three hierarchical levels are given to the risks for pair-wise risk comparisons. The

risk in hierarchical level represents the dominance of risks at its bottom. The

hierarchic levels are based on the degree of hazards from psychological to

physiological effects on a time based sequence during the evacuation procedure. The

impacts lie on a continuum from little or no effect at low level to relatively severe

incapacitation at high levels, varying in response for different individuals. The risks

at three hierarchical levels and their expected consequences (impacts) are given

below:

• Low Risk (Decision Uncertainty): Uncertainty of making decisions for

using lifts arrives at the time of building evacuation and may be caused by

excessive waiting time for lifts in the lift lobby. Decision uncertainty may be

influenced by the number of evacuees waiting in the queue. Uncertainty is the

state of belief when one is unsure (Reber, 1995) and decision uncertainty

refers to a lack of knowledge about the lift waiting period or pathways.

Decision uncertainty is considered as low level of severity. The expected

impact can be ‘anxiety for information or knowledge’ or ‘mental agony’

amongst the evacuees (psychological effects).

Lev

el o

f Sev

erity

Low

Decision Uncertainty

Pro

bab

ility o

f O

ccurr

ence

Injuries (Nonfatal and/or

fatal)

Panic

Medium

High

Lift Operational Mechanism and/ or Pre-Life Threatening

Condition)

Fire

Hazards

Physiological and/or

Psychological

Cause Level of

Severity

Risk/

Hazard Impact

Psychological and/or

Physiological

Psychological

Lift Waiting Time and/ or Evacuees’ Queue

58

• Medium Risk (Panic): Panic may arrive in the lift lobby and/or lift cabin and

may be due to faulty or unavailability of lift system or visual threat (pre-life

threatening condition). Schultz (1968) defined panic as a fear-induced

behaviour which is non-rational, non-adaptive and non-social, which serves

to reduce the escape possibilities of the group as a whole. Prolux’s (1993)

stress model demonstrated that fear can be induced during emergencies,

which may subsequently convert into panic. Ramachandran (1991) found that

if evacuees are abnormally delayed and they are likely to be exposed to

unsafe conditions, the concern of their life safety is imminent. It could cause

a sense of life threat and inflict panic. Panic is considered as medium level of

severity. Evacuees may adopt competitive behaviour and overload lifts. The

expected impact can be ‘mental agony, physical injuries or deaths

(psychological and/or physiological effects).

• High Risk (Injuries): Life threatening condition may arrive due to hazardous

conditions arising from fire hazards in the lift lobby and/or lift cabin. Injuries

(nonfatal or fatal) are considered as high level of severity. Evacuees may

adopt competitive behaviour and overload lifts. The expected impact of life

threatening condition can be ‘mental agony, physical nonfatal or fatal injuries

due to exposure to smoke, heat and toxic products (psychological and/or

physiological effects).

The levels are conservatively assumed in orders in the lift system. However, panic

may occur first without first causing decision uncertainty among the evacuees.

Likewise, physical injuries may also occur first, without causing decision uncertainty

or panic or no such issues may arise during fire emergencies. The risk levels are

considered independently.

Comparative judgments: The use of lifts during fire emergencies is evaluated in

relation to the psychological and physiological impacts. The psychological and/or

physiological impacts may occur due to the risks from “decision uncertainty, panic

and injuries (nonfatal or fatal)”. However, these risks are related and one risk may

lead to another risk during the lift evacuation procedure. A relationship is developed

among the risks and impacts for a comparative judgement (see Figure 3-3).

59

Figure 3-3 – Hierarchical Relationship for the Evaluation of Risk Priorities

Matrices are developed for assigning the priorities for three levels of risks {decision

uncertainty, panic and injuries (nonfatal and fatal)}. The priorities are the numerical

ranks measured on a ratio scale (Saaty, 1980). The priorities are obtained from the

judgment of a column divided by the sum of all the judgment.

Physiological impact is considered more important than psychological impact for

causing injuries. Miller (2005) indicated that out of 131 victims of residential fires,

15 victims acted in irrational or attention seeking ways (although their intent to cause

fires and to cause harm was ambiguous or unclear). Nearly 11% of the victims were

due to psychological impact. Therefore, physiological impact is assigned the absolute

number 8 in the (2, 1) or second-row first-column (see Table 3-3). This implies that

physiological impact is eight times more risky than psychological impact. The

reciprocal value is shown in (1, 2) as 1/8, which signifies that psychological impact

is eight times lower in risk than physiological impact. The priorities are 0.111 for

psychological impact and 0.889 for physiological impact {calculated for example, 1

divided by (1+ 8) for first column, gives the priority for psychological impact a value

of 0.1111 (≅ 0.111)}.

Physiological Impact

Psychological Impact

Decision

Uncertainty Panic Injuries

(Fatal)

Evaluation

Impact

Risk

Use of Lifts during

Fire Emergencies

Injuries

(Nonfatal)

60

Table 3-3: Matrix (2 × 2) for Priorities of Lift Evacuation

Psychological Impact Physiological Impact Priorities

Psychological Impact 1 1/8 0.111

Physiological Impact 8 1 0.889

λmax = 2; C.I. = 0; C.R.= 0

Panic is considered more risky than decision uncertainty for psychological impact

and is assigned the absolute number 5 in the (2, 1) cell of Table 3-4. Likewise,

nonfatal injury is considered to be of extreme importance than other risks and

therefore assigned intensity of importance to a value of 9. Similarly, 2 is assigned to

nonfatal injury in comparison to panic in (3, 2). The risk priorities are 0.066 for

decision uncertainty, 0.319 for panic and 0.615 for nonfatal injury for psychological

impact. The calculations of risk priorities, λmax , C.I. and C.R. are given in Appendix

B.

Table 3-4: Matrix (3 × 3) for Priorities of Psychological Impact

Decision Uncertainty

Panic Injury (Nonfatal)

Priorities

Decision Uncertainty 1 1/5 1/9 0.066

Panic 5 1 ½ 0.319

Injury (Nonfatal) 9 2 1 0.615

λmax = 3.10; C.I. = 0.01; C.R.= 0.10

Nonfatal injury is considered more risky than panic for physiological impact and is

assigned the absolute number 3 in the (2, 1) cell of Table 3-5. Fatal injury is

considered to be of extreme importance than other risks and therefore is assigned an

intensity of importance value of 9. The risk priorities are 0.077 for panic, 0.231 for

nonfatal injury and 0.692 for fatal injury for physiological impact.

Table 3-5: Matrix (3 × 3) for Priorities of Physiological Impact

Panic Injury (Nonfatal)

Injury (Fatal) Priorities

Panic 1 1/3 1/9 0.077

Injury (Nonfatal) 3 1 1/3 0.231

Injury (Fatal) 9 3 1 0.692

λmax = 3; C.I. = 0; C.R.= 0

61

Synthesis of priorities: The priority vectors from Tables 3-3 to 3-5 are combined

into a single (or global) priority vector for evaluating the risks associated with the

use of lifts in fire emergencies (see Table 3-6).

Table 3-6: Global Risk Priorities

Risk Global Priorities

Decision Uncertainty = {(0.066 × 0.111) + (0 × 0.889)} 0.0073

Panic = {(0.319 × 0.111) + (0.077 × 0.889)} 0.1039

Injury (Nonfatal) = {(0.615 × 0.111) + (0.231 × 0.889)} 0.2736

Injury (Fatal) = {(0 × 0.111) + (0.692 × 0.889)} 0.6152

The risk priorities are 0.0073 for decision uncertainty, 0.1039 for panic, 0.2736 for

nonfatal injury and 0.6152 for fatal injury (see Table 3-6 and Figure 3-4). Fatal

injury has the maximum risk priority for causing risk in the lift system. The

combined risk priority for injury (nonfatal and fatal) is 0.8888 (≅ 0.889) and is used

later for evaluating the risk relating to physiological impact.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Priority

Decision

Uncertainty PanicNonfatal Fatal

Figure 3-4 – Global Risk Priorities

The matrices indicated in Tables 3-3 to 3-5 are consistent, and the consistency ratios

(C.R.) are 0 or less than 0.1 (see Table 3-7).

62

Table 3-7: Consistency Tests of Matrices

Matrix Consistency Ratio Consistency Test

Matrix for Main Objective 0 Accepted

Matrix for Psychological Impact 0.086 Accepted

Matrix for Psychological Impact 0 Accepted

Limitation: For pair wise comparisons in a matrix, the decision maker specifies a

judgment about ‘How much more important is one risk than the other?’ Each pair

wise comparison requires the decision maker to provide an answer to the question:

“how much more important is decision Uncertainty (or Panic) than Panic (or Injury),

relative to the overall objective?” Decision makers often find it difficult to accurately

determine cardinal importance weights for a set of risks simultaneously. As the

number of risks increases, better results are obtained when the problem is converted

to one of making a series of pair wise comparisons.

3.1.3 Acceptable Level of Risk

Risk analysis of lift evacuation system can provide a framework for decision-

making. More challenge is involved with the identification of a level of acceptable

risk, which is more philosophical than technical (Watts and Hall, 2002). Vrijling et

al. (1995) proposed acceptable levels of risk from both individual and societal points

of view. Acceptable risk RA is defined in terms of the policy factor expressing the

degree of voluntariness with which the activity is undertaken and with the benefit

perceived.

410−×= volAR β 3.8

where βvol is the degree of voluntariness with which the activity is undertaken

(varying from 0.01 to 100). The value of 100 is an indicator of voluntary high risk

exposure for its thrill activities like mountaineering and 0.01 for involuntary risk

exposure due to a hazardous structure, as with spatial planning issues (Vrijling et al.,

1995). This scale can be further refined (Vrijling et al., 1995).

63

The acceptable risk can provide a rational basis for single objective attribute, but this

does not reflect the criterion for multi-objective attributes. The present criteria for

risks are not yet fully developed to address the risks in a complex environment. A

universally acceptable level of fire risk does not exist (Watts and Hall, 2002).

Researchers are confronted with the problem of deciding what risks will adhere to

safety that affect life and death of the people taking advantage of its benefits. A value

judgment can be an arbitrarily human perception of risk, which may involve error.

For considering acceptable level of risk in the evacuation route, the risks, involved in

the fire-isolated exit of BCA 2005 Deemed-To-Satisfy (DTS) provision such as

stairs, are considered.

3.1.4 Hypothetical Building and DTS Provisions

Hypothetical Building: A hypothetical building shown in Figure 3-5 with 38 floors

is considered in this research analysis. The building containing 38 floors is selected

with the intent of providing a generic situation for high-rise apartment buildings,

above which, other provisions such as lift stacking or lift zoning are generally

applicable. The typical floor area of the building is 1000 m2 approximately. The

details of enclosures are given in Appendix J. The apartment building contains

twelve dwelling units (three bedrooms, two bedrooms and one bedroom), lifts, stairs

and public corridor. The lifts and stairs run from the ground floor to the top level.

The stairs and lifts are accessible through public corridor.

Figure 3-5 – Typical Floor of a Hypothetical Building (57 m × 20 m)

64

All the access, egress and fire safety features are assumed to be compliant with the

Australian building code (ABCB, 2005).

BCA 2005 DTS Provisions (ABCB, 2005): A fire-isolated stairway (including any

associated fire-isolated passageway) serving any storey above an effective height of

25 m must be provided with:

• an automatic air pressurisation system for fire-isolated exits in accordance

with AS 1668.1 (Table E2.2a); or

• open access balconies in accordance with (Clause D2.5):

- ventilation opening to the outside air; and

- not to be enclosed on its open sides above a height of 1 m except by an

open grille or the like having a fire air space of not less than 75% of its

area.

• If more than 2 access doorways open to a required fire isolated exit in the

same storey (Clause D1.7d),

- a smoke lobby must be provided; or

- exit must be pressurised in accordance with AS 1668.1.

The hypothetical building shown in Figure 3-5 is not provided with open access

balcony (or public corridor). There are more than 2 access doorways opening to the

public corridor in the same storey. Fire-isolated exits such as stairway and public

corridor need to be provided with pressurisation. As the positive pressurisation

would not allow smoke and hot gases to fire-isolated stairway, associated

passageway and lift shaft, positive pressurisation as a mechanical system is not

considered. Only a smoke lobby in the stairway is considered. The practice of

smoke lobby is also adopted by international codes and regulations (NBS, 2000 and

NFPA 101, 2000). The provisions of pressurisation and smoke lobby are shown in

Figure 3-6.

65

Figure 3-6 – Positive Pressurisation and Smoke Lobby in a Fire Isolated Exit

The risk involved in the fire escape stair with smoke lobby is considered as a

minimum acceptable requirement for a comparative analysis.

3.1.5 Selection of a Fire Scenario

The variables related to the building spatial environment, fire dynamics environment

and human activities are important for evaluating the use of lifts for emergency

evacuation. The variables are discussed in the subsequent chapters. However, a

deterministic approach is used for deciding a worst possible fire scenario. After

selection of a worst possible fire scenario, non-deterministic approach is considered

for detailed analysis.

Event tree analysis (ETA) is used for the evaluation of the worst possible fire

scenario. ETA is based on binary logic, in which an event either has or has not

happened. It is valuable in analyzing the consequences arising from an undesired

event. An event tree begins with an initiating fire (see Figure 3-7). The consequences

of the event are followed through a series of possible paths. Each path is assigned a

probability of occurrence and the probability of the various possible outcomes can be

calculated. A smoke detector either detects the smoke or it does not. A sprinkler

either works or it does not. Sole occupancy unit (SOU) door either closes correctly

or it does not. Shaft pressurisation either works or it does not. The probabilities of

success values of these events are given below:

FD

Legend: FD – Fire Door

SD – Smoke Door

Enclosed public corridor with

number of SOU access door Enclosed

Stair

Smoke

lobby

FD SD Enclosed public corridor with

number of SOU access door Enclosed

Stair

Blower to pressurise fire-

isolated exits

Provision of a smoke lobby

66

• Smoke alarms: The estimated reliability of smoke alarms in residential

buildings is 77.8% and the estimated reliability in apartment buildings is

69.3% (Bukowski et al., 1999). Smoke alarms (AS 3786) are therefore

assigned a probability of success value of 0.693.

• Sprinklers: Marryatt (1988) has quoted sprinklers as being 100% reliable

where inspection, maintenance and testing activities were well documented.

The study by Bukowski et al. (1999) found the reliability of sprinklers to be

96.6% for residential occupancy and 94.6% for overall occupancy. The

NFPA statistics (Rohr, 2001) for ten years reporting period from 1989 to

1998 indicates the operational reliability of automatic sprinkler systems for

apartment buildings is 87.6%. Hence, sprinkler systems are assigned a

probability of success value of 0.876 for apartment buildings.

• SOU fire door: The estimated reliability of passive protection is 95% for

construction with no openings and 90% for construction with openings (with

self-closers) under pre-flashover and flashover conditions (FCRC, 1996).

Residents within the fire affected unit are assumed to escape and leave their

SOU in one of two conditions i.e. entry door open or closed. These scenarios

simulate real life conditions where building occupants have lost their lives as

a result of doors being in the open position and occupants have decided to

flee at this time. Hence, SOU entry door is assumed to be either open or

closed. The door is assigned 0.9 in the closed and 0.1 in the open position.

• Shaft pressurization: The shaft pressurisation system is assigned a

probability of success value of 0.9 (Zhao, 1998) (see Chapter 2, Section 2.5).

67

Figure 3-7 – Event Tree Analysis for a Worst Possible Path (or Fire Scenario)

(Second worst possible path is analysed in this research program; remaining paths have less severity and therefore they are not analysed)

68

Probabilities of events are shown on the tree diagram. For each path (branch), a

probability of final outcome is calculated. Assuming the events are independent, the

probabilities along each path are multiplied. High severity of consequences is

involved in the evacuation routes, if sprinkler, fire door and pressurisation are not

working or none of the events are working (for example, failure of detection,

sprinkler, fire door and pressurisation). High severity of consequences may cause

psychological and/or physiological impact (for example, mental agony, anxiety,

panic, nonfatal injuries and/or fatal injuries). The probability of occurrence for high

severity of consequences is 0.0004 and is considered very low for failure of all fire

safety measures. Next higher probability of occurrence for high severity of

consequences is 0.0009 for failure of all fire safety measures except smoke alarm.

Other paths have comparatively lower severity of consequences and have high

probabilities of occurrence. However, it can not be interpreted that the severity of

consequences leading to mental agony, anxiety, panic, nonfatal injuries and/or fatal

injuries would not occur in other paths. It can also not be interpreted that the severity

of consequences leading to mental agony, anxiety, panic, nonfatal injuries and/or

fatal injuries would definitely be caused in the high severity of consequent paths.

Keeping in view the second worst possible path (or branch), where sprinkler, fire

door and pressurisation are not working and smoke alarm is working, the path is

analysed with variables related to the issues of human behavioural response and fire

hazard.

3.1.6 Concept Design Options

Risk analysis for lift system is conducted on a comparative basis with the minimum

requirement (stairs with smoke lobby). Three concept design options along with

additional evacuation strategy (lift evacuation for 25% of the population) are

considered for risk analysis (see Figure 3-8).

69

Figure 3-8 – Comparison of Stairs and Lifts

Table 3-8 gives the purpose of comparison and strategic planning between lift and

stair systems.

Table 3-8: Lift and Stair Systems for Comparison

S. No. Lift system Stair system Purpose

1. Unprotected lift lobby

(100% population evacuation)

Stairs

(100% population)

Feasibility consideration

2. Protected lift lobby

(100% population evacuation)

Stairs

(100% population)

Feasibility consideration

3. Protected lift lobby

(25% population evacuation)

Stairs

(75% population)

Feasibility consideration

and strategic planning

4. Double protected lift lobby

(100% population evacuation)

Stairs

(100% population)

Feasibility consideration

The details of the five concept design options are given below:

Concept Design ‘A’ (Lifts with Unprotected Lobby): Generally lifts are not

protected in high-rise apartment buildings. Therefore, concept design ‘A’ is

considered without any protective measure (see Figure 3-9 a).

Evacuation routes in

buildings

Stairs (Main evacuation

route)

Lifts (Alternate evacuation

route)

Lifts with unprotected lobby for

the entire population

Lifts with protected

lobby

Lifts with double protected lift lobby for

the entire population

Lifts for the entire

population

Lifts for 25% of the

population

Stairs with double protected smoke

lobby

Stairs for the entire

population

Stairs for 75% of the

population

Parts of one evacuation strategy

70

Concept Design ‘B’ (Lifts with Protected Lobby): To restrict heat exposure to the

lift cars, concept design ‘B’ is considered with a lift lobby. The lift lobby is enclosed

with the fire resistive walls, floors and self closing fire doors (see Figure 3-9 b).

Figure 3-9 – Three Concept Designs for risk analysis Concept Design ‘C’ (Lifts with Double Protected Lobby): The concept design

‘C’ is similar to the concept design ‘B’, except one additional door is provided in the

protected lift lobby. This provision is similar to the DTS provision for stair smoke

lobby (smoke lobby between two sets of doors). The lift lobby is enclosed with the

fire resistive walls and floors, one set of door is self closing fire doors and other can

be magnetically operated sliding door or self closing smoke door (see Figure 3-9 c).

Concept Design ‘D’ (Stairs): The concept design ‘D’ is an option for which the

associated risk is deemed acceptable. This option was included as a reference for the

comparative study.

Concept Design ‘E’ (Stairs and Lifts): The concept design ‘E’ is an option for the

use of protected lift system to evacuate 25% of the population and the remaining

population by stairs.

Unprotected lift lobby Double protected lift

lobby

Protected lift lobby

Smoke lobby

FD Public corridor

(a) (b) (c)

SD

Legend: FD – Fire Door

SD – Smoke Door

71

3.1.7 Risk Quantification

In order to determine the risks associated with the use of lifts, risks are quantified

and evaluated. The risks are quantified with the help of following models:

a. Stochastic models for evacuation using lifts and stairs; and

b. Fire hazard computational models for the determination of time for fire

hazards to occur; and

c. Probabilistic risk analysis for lift reliability issues.

Stochastic evacuation models estimate the building evacuation times. The model is

developed within the context of occupant load and building space, where the

evacuation time would tend to be constrained by human factors (social, physiological

and psychological characteristics) and limited by flow rate capacities of evacuation

routes. The fire hazard computational model determines the time to exceed the

tenability limits, or the available safe egress time (ASET) (ABCB, 2005b). The

results are compared with the safe evacuation time, or the required safe egress time

(RSET) (ABCB, 2005b), for determining the impact on building occupants and the

concept of ‘safety index’ is applied. The reliability of lift operational mechanism is

determined from standard risk assessment techniques. The results from building

evacuation models, fire hazard computational model and probabilistic risk analysis

form an integrated base for risk assessment.

3.1.8 Risk Assessment

The risk assessment is basically a structured approach to decision making under

uncertainty (Watts and Hall, 2002). To determine expected impact upon the building

evacuees, Multi-Objectives Decision Analysis (MODA) based on Analytical

Hierarchy Process (AHP) is used.

The parameters causing the risks are given weights (importance or merit). They are

selected from the literature review based on professional judgments and past

experience (for example, decision uncertainty may be caused by long waiting time

for lifts and the number of evacuees standing in queue). Weights are given by a

72

simple analytical method based on survey and statistics (for example, weights are

given to each component for causing deaths from fire, toxic gases or pre-existing

health conditions). After fixing the weights to the parameters, the value of each

parameter is obtained from the stochastic evacuation model, fire hazard model and

probabilistic risk analysis model for each concept design and strategy. These

parametric values may be different for each concept design based on risk involved.

The value of each parameter is multiplied by its weight and all the weighted values

are added to give a final risk value. The weighted values represent the risks

associated with the individual concept design on a comparative basis (see Chapter 7).

The risk analysis does not cover fire-cost expectation (FCE). The FCE includes the

capital cost for the passive and active fire protection measures, the maintenance cost

of active fire protection measures, and expected loss from the fire (Meacham, 2002).

The FCE quantifies the fire cost associated with the particular fire safety system

design.

3.1.9 Selection of Design Options

After determining the weighted risk values, the selection of concept design can be

made in light of the acceptable risk. The selection is rational, transparent and

systematic for decision making.

3.2 Research Work

Figure 3-10 outlines the research work for achieving the objectives. It presents the

proposed research activities based on the following chapters.

In Chapter 4, the residents’ choice for using evacuation routes is modelled using a

pilot survey. This determines the residents’ willingness and acceptance of the use of

lifts for emergency evacuation. Data from the pilot survey are used for risk

assessment. Interviews were conducted with the fire brigade personnel. An

illustrative case of 38 storey hypothetical apartment building is prepared for risk

analysis. A stochastic model for lift evacuation is developed for determining the lift

time periods and the number of evacuees in queue. These output variables are

73

determined for the entire population and one-fourth of the population (25%) for lift

evacuation. The output variables for one-fourth of the population are determined for

evaluating a safe and efficient lift evacuation strategy. The proportion of 25%

population is based on the literature review and survey findings (see Chapter 4,

Section 4.2.3). Stochastic model for stair evacuation is also developed for

determining the stair time periods and the number of evacuees in queue. These

output variables are determined for the entire population and three-fourth of the

population (75%) in the stairs.

Figure 3-10 – Research Work Flow Diagram

Risk identification

for the use of lifts

Issues of fire hazard Issues of lift operational

mechanism

Issues of human

behavioural response

Design considerations

and strategy evaluation

Stochastic evacuation models for lifts and stairs

Fire hazard modelling for lifts

and stairs

Risk assessment

(An integrated approach)

Pilot survey and

interviews

Probabilistic risk

Analysis

Establish available safe evacuation time and parametric analysis

Establish required safe evacuation time for lifts

and stairs

Safety indices for

the lifts and stairs

Parametric analysis for ‘Decision Uncertainty’

Parametric analysis for

‘Panic’

Parametric analysis for

‘Injuries’

74

In Chapter 5, models for fire and smoke hazards in evacuation routes are proposed

for determining the time to exceed the tenability limits relating to visibility, CO,

CO2, O2 (low oxygen-causing hypoxia), temperature and radiant heat flux. Twenty

four fire scenarios are analysed after incorporating uncertainties relating to wind and

vertical location with FDS model (McGrattan et al., 2004). The safety indices are

evaluated for the lift and stair systems. The safety index estimates the probability that

the escape time would exceed the available time. The fire hazard model does not

take into account self closing device of SOU door or operation of sprinkler.

In Chapter 6, the reliability of lift operational mechanism for water damage, lift

malfunctioning and electric power failure are considered for probabilistic risk

analysis. The analysis is based on statistics and standard risk assessment techniques.

The techniques include a complex parallel and series system (a modified form of

event tree analysis) and fault tree analysis. Output variables from FDS models are

used for determining the lift malfunctioning.

In Chapter 7, the Multi-Objectives Decision Analysis method is used for risk

assessment. The parameters are assessed and statistics from various sources are used

for an integrated risk assessment approach. Each parameter is given a degree of

importance (weight). For decision uncertainty, survey data and statistics reports are

analysed for giving weights. The output variables are obtained from the stochastic

evacuation models. For panic, statistics relating to unavailability of evacuation routes

and visual threat (during pre-life threatening condition) are considered for giving

weights. For life threatening conditions, statistics relating to fire deaths are analysed

for giving weights. The risk priorities were derived from analytic hierarchy process.

After giving parametric weights and values, risk assessment is conducted for design

options and evacuation strategies. Sensitivity analysis is conducted for adjudging the

importance of variables. In Chapter 8, the feasibility and design considerations for

lift evacuation system are determined. The redundancy measures are proposed.

75

3.3 Conclusion

The research methodology involves the following main steps:

(a) risk identification, risk priorities and expected impact upon the building

evacuees.

(b) establishing acceptable level of risk for design options and evacuation

strategies.

(c) quantifying risks with the stochastic models for building evacuation, fire

hazard models for determining time to exceed tenability limit and

probabilistic risk analysis for lift reliability issues.

(d) conducting risk assessment using Multi-Objectives Decision Analysis

method and determining suitable design options and evacuation strategies.

Risk cause-effect relationship identifies the key issues to be addressed. The issues of

human behavioural response, fire hazards and lift operational mechanism give rise to

three hierarchical levels of risks i.e. decision uncertainty, panic and injuries (nonfatal

and/ or fatal). These risks are interlinked, multi-dimensional and require a complex

process for risk assessment. The research strategy involves risk management by

reducing the risk level to an acceptable level using various concept design options

and evacuation strategies. The risks are quantified by using building evacuation

simulation models, fire hazard models and probabilistic risk models. The Multi-

Objectives Decision Analysis method is proposed for risk assessment. The priorities

of the three risks are assigned on a ratio scale. Risk assessment is conducted based

on parametric values obtained from the models. The feasibility of alternative design

model is determined in light of the acceptable risk.

This chapter has addressed risk identification, expected impact upon the building

evacuees, concept design options and stairs as an acceptable level of risk. The next

chapter is related to stochastic building evacuation model.

76

4. STOCHASTIC MODELS OF BUILDING EVACUATION

4.1 Introduction

Required safe evacuation time (RSET) is an important parameter in fire safety

engineering (ABCB, 2005b). RSET is defined as the time period, subsequent to fire

alarm, required for safe occupant evacuation. RSET depends on several factors, such

as the physical dimensions of evacuation path (length and width of corridors, stairs

and lifts), fire detection and alarm, the occupant density, number and distribution,

and the occupants’ social, physiological and psychological characteristics. The

physical and human factors play a vital role in the evacuation procedure and are

often not given due consideration. The building evacuation is generally affected by

uncontrollable and random arrival pattern of occupants and causes the output to be

random as well. A considerable period of time is lost in pre-evacuation activities.

Earlier researches have been conducted to predict the building evacuation time

(Sekizawa et al., 1996 and Kuligowski, 2003). The approaches of these researches

were deterministic and did not include physical and human factors. However, the

results have demonstrated that using both stairs and lifts could provide a better

performance (Sekizawa et al., 1996 and Kuligowski, 2003). The approach used in

determining the evacuation times should include human factors for realistic

scenarios.

This chapter is focused on developing stochastic models for determining RSET using

lift and stair systems. Stochastic models reflect uncertainty due to variability or

randomness in data. The discrete event simulation using ARENA (Rockwell, 2000)

gives the capability to develop realistic time based models for evacuation using lifts

and stairs. The model simulates the lift cabin operation and computes output

variables relating to RSET (such as lift waiting time, lift transportation time and lift

evacuation time) and the number of evacuees waiting in queue. Such performance

parameters are also developed for the stair system for comparison with the lift

system. The output variables are determined for the entire building population and

for one-fourth of the building population. The output variables for one-fourth of the

building population are determined with the intent of reducing RSET so that

77

evacuees are not subjected to a long waiting period and exposed to toxic hazards of

fire effluents (RSET is reduced with the evacuation of one-fourth of the building

population). RSET is compared with the available safe evacuation time (within safe

conditions) in Chapter 5. The use of lifts for a partial population can help in

developing operational strategies.

Before evaluating the output variables relating to RSET, occupants’ attitude towards

the use of lifts need to be considered as the occupants should have inherent sense of

confidence in the mechanical evacuation facility. In Hong Kong, a survey was

conducted for high-rise apartments with the objectives of determining comforts of

high-rise living and any effect it might have on children (Mori and UHK, 2002). In

response to a question relating to the disadvantages of high-rise living, 36% reported

fire escape, 20% reported lift breakdown, 2% reported strong wind, 2% reported heat

and 4% reported lack of play areas. This indicates that a significantly large

proportion of high-rise residents are concerned of fire emergency evacuation

followed by lift breakdown. The confidence on lift systems requires further analysis.

A pilot survey was therefore conducted to gain an understanding of the occupants’

attitude towards the use of lifts. Interviews were also conducted to determine the

attitude of professional fire fighters toward the use of mechanical lift evacuation

system.

4.2 Pilot Survey

4.2.1 Pilot Survey Overview

A pilot survey was conducted for a 38-storey apartment building in Brisbane. The

pilot survey was conducted using a questionnaire form, which was distributed to 250

residents in the building, out of which 20% responded. The questionnaire form and

data collected are given in Appendix C. The objectives of the survey were to:

1. determine the residents’ preferred access and egress routes in normal

circumstances; and

2. determine the residents’ preferred exit route during fire emergencies; and

78

3. determine the residents’ awareness towards the ‘Emergency Evacuation

Procedure’ (EEP) ; and

4. determine the residents’ priority and willingness to accept lift as an

alternative evacuation facility.

The residents were adjudged for the following parameters:

• age distribution in the high-rise apartment building

• residents’ inclination toward the use of lifts for emergency evacuation

• residents’ awareness for emergency evacuation procedure

• residents’ inability to use stairs and hence they need to use lifts for prompt

evacuation

The parameters were used in the analysis to determine the necessity of lifts as an

evacuation facility in high-rise apartment buildings.

4.2.2 Pilot Survey Results

Age Distribution: Residents living in high-rise apartment building were analysed for

their age distribution. This parameter is helpful in ascertaining the necessity of lifts

as an evacuation facility for aged and disabled persons. The results presented in

Table 4-1 show the aged distribution of residents. The results show that the majority

of the baby boomers (aged from 65 to 73 years) were living above the 20th floor

level.

Table 4-1: Residents’ Age Distribution

Under 30

years

30 to 44

years old

45 to 64

years old

65 years

and older

1 to 10 storeys 18% 18% 64% -

11 to 20 storeys 13% 54% 26% 7%

21 to 28 storeys 24% 46% 15% 15%

29 to 38 storeys 27% 18% 46% 9%

Overall average 20.5% 34% 37.75% 7.75%

79

Normal Entry and Exit: Residents were surveyed for their use of entry and exit

routes in the building. This helped in determining residents’ inclination toward the

use of evacuation routes. The survey results presented in Table 4-2 show lifts as

normal access and egress routes. The pilot survey results showed that 92.75% of the

residents always use lifts, 3% of the residents use lifts sometimes, and 4.25% of the

residents use them rarely on average, for normal building access and egress. These

values were obtained from the average of grading given to individual data collected

for the use of lifts and stairs during normal entry and exit. From these results, it is

clear that residents were inclined towards the use of lifts for everyday access and

egress. Some of the residents have the tendency of using stairs.

Table 4-2: Residents’ Use of Lifts as Normal Access and Egress Routes

Storey Always Sometimes Rarely Never

1 to 10 storeys 92% 4% 4% -

11 to 20 storeys 85% 4% 11% -

21 to 28 storeys 96% 2% 2% -

29 to 38 storeys 98% 2% - -

Overall average 92.75% 3% 4.25% -

Emergency Exit: Residents were surveyed for their intended use of emergency exit

routes. Residents were asked about the use of lifts or stairs during fire emergencies.

The survey results presented in Table 4-3 show that 84% of the residents, living on

the 1st to 10th storeys, 93% living on the 11th to 20th storeys, 92% living on the 21st to

28th storeys and 92% living on the 29th to 38th storeys, intend to use stairs during fire

emergencies. This means that on average 90.25% of the residents intend to use

stairs, whereas 3.75% of the residents rely on fire brigade facilities and 6% of the

residents intend to use lifts and stairs. The option of using both lifts and stairs was

also given to the residents, which they may like to use depending upon the

circumstances. However, no one intends to use lifts solely or stay in their unit. From

these results, it is clear that majority of the residents intended to use stairs in the case

of fires. However, 9.75% of the residents (mainly include aged or disabled) were not

capable of evacuating the building using stairs and they have to rely on using a

combination lifts and stairs or fire brigade facility. Permanently disabled residents

80

were on the 11th and 21st floors, while the temporarily disabled residents were on the

3rd and 18th floors.

Table 4-3: Residents’ Preferred Exit Routes during Fire Emergencies

Storey By stairs By lifts By lifts and

stairs

By fire brigade

appliance

Remain

in unit

1 to 10 storeys 84% - 8% 8% -

11 to 20 storeys 93% - - 7% -

21 to 28 storeys 92% - 8% - -

29 to 38 storeys 92% - 8% - -

Overall average 90.25% - 6% 3.75 -

Awareness of Emergency Evacuation Procedure: Residents were surveyed for

their awareness about the emergency evacuation procedure (EEP). The awareness

level helps the residents in taking rational decisions during fire emergencies. The

survey results (see Figure 4-1) show that 69% of the residents admitted that they

were not trained in EEP on average. Only 42% of those living on the upper top

levels between 29th and 38

th storeys were trained. Majority of residents living in the

middle levels between 21st and 28

th storeys stated that they were not trained in EEP.

Untrained, 83%

Untrained, 57%

Untrained, 77%

Untrained, 58%

Untrained,

69.00%

Trained , 17%

Trained , 43%

Trained , 23%

Trained , 42%

Trained , 31%

0% 20% 40% 60% 80% 100%

1 to 10 storey

11 to 20 storey

21 to 28 storey

29 to 38 storey

Overall average

% of Residents

Untrained Trained

Figure 4-1 – Residents’ Awareness of Emergency Evacuation Procedure

81

Experience with Fire Drill: Residents were asked if they participated in fire drills in

the building. Experience with fire drill determines if the residents were trained in

evacuating the building confidently. The survey results showed that 69% of the

residents revealed that they did not experience any fire drill in the building (see

Figure 4-2). Ten percent of the residents felt that they faced difficulties during the

fire drills and expressed a sense of crowdedness, bottleneck effect in movement,

narrow width of stairwell, different walking speed, knee trouble and wheel chair

movement problem. Twenty one percent of the residents were confident and

experienced no difficulty during fire drills. The survey results also indicated that 8%

of the residents observed that stair doors were locked at all times or some times.

No difficulty,

21%

No fire drill

experience,

69%

Difficulty

encountered,

10%

Figure 4-2 – Residents’ Experience during Fire Drill

4.2.3 Discussion and Conclusion

The confidence intervals were determined for the use of lifts during normal

circumstances and intended use of stairs during fire emergencies. The confidence

intervals were calculated with the help of BETAINV (see Appendix C). A 95%

confidence interval for the use of lifts during normal circumstances is determined to

be in the range of 0.849 and 0.961, whereas 95% confidence interval for the stairs as

an emergency exit is determined to be in the range of 0.8069 and 0.9526. The use of

lifts during normal circumstances is slightly higher than the use of stairs during fire

emergencies, which indicates that residents were more inclined to use lifts as a mode

of vertical transportation during normal circumstances than the use of stairs during

emergencies. The occupants can be more confident in using lifts as an alternative

evacuation facility during fire emergencies.

82

The small percentage (9.75%) of residents intending to use lifts during an emergency

indicates that the burden of managing the lift evacuation during an emergency is

small, since most of the residents will be using stairs for evacuation. However, this

small percentage does not represent the willingness of the entire population for the

use of safe and reliable lifts. If lifts are permitted as an alternative evacuation facility,

the number of evacuees using lifts may increase as others may also join during

emergency evacuation. Therefore, if the existing lift infrastructure does not provide

adequate safety to the majority of population in the building, only aged and disabled

can be considered for lift evacuation (with the limited number of occupants, output

variables relating to RSET can be reduced to achieve adequate safety). Aged and

disabled revealed their concern of life safety while living on upper levels. Some of

them commented that they had to wait in the building for fire brigade help. The total

population of aged and disabled was approximately 16% based on this survey and the

result agrees with the data published in the literature (ABS, 2004 and Pauls, 1977).

However, the population of aged persons was only 7.75%, whereas the literature

review indicated 13%. The population of disabled persons was 8%, whereas the

literature review indicated only 3% are living in high-rise apartment buildings in

Australia. If aged and disabled representing 16% of the population are using lifts and

it is assumed that half of them are assisted or helped by others (say 9%) in using the

lift evacuation route, about 25% of the population may require an alternative

evacuation facility in the building.

The survey findings also showed that the majority of residents were not trained in the

EEP. A significant population of residents reported that they never experienced or

witnessed fire drill in buildings. In such circumstances, residents may not be

confident in building evacuation. This may add complexity to their evacuation

procedure and a state of confusion may arise. It is concluded that the residents may

need to use lifts for emergency evacuation in high-rise apartment buildings. Further,

there are also chances that a significant population may also join 25% of the

population for lift evacuation.

Above discussion shows that (a) the majority of residents are inclined to use lifts in

their routine life (b) aged and disabled are concerned about their emergency

evacuation as the majority of them are residing on mid or upper top levels (c) the

83

majority of residents are not trained in EEP (d) necessity to investigate an alternative

evacuation facility for the entire population. If it is not feasible, the use of lifts can be

investigated for at least 25% of the building population. Regular drills and practices

are required to be conducted to avoid ambiguity in evacuation procedures as there are

chances that a few occupants from 75% of the population may also join 25% of the

building population for lift evacuation.

The residents near the ground floor of the building did not show their intent for using

stairs as their prevalent ingress or egress route. Thus, efforts were also made to

conduct surveys in other buildings of similar nature and height, but building

corporate management did not give permission. Further, there were a few other

limitations of this pilot survey, which are given below:

- The sample size was limited. The results generated from a sample of 50

residents in one apartment building did not necessarily present adequate

information for the general population in apartment buildings.

- The data on the choice of emergency exit route reflected the building

occupants’ intentions expressed when filling in the questionnaire, which may

not be the true indication of what they will do in a real emergency.

4.3 Interviews

Some proponents indicated that the lifts in protected lift lobby can be used as an

evacuation facility (Sekizawa, et al., 1996) and that protected lift lobby space can be

used as staging area for fire department (Kuligowski and Bukowski, 2004).

Opponents argued that the lift lobby protection would not provide additional safety.

One of the papers presented by the regular fire brigade officers strongly disagreed

with the intent of proposing lift as an evacuation facility in the buildings

(O’Donoghue and O’Donnel, 2003). Although, the use of lifts has always been a

controversial issue, the occupants have used lifts as an escape route on many

occasions (Barnett et al., 1992, see also Appendix A). Therefore, interviews were

conducted with fire brigade personnel to determine their willingness toward the use

of lifts as an evacuation facility. Three interviews were conducted with the officers

of regular fire services in Australia.

84

The objectives of the interviews were to:

1. determine the interviewees’ perception toward the use of lifts in high-rise

buildings; and

2. know the fate of general public, disabled and aged persons during fire

emergencies.

The questions asked from the emergency officials and their replies are given in

Appendix D. The summary of the interview findings are given below:

1. All interviewees expressed their concern about the impeded movement of

occupants in stairs as a result of over-crowding, bottle neck, counter current

flow between fire brigade personnel and evacuees.

2. Interviewees stated that disabled, aged, children and physically weak persons

can go to refuge areas (such as lift lobbies) for temporary staging; then use

lifts and/or stairs with assistance as needed. Interviewees also recommended

that lifts can be explored as an option in high-rise buildings.

3. Interviewees mentioned that experienced people act rationally and occupants

use stairs for emergency evacuation. One of the interviewees stated that

about half of the population behave rationally depending upon the

circumstances.

4. All interviewees expressed that lifts can be considered as an option for

disabled and elderly people. However, lift systems should satisfy design,

commissioning and maintenance requirements, which can prove to support

safety of the evacuees.

5. Interviewees informed that fire emergency evacuation drills are not regularly

held in apartment buildings as fire drills are not mandatory as per the

provisions of Building Fire Safety Regulation.

6. Comfortable travelling via stairs depends upon the safety measures in the

stairs and individual’s strength. However, there are re-entries from stairs to

building floor levels at every 4th level in high-rise buildings.

7. In a working fire environment, fire fighters use lifts or stairs depending upon

the fire conditions in the building.

8. Electric power supply is generally reliable in Australia, although power

outage can not be ruled out.

85

9. Interviewees recommended that provisions of pressurisation and lift stack

arrangement can be considered for lift evacuation system.

It should be pointed out that the information obtained from the interviews is

qualitative. The accuracy of the figures given by the interviewees was not verified.

4.4 Analysis of Building Evacuation Periods

To determine the risks relating to human behavioural response and fire hazards, the

variables relating to building evacuation periods (or RSET) are analysed. The

objectives of the analysis are to:

1. develop a model for the building lift evacuation under uncertainties

associated with human social, behavioural and physical movement along with

a priori heuristics of the lift domain and determine the probable time for safe

evacuation; and

2. develop a model for the building stair evacuation under uncertainties and

determine the probable time for safe evacuation; and

3. compare lift evacuation and stair evacuation times for ascertaining human

behavioural response (decision uncertainty); and

4. examine the potential ways of reducing the building evacuation time periods.

4.4.1 Methodology

Two stochastic models of lift and stair evacuation systems are developed to

determine the probability of successful evacuation. The model analyses the passenger

optimum service level for lifts in comparison to stairs. The following variables are

determined:

• lift waiting time, lift transportation time, lift evacuation time and number

of evacuees in queue

• stair waiting (queuing) time, stair travelling time, stair evacuation time

and number of evacuees in queue

86

Lift evacuation time: Lift evacuation time is the total time period for evacuating the

building occupants using lifts. The lift evacuation time tLE for the number of

evacuees in a building can be expressed as:

xiLE tttt = t ∪∪∪∪∪ ..........21 4.1

where x is the number of evacuees

(Union symbol ∪ shows that lift evacuation time includes the evacuation time of all

individual evacuees)

The lift evacuation time of an individual ti can be expressed as:

LILTLWLMLPMFDi tttttt = t +++++ 4.2

where

tFD is the fire detection time (second)

tLPM is the lift pre-movement (coping and response) time for an evacuee (second)

tLM is the movement time for an evacuee (second)

tLW is the lift waiting time for an evacuee (second)

tLT is the lift transportation time for an evacuee (second)

tLI is the evacuee intermittent floor movement time via lift (second)

A combination of lift pre-movement time, movement time and lift waiting time is the

lift pre-evacuation time. The lift pre-evacuation time can be expressed as:

LWLMLPMLPE tttt ++= 4.3

Lift evacuation time of an individual can be considered as the lift evacuation time

from the building, if that individual takes the maximum time to evacuate the

building. It can be expressed as:

87

tt LE imax= 4.4

Stair evacuation time: Stair evacuation time is the total time period for evacuating

the building occupants using stairs. The stair evacuation time tSE for the number of

evacuees x in a building can be expressed as:

xiSE tttt = t ∪∪∪∪∪ .........21 4.5

while the evacuation time of an individual ti can be expressed as:

SISTSMSPMFDi ttttt = t ++++ 4.6

where

tSPM is the stair pre-movement (coping and response) time for an evacuee (second)

tSM is the movement time for an evacuee (second)

tST is the stair travelling time for an evacuee (second)

tSI is the evacuee intermittent floor movement time through stairs (second)

Further, stair evacuation time of an individual can be considered as the stair

evacuation time from the building, if that individual takes the maximum time to

evacuate the building. It can be expressed as:

tt SE imax= 4.7

Pre-evacuation activities are considered the same as the evacuees are likely to choose

lift or stair at a later stage. Therefore, stair pre-movement time and lift pre-movement

time are the same for both the evacuation routes. Similarly, stair movement time and

lift movement time are considered the same as both evacuation routes are considered

equidistance from the dwelling units.

The methodology used in this analysis is given in Figure 4-3. The lift simulation

model is based on the conventional lift group controller (see Section 4.5.2), which is

88

the most prevalent lift group controller system used in apartment buildings. The stair

simulation model is based on the most prevalent stair design in apartment buildings.

The lift simulation model generates the results of performance parameters (RSET

and queue length) during down peak traffic. The results from simulation are analyzed

in comparison with that from the stair simulation model. Output variables are plotted

with the help of a computer program @RISK (Palisade Corp, 1996) for obtaining the

mean and standard deviation (see Chapter 2, Section 2.10), which are used in

Chapter 5 in determining the fire hazards.

Figure 4-3 – Flow Diagram for Analysing the Output Variables of Models

If the results generated by the simulation models are not acceptable, the parameters

(lift performance parameters or lift logic controller) can be developed/ modified, for

which the advanced technologies such as genetic algorithm, reinforcement learning,

Lift supervisory

controller and variables

Lift simulation model

Determine variables

relating to stair evacuation

Hypothetical building

model and variables

No

Yes

Determine variables

relating to lift evacuation

Comparing lift and

stair evacuations

Acceptable level

Establish parameter for

simulation model

Stair simulation model

Develop performance parameters for future

research

Analysis of output

variables

Verification of

proposed models

Lifts for a limited number

of evacuees

Modify lift simulation

model

89

fuzzy logic or neural networks lift group controller need to be considered for

apartment buildings. These technologies are not currently used in apartment

buildings; however, they are used in office buildings for efficient transportation.

Researches have shown that these technologies have significant time saving of up to

25% to 40% in transportation (Siikonen, 1997 and Cortes et al., 2004). The

technologies include advanced adaptive systems with short-term and long-term

memories and ongoing calculation for how much time has elapsed between initial

call and arrival of the car in order to prioritize call. If advanced technologies are used

in the lift evacuation system, the performance parameters need to be modified for the

evacuation model. This research begins with the lift time periods and the number of

evacuees in queue determined within the acceptable level for a fraction of population

(25% of the population, which may include aged and disabled). The partial

evacuation of 25% of the population is estimated from the literature review (ABS,

2004 and Pauls, 1977) and pilot survey (see Section 4.2.3).

The simulation models can never be validated over the whole range of their

behaviour (Phillips, 1995). Hence, the lift evacuation model was verified using the

results from a building egress model, ELVAC (Klote et al., 1991) while the stair

evacuation model was verified with a mathematical expression. ELVAC is a

deterministic model and is not capable of incorporating random variables and

distributions. Hence the lift and stair evacuation models were verified without

incorporating random variables. The verification is limited to the output parameters

that are common to stochastic and deterministic models. The common parameters of

both the models are identified in Section 4.9.

4.4.2 Discrete Event Simulation

The building egress models such as ELVAC (Klote et al., 1991) and EVACNET

(Kisko et al., 1998) are based on deterministic approach and assume exact

knowledge of input parameters. This causes a discrepancy between the evaluated

results and real situations. An alternative approach is to model the uncertainty by

introducing random variables, in which many numerical observations are made as a

probability distribution for obtaining a condensed approach of final results. The

90

condensed approach is the average value of the results with minimum, maximum and

half width (or 95% confidence level) of multiple replications.

The non-deterministic stochastic modelling reflects inherent variability found in

physical system parameters that demonstrate random behaviour (Hoeksema and

Kitandis, 1985). This modelling approach is intended to incorporate the uncertainties

associated with physical system in predicting the system behaviour. In a

probabilistic risk assessment, He et al. (2003) adopted a simple approach where the

building occupant evacuation was treated as a Poisson process and no differentiation

was made between pre-movement and movement activities. Vistnes et al., (2005)

later employed a stochastic approach to estimate the time associated with the pre-

movement activities.

The modelling is intended to represent all the uncertainties associated with physical

system. Lift evacuation time and stair evacuation time (or RSET) are quantified by

defining random variables relating to pre-evacuation activities, human social and

physical movement. ARENA (Rockwell, 2000) is a commercially available package

for stochastic modelling. ARENA is the animation component of the SIMAN

language, which is a powerful general purpose simulation language for modeling

discrete, continuous and combined systems. SIMAN has a logical modeling

framework in which the simulation problem is divided in to model and experiment

components (see Figure 4-4). The model describes the physical and logical elements

of the system. The experiment specifies the experimental conditions (inputs) under

which simulation runs including the initial condition, attributes, running time and

replications. Once a model and an experiment have been defined, they are linked and

executed by SIMAN language to generate the simulated response of the system. The

SIMAN output analyzer is used to generate plots, tables and bar charts. ARENA has

been used successfully in many major projects including railway stations, airport

terminals, fire departments, manufacturing and processing industries and high rise

buildings (for evacuation purpose). This discrete model is used in the analysis of

output variables relating to the issues of human response and fire hazards where it

handles the behaviour of thousands of occupants, hundreds of queues and the

dynamics of tens of lifts.

91

Figure 4-4 – SIMAN Flow Diagram

4.5 Model Framework

In order to determine the maximum efficiency of lift and stair evacuation systems,

ARENA version 10.0 software was used to simulate the possible events. Passenger

traffic in the building can be described as down peak traffic during fire emergencies.

The occupants arrive at the lift lobby or stair lobby on upper floor levels and travel

down in lifts or stairs to the ground floor. This section presents the details of a

hypothetical building model, and experimental (variables) components used in

simulating the lift and stair evacuation systems.

4.5.1 Hypothetical Building and Parameters

A hypothetical building shown in Figure 4-5 with 38 floors was considered in the

analysis. The building was analyzed with the following assumptions:

• A symmetrical fire occurrence on three levels as different scenarios viz.

lower level (2nd

floor), middle level (19th floor) and top most level (38

th

floor) in the building. Fires on three floors are considered independently

Model components

Complier

Model object file Experiment object file

Complier

Experiment components

Linker

SIMAN program file

Execute

Results

92

to determine the effects on evacuation times. Evacuation times will be

used in Chapter 5.

• Emergency evacuation routes on the ground floor are suitably

compartmented to restrict the spread of smoke and hot gases. It complied

with the deemed to satisfy provisions of the building code.

• Pre-movement activities commence following a fire alarm.

• Evacuating population is evenly distributed at both the stairs.

• The service levels for both stairs and lifts are considered to be their

maximum capacity.

Figure 4-5 – Typical Floor of a Hypothetical Building (57m × 20m)

The test problems based on the generic apartment building, lift configurations and

traffic scenarios have been generated. Table 4-4 shows the details of building,

occupants, lifts and stairs. The generic design of the building and variables were

prepared with the intent of giving equal evacuation opportunities for lifts and stairs.

The main terminal (ground floor) is indexed 1. Occupants’ characteristics, such as

walking speed and behavioural response, are given in Sections 4.5.3 and 4.5.4.

93

Table 4-4: Model Parameters

Parameter Building data

Total floors 38

Floor population (each level) 32

Height between floors (m) 3

Number of lifts 4

Max lift capacity (persons) 16

Max velocity (m/s) 3.15

Acceleration (m/s2) 1.0

Number of stairs 2

Width of stairs (m) 1.2

4.5.2 Lift Supervisory Controller for Lift Simulation Model

The orders in which the lift cars respond to the landing calls play a vital role in the

performance of a vertical transportation system. Normally apartment buildings are

provided with a simplistic supervisory lift group controller. The proposed simulation

model is based on a conventional lift group control system. The flow diagram for lift

operation used in the model is shown in Figure 4-6.

The controller implements dispatch rules that make use of an IF-ELSE logical

command set (Cortes et al., 2004). The dispatch rules are simulated in the computer-

aided design suite LSD (Lift Simulation and Design), under the designation of the

THV algorithms (implemented under this designation at the University of

Manchester Institute of Science and Technology). This system collects the

information in most common rules in duplex or triplex algorithms. The duplex or

triplex algorithm assigns the hall call to the nearest lift in the travelling direction.

This method is bound to expert knowledge and a priori heuristics of the lift

application domain. Whenever an evacuee presses the hall call button for a lift,

micro-processor type lift group controller mounted in the lift machine room receives

the request and logs it for future reference and work on the principle of moving in

one direction at a time. Lift heading to higher floors is programmed to ignore calls

for lower floors until they have reached the top. Each hall call is attended by only

one lift car. The lift can stop at a floor only if there is a hall call from or a cabin call

94

to that floor. The cabin calls are served in accordance with the cabin’s current

travelling direction.

Figure 4-6 – Lift Controller Logic Diagram used for ARENA Simulation Model

There are two sets of doors, which work in tandem in order to safely allow evacuees

to enter and exit the cars. One set of doors remains shut until a lift car's presence is

detected. The other door is controlled by the lift's controller and opens sideways as a

powerful electric motor pulls the first panel. Once both doors are open, evacuees

95

leave the lift to allow new evacuees to board and more calls to be answered. The lift

car door possesses a sensor that detects if someone is obstructing the door and door

closing mechanism stops the closing of doors and remains in the open position until

the obstruction is removed. It is noted that the lift doors are open for 5 seconds. A

25% of additional delay is conservatively taken during emergencies as evacuees may

hold the lifts. Acceleration and deceleration functions as indicated in Table 4-4 are

added. If the lift car exceeds its capacity (weight), lift sends a signal to the lift group

controller not to pick up any more evacuees until the evacuees disembark the car.

4.5.3 Lift Simulation Variables

In order to consider a realistic situation, certain variables such as occupant arrival

rate, occupants’ characteristics (social), response and movement (physical – based on

age factor) are incorporated in the model. The times between occupants’ arrival to

the common evacuation route (such as lifts and stairs) are determined from fire

detection time, pre-movement (response and coping) time and movement time.

Occupant movement: Social groups within the building stay together throughout

the duration of the evacuation. Someone would like to ascend the lift to look for his

relatives or to travel inter-floor in the building. Building body corporate or

emergency crew would also like to ascend upper floor. For the current model, the

down-movement, inter-floor movement and up-movement are specified below:

� From the upper floors – all 1184 (32 × 37) occupants would be using lifts to

go to ground or intermittent floors

� Intermittent floors – 12 occupants would be using lifts for intermittent floor

movement (approximately 1% of total population – 1184 occupants)

� From the ground floor to the upper floors – 22 outsiders would like to go to

upper levels (approximately 2% of total population)

– Total number of occupants using lifts = 1206

Fire detection time (tFD): FDS model is used to determine the fire detection time,

which is 90 seconds for fire compartment (smoke alarm in SOU – individual alarm

96

only and not connected to public alarm) and 140 seconds for corridor (smoke

detector in public corridor and connected to public alarm) (see Chapter 5, Section

5.6).

Pre-movement (response and coping) time (tLPM): The pre-movement time is the

time spent by the occupant before starting the evacuation. It includes the time

required to perceive the event and the time required to respond to the event. The pre-

movement times can vary from seconds to many minutes (occupants are awake,

trained, familiar with the building, alarm systems and procedures). This is arguably

one of the poorest documented attributes in fire safety engineering. The existing

methods used to determine occupants’ response and coping are often inadequate and

insufficient to predict patterns of occupants’ spatial movement. Brennan and Thomas

(2001) argue that there is a difficulty with the traditional assumption that occupants

confronted with fire will react to fire. They proposed a paradigm shift from a reactive

to an interactive model of behaviour in fires. Occupants are likely to be intimately

involved in interactions with fire. However, occupants may have minimal interaction

with fire due to the speed it developed and these occupants are usually asleep or

otherwise unaware until escape is impossible.

Based on interactive model of behaviour of fires, the pre-movement (response and

coping) time of occupants in apartment buildings is selected from Fire Engineering

Guidelines (FCRC, 1996). The method considers the occupants’ physical, mental and

social conditions during the fire response and coping period and evaluates a

deterministic time period. The occupant response time based on primary weighting

factors such as alertness, commitment and familiarity and secondary weighting

factors such as role, focal point, mobility, social affiliation and position is determined

(see Appendix E). The coping time based on primary weighting factors such as role,

mobility and social affiliation and secondary weighting factors such as alertness,

position, commitment, focal point and familiarity is also determined. The calculated

response time is 186 seconds while the coping time is 255 seconds (see Appendix E).

Movement time (tLM): The horizontal walking speed of occupants is based on the

density. If the density is smaller than 0.54 persons/m2, the walking is free to a value

97

of 1.1 m/s. If the density is greater than 3.8 persons/m2, the occupants can not move

(Nelson and MacLennan, 1995). An average horizontal walking speed of 0.6 m/s for

the aged and disabled is conservatively considered (Nelson and MacLennan, 1995).

The travel distance varies between 9 m and 27 m. Therefore, the movement time of

occupants travelling to the lift lobbies is taken from 15 seconds (minimum) to 45

seconds (maximum). During this period, all the occupants are assumed to travel to

the lift lobby.

Time to arrive in lift lobby: Fire detection time, pre-movement time and movement

time are added to determine the time to arrive in the lift lobby. However, occupants

may also arrive without engaging themselves in response and coping activities.

Therefore, on the fire-affected floor:

• Minimum time to arrive in lift lobby = tFD + tLM

= 90 + 15 = 105 seconds

• Maximum time to arrive in lift lobby = tFD + tLPM + tLM

= 140 + 186 + 255 + 45 = 626 seconds

On other floors:

• Minimum time to arrive in lift lobby = tFD + tLM

= 140 + 15 = 155 seconds

• Maximum time to arrive in lift lobby = tFD + tLPM + tLM

= 140 + 186 + 45 = 371 seconds

(Coping time is not included here since the residents do not encounter fire).

The occupants arrive between the minimum and maximum times. These times are

considered as random variables in the stochastic modelling.

Occupant arrival rate: Occupant arrival rate is the average number of occupants

arriving in the lift lobby per unit time. Lift system provides discrete mode of vertical

transportation. The occupants arrive for lift evacuation in groups or one at a time in a

period t according to a Poisson process (Alexandris, 1977). Assuming that the

98

probability of x calls are registered for lifts in time t for an average rate of occupants’

arrivals is λ, then

( )!

)(

x

etxp

tx λλ −

= 4.7

where

t is the time period of the counting process (second)

x is the number of occupants who arrive at designated point

p(x) is the probability mass function (PMF) for x number of people arriving within

the time period t

λ is the average number of occupants arriving per unit time (s-1)

The PMF given by Eq. 4.7 is the Poisson (λ t) distribution. The values of occupant

arrival in the lift lobby are expressed in Poisson distribution with the help of ARENA

input analyser. ARENA input analyser is a component of ARENA model to

determine the quality of fit of probability distribution function to input data (number

of occupants and time variables). Figure 4-7 shows the occupant arrival for the entire

floor population and one-fourth of the floor population. The Poisson distribution

predicts occupants’ arrival to lift lobby as a peak load at any given time. The

probability of occupant arrival is uniformly distributed between two extreme time

intervals. The calculated values are between 105 and 626 seconds on fire-affected

and ground floors, and 155 and 371 seconds on other floors for 32 occupants (entire

population of the floor) and 8 occupants (one-fourth of the floor population). The

expressions are given as Poisson (16.6) on the fire-affected floor, Poisson (6.81) on

the rest of the floors and Poisson (23.5) on the ground floor for the entire floor

population {Poisson (16.6) – distribution function contains the value of mean}.

Similarly the expressions are given as Poisson (67.6) on the fire-affected floor,

Poisson (27.6) on the rest of the floors and Poisson (89.5) on the ground floor for

one-fourth of the floor population.

99

Figure 4-7 – Poisson Distribution for Occupant Arrival (Lifts)

A discrete function of 25% delay is added to the model, which may be due to

occupants holding the lift for their next of kin (the discrete function in ARENA

returns a sample from a user-defined discrete probability distribution). A 25% of

additional delay is conservatively considered keeping in view the aged and disabled

population and their assistance. However, a general population can also hold the lift.

The inter-floor movement is considered from occupants’ movement time from lift to

apartment unit in 20 seconds, staying in apartment for 20 seconds and return in 20

seconds on average and therefore an exponential distribution function with a mean of

20 seconds is used for all. Exponential distribution contains a value of mean and

provides a random variable for stochastic modelling. These times are added for

social reasons.

Lift movement speed: Lift movement speed varies with the hall call and cabin call.

Velocity and acceleration are incorporated in the stochastic modelling.

100

4.5.4 Stair Simulation Variables

The building indicated in the lift model was considered for stair analysis. On each

floor, there are two stairways. Arrival patterns of the occupants are the same at lift

lobbies and stair lobbies. However, the floor population is equally distributed on both

the stairs with half the floor population using one stair (whereas there is only one lift

lobby and the entire floor population is arriving at one lift lobby). Total time frame

for evacuees’ arrival at lift lobby or stair lobby is the same.

Occupant movement: The down-movement, upper floor movement and inter-floor

movement are defined below:

� From the upper floors – 50% of building population, 592 (16 × 37) occupants

would be using one stair to go to ground or intermittent floors

� Intermittent floors – 6 occupants would be going to intermittent floors

(approximately 1% of total population – from 592 occupants)

� From the ground floor to the upper floors – 6 outsiders would like to go to

upper levels (approximately 1% of total population – 592 occupants)

– Total number of occupants using one stair = 598

Movement from the ground floor to the upper floors in the stairs is considered for

building corporate management and general public, not the emergency crew

members as they are permitted to use lifts during fires. Disabled and aged persons

can go to stair lobbies for temporary staging and then can be assisted by others.

Pre-movement and movement times: The pattern of occupant arrival for stairs is

considered similar to lift pre-movement (response and coping) and movement time.

The calculated response time is 186 seconds and coping time is 255 seconds. The

movement time of occupants travelling to the stairs is taken as 15 seconds

(minimum) to 45 seconds (maximum). During this period, all the occupants would

have travelled to stairs.

Occupant arrival rate: Figure 4-8 shows the occupant arrival rate for the entire

floor population and three-fourths of the floor population. The expressions Poisson

101

(34) on the fire-affected floor, Poisson (14.5) on other floors and Poisson (33) on

ground floor are considered for the entire floor population (see Figure 4-8).

Similarly, the expressions Poisson (44.8) on the fire-affected floor, Poisson (18.5) on

other floors and Poisson (49.3) on ground floor are considered for the three-fourths

of the floor population. These distributions are given after equal distributions of

population between the two stairs in the building.

Figure 4-8 – Poisson Distribution for Occupant Arrival (Stairs)

Stair capacity: It is generally considered that a walker on stairs needs to perceive

two vacant treads ahead and occupies an area of approximately 0.7 m2. Thus free

flow design is possible at a density of 0.6 P/m2 and full flow design is possible at a

density of 2.0 P/m2. The walking speed is considered to be 1.25 P/m2, which will

permit passer-by, intermittent floors movement or reverse flow, although severely

restricted (Kisko et al., 1998). The stairway is 1.2 m wide and considering a 27

degree stairway with 3 m between floors and with an intermediate landing, the

102

horizontal distance walked in descending one floor is the ratio of 3 m to the tangent

of 27 degrees, which works out to be 11.4 m. Therefore the area is 13.68 sq.m.,

which would hold approximately 17 persons in the stair in one floor height.

Stair walking speed: Walking speed in stairs varies from 0.52 to 0.62 m/s (0.57) for

general public and 0.22 to 0.79 m/s (0.505) for occupants carrying children on

average in non-crowded stairs in apartment buildings (Proulx, 1995). The walking

speed of occupants carrying children varies to a large extent when compared to

general public. The speed was 0.45 m/s for small children (aged 2-5, not carried by

adults) and 0.43 m/s for over 65 years old (Proulx, 1995). The research data of stair

walking speed of aged people is over optimistic. It is not expected that aged and

physically weak people can maintain their travel speed all the way to the ground.

Aged and physically weak people may require rest in the stairs. This may cause

impedance to the traffic flow. There is a percentage of population who are physically

immobile and can not use stairs at all. If lifts are not permitted, they have to wait for

fire brigade intervention and therefore take much longer time to evacuate.

Keeping in view the emergency situation for flowing condition, a walking speed

varying from 0.43 to 0.79 m/s with an average of 0.54 m/s is considered at a density

1.25 P/m2 for stair analysis. The average of 0.54 m/s is obtained based on the 75%

of normal population having a walking speed of 0.43 m/s and the 25% population,

including aged and disabled persons, having a walking speed of 0.43 m/s. A

triangular distribution for stair travelling time of 32 travellers with a minimum of

13.4 seconds, mode (most likely value) of 19.6 seconds, and maximum of 24.6

seconds is obtained for one floor level (see Figure 4-9). Triangular distribution is

skewed with a mode of 19.6 seconds. This distribution is applied to all floor levels.

Figure 4-9 – Triangular Distribution for Occupants’ Stair Travelling Time

103

4.6 Simulation Models

4.6.1 Lift Simulation Model

The block diagram indicated in Figure 4-10 shows the lift simulation model for the

hypothetical building. The model is comprised of four major zones: controller and

lift zone, passenger zone, re-entry zone and exit zone. Animation Zone is supportive

for demo and is not shown here. The various input components relating to the lift

system used in these zones are such as entity – occupants; attributes – upper levels

and main terminal (ground level); activities – vertical traveling; events – arrival at

the lift lobbies and arrival at the main terminal; state variables – number of occupants

waiting at each floor level and number of evacuees in transit and resource – number

of lifts.

Elev ato r Ariv a lS tation 1As s ign 1

Arriv a l Floor 01Pas s enger

Arriv a l Floo r 02Pas s enger

Tr ue

False

Dec ide 1

S tation 2 Dis pos e 1

A partment 02 D elay 1 R oute 1

Floor 02. Q ueue

Queue

Floor 01. Q ueue

Queue

ReleasedAt Floor ==What Floor

Floor Num ber ==What Floor

( Elevat or _Load < Elevat or _Capacit y) * (What Floor > Floor Num ber )Else

Dec ide 2

As s ign 2

Can_Fi t_1_Pas s enger

Tr ue

False

R oute 2

As s ign 3

As s ign 4

As s ign 5

O r iginal

Member s

D ropoff 1

S tation 8

Passenger Zone

Controller and Lift Zone

Re-entry Zone

Route 5

As s ign 6

Elevat or _Rem ainingCapQ ueues_2(What Floor )

P ickup

As s ign 8

TWhat Floor ( 1) | | ( NQ (Q ueues_2( Floor Num ber ) ) < 1) * ( Dispose 1. Number out <1210) * EWhat Floor ( 2)

W hi le

E ndW hi le

FloorD ropP ickD elay

( Q E( 1) + Q E( 2) + Q E( 3) + Q E( 4) + Q E( 4) + Q E( 5) + Q E( 6) + Q E( 7) ) + Elevat or _Load >= 1

S can

Create 43 As s ign 9 Rec o rd 1H old 1 Dis pos e 2As s ign 10

As s ign 11

Exit Zone

0

0

0

0

0

0

0 0

Figure 4-10 – Various Zones in Lift Simulation Model (two floors only)

Assign modules have attributes of lift number, lift load, floor number, released at

floor, lift remaining capacity and acceleration. Station module contains the set of

floors. Decide modules have conditions of expression for lift load, lift capacity, floor

number and where to go. Delay module assigns the value for pick-drop delay during

the lift movement and occupants’ momentarily staying in an apartment during

intermittent floor movement. The passenger exit zone consists of the location of

104

evacuees with an arrival expression on various floor levels. The evacuees are in the

waiting queue for lifts in the lobby. The passenger re-entry zone consists of the

location for evacuees’ re-entry in the apartment unit and lift waiting station.

4.6.2 Stair Simulation Model

The block diagram indicated in Figure 4-11 shows the stair simulation model for the

hypothetical building.

S tation 2 Dispose 1

A partment 02 Delay 1 Route 1

Floor 37Passenger Arriva l

S tation 65

Floor 38Pas senger Arriva l

S tation 66

A partment 38 Delay 39 Route 39

Route 40

S tation 80

Route 41

S tation 81 R oute 42

Route 43

S tation 117

Route 114

S eize 1

S eize 2 Release 1

Decide 1

Want edFloor == 37

Floor _ent er ed == 37 && Want edFloor > 37

Floor _ent er ed == 37 && Want edFloor < 37

Want edFloor > 37Else

Ass ign 2

Release 39

Route 115S eize 78

Route 117S eize 79

Route 118S eize 80 Release 40

Release 41 Route 119

Ass ign 74

Ass ign 109

Decide 73

Want edFloor == 38

Else

Release 147 Route 295

Floor 01Passenger Arriva l

S tation 155 R oute 296Ass ign 110

Decide 74

Floor _ent er ed == 01 && Want edFloor > 01

Else

Route 297S eize 221

Stair Traveller Zone

Re-entry Zone

Exit Zone

0

0

0

Figure 4-11 – Various Zones in Stair Simulation Model (three floors only)

The model is comprised of three major zones: stair traveler zone, re-entry zone and

exit zone. These zones contain several modules representing various attributes and

variables. Assign modules have attributes of stair capacity and walking speed.

Station module contains the set of floors. Decide modules have conditions of up or

105

down movement. The traveller exit zone consists of the location of travellers with an

arrival expression on various floor levels. The travellers’ re-entry zone consists of the

location for travellers’ re-entry in the apartment unit. The delay module indicates the

travellers’ momentarily staying in apartment during intermittent floor movement.

Appendix F gives the details of SIMAN ARENA language and the overall model

used for lift and stair systems (Figures F1 and F2).

4.7 Simulation Results

For down peak traffic scenario, 100 replications of two hour traffic for the maximum

evacuation period were generated. The test results are based on full building

evacuation using lifts and stairs separately. Detailed results are given in Appendix G.

The animations of both simulation models are shown in Figure 4-12.

Figure 4-12 – Animation of Lift and Stair Simulation Models at 300 seconds

106

4.7.1 Lift Simulation Model

Lift for Total Evacuation: Issues relating to lift evacuation system were adjudged

for three floor levels (viz. bottom level, middle level and top level). Output variables

relating to RSET (lift waiting time, lift transportation time and lift evacuation time)

and the number of evacuees in queue were analyzed for 100 replications (see

Appendix G, Tables G1, G2 and G3). Tables 4-5 to 4.7 give the values of the lift

time periods and the number of evacuees in queue for fire occurring at three levels.

Table 4-5: Lift Time Periods and Number of Evacuees in Queue (2nd floor fire)

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 721.24 8.11 601.69 827.20 0 2371.36

Lift Transportation

Time, tLT (second) 41.83 0.21 39.73 44.31 1.34 251.80

Lift Evacuation

Time, tLE (second) 2321.74 105.48 1158.68 3169.28 51.69 3516.24

Number of Evacuees in Queue

8.17 0.87 0.90 19.96 0 34

Table 4-6: Lift Time Periods and Number of Evacuees in Queue (19th floor fire)

Identifier Average

Half

Width

Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 719.66 7.89 610.03 810.20 0 2402.26

Lift Transportation

Time, tLT (second) 41.92 0.23 39.49 44.37 1.50 297.59

Lift Evacuation

Time, tLE (second) 2269.30 111.36 1176.37 3291.86 4.03 3564.95

Number of Evacuees in Queue

8.14 0.86 1.02 19.48 0 34

Table 4-7: Lift Time Periods and Number of Evacuees in Queue (38th floor fire)

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 721.95 8.50 605.19 839.67 0 2406.00

Lift Transportation

Time, tLT (second) 41.83 0.25 39.57 42.12 1.49 279..78

Lift Evacuation

Time, tLE (second) 2270.00 116.39 1180.71 3403.47 20.85 3403.47

Number of Evacuees in Queue

8.19 0.86 0.94 20.00 0 34

107

The average is the mean time period of 100 simulations. The "Half Width" column is

included to determine the reliability of the results from all the replications. A value is

interpreted by saying "in 95% of repeated trials, the sample mean would be reported

as within the interval sample mean ± half width". The maximum and minimum

averages are the averages of simulations. The maximum and minimum values

indicate the maximum and minimum of simulations. It may be inferred from the

similar results presented in the above three tables that the model did not consider the

impact of the location of the fire floor on evacuation.

Lift for 25% Population: Considering the use of lifts for a limited number of

evacuees, the lift time periods and the number of evacuees in queues were calculated

for three floor levels (see Appendix G, Tables G1, G2 and G3). The results are

shown in Tables 4-8 to 4-10.

Table 4-8: Lift Time Periods and Number of Evacuees in Queue (2nd

floor fire) –

25% population

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 299.91 7.36 211.23 424.95 0 1336.34

Lift Transportation

Time, tLT (second) 40.09 0.30 36.61 44.16 1.96 237.43

Lift Evacuation

Time, tLE (second) 1613.20 74.88 785.97 2319.05 0.88 2349.90

Number of Evacuees in Queue

1.28 0.14 0.22 3.48 0 9

Table 4-9: Lift Time Periods and Number of Evacuees in Queue (19th floor fire) –

25% population

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 294.96 7.34 208.79 387.43 0 1391.90

Lift Transportation

Time, tLT (second) 40.30 0.30 36.48 43.89 1.95 283.52

Lift Evacuation

Time, tLE (second) 1586.56 71.77 706.84 2401.98 8.52 2459.25

Number of Evacuees in Queue

1.28 0.14 0.22 3.52 0 9

108

Table 4-10: Lift Time Periods and Number of Evacuees in Queue (38th floor fire)–

25% population

Identifier Average

Half

Width

Minimum

Average Maximum

Average Minimum

value

Maximum

value

Lift Waiting Time,

tLW (second) 298.13 6.55 241.85 403.91 0 1261.44

Lift Transportation

Time, tLT (second) 40.36 0.34 36.91 45.43 1.49 254.90

Lift Evacuation

Time, tLE (second) 1560.94 80.42 874.57 2306.55 3.63 2306.55

Number of Evacuees in Queue

1.26 0.14 0.24 3.47 0 9

It can be seen from the lift simulation results that building population was reduced by

three quarters but the average lift waiting time, lift transportation time and lift

evacuation time were only reduced by just over half. However, the number of

evacuees in the queue has been reduced considerably.

4.7.2 Stair Simulation Model

Stair for Total Evacuation: Evacuees’ using stairs was analyzed for evacuation

periods for 100 replications (see Appendix G, Tables G4 and G5). Tables 4-11 to 4-

13 give the values of stair travelling time and stair evacuation time. The stair waiting

time varied from 0 to 9 seconds while the number of travellers in queue was less than

0.1. The stair waiting time and the number of travellers in queue were not significant

in relation to bottleneck or queuing, and are therefore not given.

Table 4-11: Stair Time Periods (2nd

floor fire)

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 373.10 0.45 367.77 377.90 14.62 1554.79

Stair Evacuation

Time, tSE (second) 1556.03 60.05 800.35 2114.19 17.58 2114.19

109

Table 4-12: Stair Time Periods (19th floor fire)

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 372.69 0.48 367.40 379.44 14.62 1534.87

Stair Evacuation

Time, tSE (second) 1544.13 62.35 917.38 2157.91 18.71 2157.91

Table 4-13: Stair Time Periods (38th floor fire)

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 372.82 0.74 366.30 378.38 14.62 1510.38

Stair Evacuation

Time, tSE (second) 1559.61 57.42 822.04 2002.17 10.85 2002.17

Stair for 75% Population: For strategic evaluation, occupants’ evacuation for 75%

of the population was also analyzed for evacuation periods for 100 replications (see

Appendix G, Tables G4 and G5). Tables 4-14 to 4-16 give the values of stair

travelling time and stair evacuation time.

Table 4-14: Stair Time Periods (2nd floor fire) – 75% population

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 366.98 0.35 363.40 372.06 14.52 1058.87

Stair Evacuation

Time, tSE (second) 1102.51 39.68 619.42 1476.53 11.44 1476.53

Table 4-15: Stair Time Periods (19th floor fire) – 75% population

Identifier Average

Half

Width Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 367.72 0.38 362.63 373.80 14.62 1192.78

Stair Evacuation

Time, tSE (second) 1153.90 42.06 505.61 1462.94 66.98 1631.48

Table 4-16: Stair Time Periods (38th floor fire) – 75% population

Identifier Average

Half

Width

Minimum

Average Maximum

Average Minimum

value

Maximum

value

Stair Travelling

Time, tST (second) 367.14 0.49 362.42 373.72 14.69 1382.95

Stair Evacuation

Time, tSE (second) 1312.93 50.78 672.45 1701.60 86.57 1701.60

110

It can be seen from the stair simulation results that even when the building

population was reduced by one quarter, the average stair travelling time and stair

evacuation time were only reduced marginally.

4.8 Analysis of Results

The parameters i.e. lift waiting time (tLW), lift transportation time (tLT), lift pre-

evacuation time (tLPE), lift evacuation time (tLE) and number of evacuees in queue,

were analysed for the hypothetical 38-storey building. The occupant profile, building

symmetry and evacuation strategy were the same for both lift and stair evacuation

systems. The results show that the lift and stair evacuation times are not varying

significantly irrespective of the origin of fire at floor levels in the building.

4.8.1 Lift Waiting Time

While evacuating the entire building using the lifts, a few evacuees are required to

wait for a considerable time near the lift lobby before the first lift arrived (see

Appendix G, Tables G8, G10 and G12). However, the minimum and maximum

waiting time values reaching from 0 to 2400 seconds can not be considered an

isolated lowest or highest values (see Tables 4-5 to 4-7). These are low or high

values and represents an approximation to the most efficient or worst case

eventualities for all extreme values on the floors. The maximum average lift waiting

time is approximately 1000 seconds and is considered justified (see Figure 4-13).

0

200

400

600

800

1000

1200

0 5 10 15 20 25 30 35 40

Floor Level

Time (second)

2nd floor fire

19th floor fire

38th floor fire

Figure 4-13 – Lift Waiting Times during the Fire Occurrences at Three Levels

111

During down peak traffic in the event of fires, a lift car may fill at two, three or four

floors at different levels and then makes a run to the ground floor. Although this

reduction in number of stops results in a shorter round trip time and higher handling

capacity, the result predicted that the evacuees on the lower floors may face a slightly

longer waiting period. Figure 4-13 also shows that the waiting time is lower at 30th

level (dip in the graph), which is due to the fact that the evacuees arrive at the early

stage than the upper levels and use the lifts for downward movement (a stochastic

input). Waiting time is also slightly lower at the fire-affected floor, which is caused

by longer time to evacuate from the unit due to occupants’ coping action.

Mean lift waiting time was approximately 720 seconds from 100 replications. The

lifts served within 297 seconds with the reduction in number of evacuees (see

Appendix G, Tables G6 and G7). The probability distribution functions (PDFs) were

given with the help of @RISK (Palisade, 1996) for determining the mean and

standard deviation. The PDF and a comparison are shown in Figure 4-14 (a), (b) and

(c).

Logistic(720.913, 16.582)

672.1

5.0%

769.7

95.0%

0

0.004

0.008

0.012

0.016

640 660 680 700 720 740 760 780 800

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(a) PDF for Lift Waiting Time (100% population)

112

Weibull(2.9525, 72.645) Shift=+232.880

259.4

5.0%

338.2

95.0%

0

0.005

0.01

0.015

0.02

240 260 280 300 320 340 360

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(b) PDF for Lift Waiting Time (25% population)

0

100

200

300

400

500

600

700

800

900

0 20 40 60 80 100

Replication

Time (second)

Lift Waiting Time (100% Population)

Lift Waiting Time (25% Population)

(c) Comparing the Average Lift Waiting Times

Figure 4-14 – Lift Waiting Times

4.8.2 Lift Transportation Time and Stair Travelling Time

Mean lift transportation time was approximately 42 seconds for 100% population and

approximately 40 seconds for 25% population. The mean stair travelling time was

approximately 372 seconds for 100% population and 366 seconds for 75%

population (see Appendix G, Tables G6 and G7). The lift transportation times and

stair travelling times are not varying significantly with the change in evacuation

proportion. This shows that there is unnecessarily withholding of lift and no

bottleneck in the stair. The PDFs and comparison are shown in Figure 4-15 (a) to (e).

113

InvGauss(13.901, 4826.739) Shift=+27.966

40.675

5.0%

43.127

95.0%

0

0.1

0.2

0.3

0.4

0.5

0.6

40 40.5 41 41.5 42 42.5 43 43.5 44

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(a) PDF for Lift Transportation Time (100% population)

Logistic(40.22769, 0.55930)38.581

5.0%41.875

95.0%

00.050.10.150.20.250.30.350.40.45

37.5 38 38.5 39 39.5 40 40.5 41 41.5 42 42.5

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(b) PDF for Lift Transportation Time (25% population)

LogLogistic(351.289, 21.534, 25.466)

370.471

5.0%

375.462

95.0%

00.050.10.150.20.250.3

0.350.4

369 370 371 372 373 374 375 376 377

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(c) PDF for Stair Travelling Time (100% population)

114

Logistic(367.24305, 0.76537)

364.989

5.0%

369.600

95.6%

0

0.1

0.2

0.3

0.4

363 364 365 366 367 368 369 370 371 372

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(d) PDF for Stair Travelling Time (75% population)

0

100

200

300

400

0 20 40 60 80 100

Replication

Time (second)

Lift Transportation Time (100% population)

Lift Trasnportation Time (25% population)

Stair Travelling Time (100% population)

Stair Travelling Time (75% population)

(e) Comparing the Average Lift Transportation and Average Stair Travelling

Times

Figure 4-15 – Lift Transportation and Stair Travelling Times

4.8.3 Lift Pre-Evacuation Time and Stair Pre-Evacuation Time

The lift pre-evacuation time was obtained from lift pre-movement time, lift

movement time and lift waiting time. The lift pre-evacuation time is basically the

RSET on the fire-affected floor (if the lift shafts are free from fire effluents), and is

used in the next chapter. The pre-movement time was calculated from simulation

timer on a fire-affected floor. Mean lift pre-evacuation time was approximately 1346

seconds for 100% population and 924 seconds for 25% population. The PDFs and

comparison are shown in Figure 4-16 (a) and (b).

115

Logistic(1346.913, 16.582)

1298.1

5.0%

1395.7

95.0%

0

0.004

0.008

0.012

0.016

1260 1280 1300 1320 1340 1360 1380 1400 1420

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(a) PDF for lift pre-evacuation time for 100% population

Weibull(2.9525, 72.645) Shift=+858.880

885.4

5.0%

964.2

95.0%

0

0.005

0.01

0.015

0.02

860 880 900 920 940 960 980 1000

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(b) PDF for lift pre-evacuation time for 25% population

Figure 4-16 – Lift Pre-Evacuation Times

As the stair waiting time was 0 to 9 seconds and did not reflect any significance of

bottleneck or queuing, stair pre-evacuation is confined to stair pre-movement time

(coping and response times) and stair movement only. Therefore, stair pre-movement

time is not given here.

116

4.8.4 Lift Evacuation Time and Stair Evacuation Time

The mean lift evacuation time was approximately 2288 seconds while the mean stair

evacuation time was approximately 1550 seconds, which is less, in comparison. The

mean lift evacuation time for 25% of the population was approximately 1624

seconds and close to the stair evacuation time of the entire population (see Appendix

G, Tables G6 and G7). The mean stair evacuation time for 75% population was

approximately 1193 seconds. The PDFs and comparison are shown in Figure 4-17 (a)

to (e).

BetaGeneral(1.4834, 1.2550, 1259.2, 3158.9)

1469

5.0%

3036

95.0%

0123456789

1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200

Time (second)

F(t)

Values x 10^-4

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(a) PDF for Lift Evacuation Time (100% population)

Triang(877.79, 1856.6, 1914.5)

1103

5.0%

1860

95.0%

0

0.0007

0.0014

0.0021

0.0028

900 1120 1340 1560 1780 2000

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(b) PDF for Lift Evacuation Time (25% population)

117

Triang(772.11, 1971.5, 2128.0)

1057

5.0%

2025

95.0%

00.20.40.60.81

1.21.41.61.8

800 1000 1200 1400 1600 1800 2000 2200

Time (second)

F(t)

Values x 10^-3

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(c) PDF for Stair Evacuation Time (100% population)

BetaGeneral(2.9555, 1.0058, 502.56, 1428.40)

837.6

5.0% 1412.2

95.0%

0

0.0021

0.0042

600 700 800 900 1000 1100 1200 1300 1400 1500

Time (second)

F(t)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(d) PDF for Stair Evacuation Time (75% population)

0

500

1000

1500

2000

2500

3000

3500

0 20 40 60 80 100Replication

Time (second)

Lift Evacuation Time (100% population)

Lift Evacuation Time (25% population)

Stair Evacuation Time (100% population)

Stair Evacuation Time (75% population)

(e) Comparing the Average Lift Evacuation and Average Stair Evacuation Times

Figure 4-17 – Lift and Stair Evacuation Times

118

4.8.5 Number of Evacuees in Queue

The number of evacuees in queue ranged from 8.14 to 8.19 during the use of lifts for

the entire population (see Appendix G, Tables G9, G11 and G13). No queue was

observed for 25% of the population as the number of evacuees was less than 2 (see

Figure 4-18 (a), (b) and (c)).

Lognorm(3.2721, 0.43962) Shift=+4.9022

7.505

5.0%

8.943

95.0%

00.10.20.30.40.50.60.70.80.91

7 7.5 8 8.5 9 9.5

Number of Evacuees in Queue

F(x)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(a) PDF for Number of Evacuees in Queue (100% population)

ExtValue(1.22448, 0.11107)1.1026

5.0%1.5544

95.0%

00.51

1.52

2.53

3.54

4.5

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7

Number of Evacuees in Queue

F(x)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

(b) PDF for Number of Evacuees in Queue (25% population)

119

0

1

2

3

4

5

6

7

8

9

10

0 20 40 60 80 100

Replication

Number in Queue (person)

Number in Queue (100% population)

Number in Queue (25% population)

(c) Number of Evacuees in Queue

Figure 4-18 – Number of Evacuees in Queue in Lift System

4.8.6 Findings

It is determined that lifts provide longer evacuation time periods in comparison to the

use of stairs only. However, using lifts for a limited number of persons (25%

population) gives approximately equal evacuation time periods. The use of lifts for

evacuation can be explored for a limited population because

• there is no queue (less than 2 persons in queue – nearly 1.2 persons only).

Evacuees will not have competitive behaviour for lifts; and

• a total of lift waiting and transportation times (298 second + 40 seconds) is

less than the total stair travelling time (373 seconds). Hence evacuees’

anxiousness or decision uncertainty would be minimal.

The results also showed that for apartment building having 38 storeys with 32

persons on each floor, the stairs are the fastest way of evacuation for the entire

population. Figure 4-17 gives the means and standard deviations of output variables

derived from the PDFs with the help of @RISK. These variables will be used in

Chapter 5. The results presented in the table did not demonstrate much benefit with

regard to timings in using lifts. In a residential building, the evacuation times by lifts

become more favorable for buildings with more than 100 persons per floor (50

persons per staircase) and with more than 50 floors, or for 200 persons per floor and

with more than 25 floors (Siikonen and Hakonen, 2003). The generic residential

120

building is lower than 50 floors and the population per floor is less than 200. As per

observations made by Klote et al., (1993a) (see Chapter 1, Section 1.1) the use of

lifts for emergency evacuation is also for mega high-rise buildings and are not

applicable to normal high-rise buildings.

Table 4-17: Means and Standard Deviations of Output Variables

Output Variables Evacuation Strategy Mean µL (second)

Standard

Deviation σL (second)

100% population 720.91 30.07 Lift Waiting Time (tLW)

25% population 297.70 23.90

100% population 41.86 0.74 Lift Transportation Time (tLT)

25% population 40.22 1.01

100% population 1346.91 30.07 Lift Pre-Evacuation Time (tLPM)

25% population 923.70 23.90

100% population 2288.30 489.56 Lift Evacuation Time (tLE)

25% population 1623.9 302.83

100% population 8.17 0.43 Number of Evacuees in Queue (x)

25% population 1.28 0.14

100% population 372.87 1.54 Stair Travelling Time (tST)

75% population 367.24 1.38

100% population 1549.6 237.83 Stair Evacuation Time (tSE)

75% population 1187.81 181.8

4.9 Model Verification

Computer model ELVAC (Klote et al., 1991) was used to verify the results of lift

evacuation model. One hundred replications were made in order to analyze the

performance of ARENA model. During 100 replications, a deterministic approach

was used without incorporating variables relating to human movement, inter-floor

and upward movements. A pilot run was made in order to perform the system

verification from ELVAC. The parameters for ELVAC model include number of

persons, number of floors, floor height, number of lifts, lift capacity, velocity,

acceleration, trip inefficiency and type of doors. The parameters of human movement

such as pre-movement time and movement time are not included in the model

verification. The details of ELVAC model and results are shown in Appendix H.

121

ELVAC determined the time taken to evacuate the entire building as 1590.5 seconds

while ARENA determined it to be 1618.8 seconds.

A mathematical derivation was used to verify the results of stair evacuation model.

The travel time via stairs was calculated for the evacuees in the building. The

formula established by Melinek and Booth (1975) for calculating the evacuation time

(Eq. 2.8, Chapter 2, Section 2.8.2) was used for the evacuees travelling down the

stairs (see Table 4-18). Hand calculations are given in Appendix H. Hand

calculations of the time taken to evacuate the entire building gave 1116 seconds

while ARENA determined it to be 1111 seconds (Table 4-18). The variations in

simulation results for both the systems (lift simulation model and stair simulation

model) are less than 2%.

Table 4-18: Verification of ARENA model for Lifts and Stairs

Building Evacuation Time (second) Mode of

Evacuation ARENA Model ELVAC/ Hand Calculation % Variation

By lifts 1618.8 1590.5 (ELVAC) + 1.78

By stairs 1111 1116 (Hand calculation) - 0.44

4.10 Conclusion

The pilot survey demonstrated that residents were inclined to use lifts as an

alternative evacuation facility in buildings. The confidence interval of lift use during

normal circumstances was slightly more than the use of stairs during emergencies.

The occupants can be more confident in using lifts as an alternative evacuation

facility during fire emergencies. Aged and disabled revealed their concern for life

safety while living on upper levels and can be considered for lift evacuation. At least,

25% of the population can be considered using an alternative evacuation facility in

the building. The majority of residents were not trained in the Emergency

Evacuation Procedure. The use of lifts for limited population may add complexity in

the evacuation procedure. Fire brigade personnel also expressed their concern about

the current situation in high-rise building evacuations.

122

With the help of stochastic evacuation models, the lift time periods and the number

of evacuees in queue were determined for the lift evacuation system. Random

variables relating to human social, physical characteristics and a priori heuristics of

the lift domain were considered. The results did not demonstrate much benefit with

regard to total times in using lifts. The time for which occupants would wait at a lift

station was uncertain. However, if lifts are permitted for use by 25% of the

population, the lifts serve the individuals within the acceptable time frame in

comparison with stairs. The number of evacuees in queue was less than 2 persons.

This signifies that the human behavioural response such as decision uncertainty

would be minimal for one-fourth of the population using the lift system. This aspect

will be further analysed as an integral part of other risks in Chapter 7. Lift evacuation

time is related to the total building evacuation that needs to be compared with time to

exceed the tenability limits of life threatening conditions in the evaluation of safety

indices. With the help of field model (FDS), available safe evacuation times (ASETs)

will be determined in the next chapter.

123

5. FIRE HAZARD MODELS OF LIFE THREATENING CONDITIONS

5.1 Introduction

Available safe evacuation time (ASET) is an important parameter in fire safety

engineering (ABCB, 2005b). ASET is defined as the time period between fire

initiation and onset of life threatening conditions. ASET depends on a variety of

variables associated with the fire scenario, including the intensity of the fire and the

geometry of the fire (distribution of fire load), fire protection measures and building

spatial environment. ASET should be reasonably greater than the required safe

evacuation time (RSET) for a successful building evacuation during fire

emergencies. In the absence of adequate fire safety measures, smoke and hot gases

may spread to evacuation routes (for example, lift lobbies, lift shafts and stair shafts)

and occupants at the fire-affected floor and other floors may expose to hazardous

conditions leading to psychological and physiological effects. The parameters such

as smoke visibility, CO, CO2 and O2, temperature and radiant heat exposure are used

in evaluating hazardous conditions. Such parameters can be obtained from fire tests

or computational models. The time to untenable conditions that may lead to

incapacitation or death is predicted from the combination of smoke toxic products

and heat exposure.

This chapter is focused on the development of fire hazard models under variable

conditions to determine ASET. Given a typical configuration and fuel load in an

apartment building (BCA 2005 Specification Figure 2.1 specifies 5 MW fire for

residential buildings), the time to attain incapacitating dosage is obtained. With

known intake rate of intoxicants by human, times to exceed tenability limits are

computed using the field model ‘FDS’ (McGrattan et al., 2004). The parameters

relating to fire and smoke are probabilistic in nature. By specifying uncertainty

(stochastic) in the input variables, the variable outcomes are obtained for defining a

probability distribution. Wind speed and stack effect influence the smoke spread in

high-rise buildings (Klote, 2003). Therefore, wind speed and varying vertical

location of fire are considered as input variables. The piston effect by lift car

movement is not significant in multiple shafts in high-rise buildings (Klote and

124

Tamura, 1986) and is therefore not considered. The concept of safety index is used in

the evaluation of evacuees’ safety.

5.2 Analysis of Fire Hazards

With the help of FDS, a fire hazard model was prepared using stochastic uncertainty

that is capable of predicting hazards in the evacuation routes. Various phases

occurring sequentially in space and time during the fire growth period and occupant

evacuation were specified together with the associated probability of occurrence.

The objectives of this analysis are to:

1. develop a model for the fire and toxic hazards under variable conditions and

determine the probable time when evacuees are predicted to be exposed to the

incapacitation dose of fire effluents (ASET); and

2. determine the safety of evacuees in evacuation routes under variable

conditions; and

3. analyse the effect of wind speed and stack effect on the spread of smoke and

toxic gases in the lift system; and

4. analyse potential ways of reducing the risk associated with the use of lifts.

5.2.1 Effects of Fire Effluents and Evaluation Criteria

Psychological and Physiological Effects: There are psychological and physiological

effects of fire effluents. The evacuees may be exposed to incapacitated dose of fire

effluents in evacuation routes. Smoke reduces the visibility for evacuees and there

may be needless psychological unrest or panic (Jin, 2002). Evacuees may be trapped

in the early stage of fire. However, the loss of visibility is not a direct cause of death.

Continuous exposure of smoke, heat and toxic products affect the evacuees’

capability to escape and can lead to physiological effects. The physiological effects

may be varying degrees of impaired judgment, disorientation, reduced capability of

performing work, loss of motor coordination, unconsciousness and deaths. The

physiological effects of fire effluents causing deaths are divided into the following

categories (Miller, 2005) and are used in this research:

125

– inhalation of toxic products of combustion {smoke, CO2, CO, and other

poisonous gases, hypoxia (lack of oxygen) and asphyxia}; and

– consequences of exposure to fire (burns, thermal injuries to airways and

incineration); and

– shock from injuries that precipitate death from pre-existing health conditions

(cardiac failure and respiratory diseases).

Evaluation Criteria: Evaluation criteria are based on the concept of fractional

effective dose (FED) of human incapacitation model (see Chapter 2, Section 2.6.3).

It considers concentration and time and assumes that incapacitation occurs after

adding the effects of exposure to a toxic concentration at each time period. The

following FEDs are used in calculating the effects of fire effluents:

• FED smoke: FED for smoke incapacitation is the summation of the acquired

doses of smoke. When the FED smoke reaches unity, visual obscuration is

assumed to occur. At an FED smoke value of 1.0, many evacuees are likely to

be visually obscured or impaired by smoke. The extinction coefficient should

not be higher than 0.5 m-1. Visibility is determined at 4 m horizontal

distance at a focal point in small compartments (see Appendix L).

• FED asphyxiant: FED for asphyxiant incapacitation is the summation of the

acquired doses of asphyxiant toxicants. When the FED asphyxiant reaches

unity, incapacitation is assumed to occur. At an FED asphyxiant value of 1.0,

many evacuees are likely to be overcome by the combined effects of

asphyxiant toxicants. The combined effects of asphyxiant toxicants should

not exceed one-tenth of the dose.

• FED heat: FED for heat incapacitation is the summation of the acquired doses

of convective heat and radiant heat. When the FED reaches unity,

incapacitation is assumed to occur. At an FED heat value of 1.0, many

evacuees are likely to be overcome by the combined effects of convective

heat and radiant heat. Temperature more than 60°C and radiant heat flux

more than 2.5 kW/m2 are required consideration for the effects of convective

heat and radiant heat (Purser, 2002). The combined effects of convective heat

126

and radiant heat should not exceed one-tenth of the exposure (similar to

asphyxiant).

The FED safe criterion of one-tenth of the exposure is used in the evaluation of

ASET. The criterion of FED incapacitation is not used to allow uncertainty relating

to parameters in fire models (for example, variable fire load and intensity and

building spatial environment) and the effect of irritants.

Parameters: To determine the time for occupants to receive incapacitating doses of

smoke, asphyxiant toxic gases and heat exposure, the following parameters are

determined at each discrete increment of time:

• smoke in the lift lobby, lift shaft and stair shaft

• asphyxiant toxic gases in the lift lobby, lift shaft and stair shaft

• temperature and radiant heat flux in the lift lobby, lift shaft and stair shaft

In addition, temperature in LMR is also determined. Temperature may cause

malfunctioning of lifts due to which evacuees may be trapped in lift cabins and be

exposed to hot and toxic gases.

Smoke

Fractional effective dose of smoke visibility is measured from extinction coefficient.

The results from FDS modelling give the values of extinction coefficient Cs for

direct measurement. The FED is calculated from the following expression:

5.0

ssmoke

CFED = 5.1

Asphyxiant Toxic Gases

Asphyxiant toxic gas carbon monoxide is present to some extent in all fires.

Fractional incapacitating dose of asphyxiant gas carbon monoxide (CO) is

determined. Fractional incapacitating dose (FID) refers to the state of physical

127

inability to accomplish a specific task. FIDCO is calculated using the following

expression (Purser, 2002):

30

102925.8 036.14 ppmCOFIDCO

××=

−

5.2

Carbon dioxide (CO2) is itself asphyxiant and increases the rate of uptake of other

toxic gases (hyperventilation). Where the concentration exceeds 2%, the total

fractional effective dose (FED) for asphyxiant at each time increment is multiplied

by a factor, to allow for the increased rate of asphyxiant uptake due to

hyperventilation (Purser, 2002). Therefore, concentration of CO2 is expressed in

terms of multiplication factor VCO2 using the following expression (Purser, 2002):

=5

%exp 2

2

COVCO 5.3

Oxygen (O2) is a required consideration for the effects of oxygen vitiation lower than

13% (Purser, 2002). The fractional incapacitating dose of low-oxygen hypoxia

FIDO2 is calculated using the following expression (Purser, 2002):

( )[ ]2%9.2054.013.8exp

12 O

FIDo −−= 5.4

A combined FED is calculated from the associated risks of asphyxiant gas CO, CO2

and O2. The concentrations are determined at an interval of one minute and the

cumulative effects are determined to cause incapacitation. The fractional effective

dose of incapacitation FEDasphyxiant is calculated using the following expression

(Purser, 2002):

( )22 OCOasphyxiant FIDVCOFIDFED +×= 5.5

128

The effects of other asphyxiant toxicants and irritant gases are not considered.

Heat

The evacuees may be exposed to hot environment. Purser (2002) proposes the

following two expressions for the relationship between the time to exceed tenability

limits for radiant heat and convective heat:

33.1

80

qt Irad = 5.6

where q is the radiant heat flux (kW/m2)

4.37105 −××= Tt Iconv 5.7

where T is the air temperature (°C).

The temperature and radiant heat flux are determined at a time interval and

cumulative effects are determined to cause incapacitation. The fractional effective

dose for heat is calculated by the following expression (Purser, 2002):

ttt

FEDt

t IconvIrad

heat ∆

+=∑

2

1

11 5.8

where ∆t is the time increment between t1 and t2 (minute)

129

5.2.2 Safety Index

If the probability of occupant evacuation is higher for a place of safety, the

probability of the hazardous exposure would be lower. The risk is expressed in terms

of probabilities for unsuccessful evacuation. A First Order Second Moment (FOSM)

method is used to determine the lift evacuation safety for evacuees in apartment

buildings. In the FOSM method, safety is calculated using the first two statistic

moments of the parameters viz. mean and standard deviation and by a first order

linearization of the limit state equation (hence termed as FOSM). The safety index

β quantifies the safety associated with the evacuation system during emergencies.

The safety index β is a measure of the uncertainty in the output variables and can be

used to estimate the probability that the escape time will exceed the available time.

This approach has been used by many researchers (Frantzich, 1997a, Frantzich,

1997b, Frantzich et al., 1997, Magnusson, 1997, Hasofer and Beck, 2000).

The evacuation is described by the escape time margin for the last person leaving the

threatened area. The escape time margin, expressed by the limit state function, is the

difference between the available safe evacuation time (ASET) and the required safe

evacuation time (RSET). By subjecting some variables, in the limit state function, to

uncertainty the safety inherent in the function can be determined. The escape time

margin is calculated as M = S – L where S and L are independent stochastic

parameters. The parameter S is interpreted as a strength variable (ASET) and L as a

load variable (RSET). The escape time margin is described by means and standard

deviations. The system is functional if the margin is positive i.e. the strength is

higher than the load. The mean and standard deviation of the margin can be

described as:

LSM µµµ −= and 22

LSM σσσ += 5.9

The safety index (Cornell, 1969) is calculated by:

130

M

M

σµ

β = 5.10

If the parameters S and L are normally distributed, the margin M will also be

normally distributed. The values of safety indices 0, 1, 2 and 3 are roughly equivalent

to 50, 15, 2, and 0.1% probability of failure respectively (Frantzich, 1997c). The

discrete event simulation model (ARENA) and the field model (FDS) are used in

calculating the safety index (see Figure 5-1).

Figure 5-1 – Flow Diagram for Calculating Safety Index

ARENA has already been discussed and the relevant results and analysis of load

variables (RSET) are given in Chapter 4. The strength variables (ASET) are

determined using fire hazard models (FDS models). The calculations are performed

on three concept design options viz. unprotected lift lobby, protected lift lobby (with

additional evacuation strategy for one-fourth population) and double protected lift

lobby. Safety indices are determined for both lift and stair systems.

Determine strength variables (ASET) for lift

and stair systems

Determine load variables (RSET) for

stair system

Safety Index

‘β’

Establish parameters

for FDS simulation

Determine load variables (RSET) for

lift system

Concept designs for

FDS model

Analysis of Safety Indices

Analysis of FDS output variables

Analysis of load variables (lift system)

Analysis of load variables (stair system)

131

5.2.3 Safety Index for Three Locations

Lift car follows the discrete mode of transporting the evacuees, serves the floors and

leaves a few evacuees behind. These evacuees may be incapacitated due to exposure

to fire effluents in the lift lobby. The evacuees may also encounter toxic smoke

products in the lift and stair shafts. It is important to consider the effects of fire at

three locations:

• Lift lobby – when the evacuees are waiting for lifts on the floor. Strength

variable (ASET) is determined for the location of lift lobby. Lift pre-

evacuation time tLPE is considered for load variable L1 (RSET) i.e.

t L LPE=1 5.11

and tLPE is given by Eq. 4.3

• Lift shaft – when the evacuees are in the lift shaft. Strength variable (ASET)

is determined for the location of lift shaft. Lift evacuation time tLE is

considered for load variable L2 (RSET) i.e.

tLLE

=2 5.12

and tLE is given by Eqs. 4.1 and 4.4

• Stair shaft – when the evacuees are in the stair shaft. Strength variable

(ASET) is determined for the location of stairs. Stair evacuation time tSE is

considered for load variable L3 (RSET) i.e.

tLSE

=3 5.13

and tSE is given by Eqs. 4.5 and 4.7

132

Figure 5-2 illustrates the difference between the load variables of lift lobby and lift

shaft.

Figure 5-2 – Load Variables for Safety Index for the locations of Lift Lobby, Lift

Shaft and Stair Shaft

Load variable of stair shaft is similar to that of lift shaft. Lift pre-evacuation time and

lift evacuation time are used to calculate the load variables for lift lobby and lift

shaft, respectively. Stair evacuation time is used to calculate the load variables in the

stair system. With this approach, safety index β is determined for these three

133

locations. The average safety index for the lift system is used in the comparison with

that for the stair system. The corridor is common for the lift and stair systems and

the load variables have equal effects on both of them. The load variables in the

corridor have little significance when compared with the lift and stair systems.

Hence, load variables in the corridor are not determined.

5.2.4 Field Model ‘FDS’

In order to evaluate the fire and smoke spread in the buildings, the field model ‘Fire

Dynamics Simulator’ Version 4.07 (McGrattan et al., 2004) was used. This model

solves numerically a form of the Navier-Stokes equation appropriate for low-speed,

thermally driven smoke and heat transfer from fire. Radiant heat transfer is included

via the solution of the radiation transport equation for a non-scattering gray gas. All

solid surfaces are assigned thermal boundary conditions and information about the

burning behaviour of the material. Material properties are stored in the database and

used in the simulation. Heat and mass transfer to and from solid surfaces is usually

handled with empirical correlations. This model is used to determine the flow of hot

gases in the stair shafts, lift shafts and lift machine room; thus it predicts the

conditions in the buildings.

5.3 Model Framework and Variables

By specifying the variability in the location of fire, wind speed and

compartmentation, the fire hazards in the lift system are determined in terms of

probability of the output variables for the safety indexes. The parameters describing

the variables are chosen according to judgment based on experiments and statistics.

Each distribution is then described with its mean and standard deviation.

5.3.1 Hypothetical Building Model

The hypothetical building shown in Chapter 4 is considered for detailed analysis.

Figure 5-3 shows the typical floor of the hypothetical building used in the FDS

model. Dark lines in the model show the framework of the FDS model to avoid long

computational time. All the fire safety features are assumed to be compliant with the

134

building code. The horizontal and vertical separations are provided between fire-

affected unit and other dwelling units with fire rated construction, so that fire size is

confined to fire-affected unit only. The distance between the fire-affected unit door

and the lift door is equal to the distance between the fire-affected unit door and the

stair door. Therefore the effects of fire effluents can be equally distributed.

Figure 5-3 – Typical Floor of a Hypothetical Building for a Generalised Fire

Scenario (dimensions not to scale)

Due to faulty sprinklers, sprinkler controlled fire is not considered in the analysis. A

battery operated smoke alarm (AS 3786) is provided in each apartment. Smoke

detectors (AS 1670) are provided on each level in the public areas such as public

corridor, staircase and utility ducts and services (e.g. electric shaft, communication,

air handling equipment enclosure). The sounders of the smoke detectors are located

in the public corridors in the building. The building is analyzed with the following

assumptions and justifications in the FDS model:

• Fires have equal probability of occurrence on all floors. The generic

building studied was a residential building and have identical floor layout

throughout. In order to investigate the influences of stack effect and wind

speed on the spread of smoke and hot gases in the building, three typical

vertical fire locations viz. the lower level (2nd floor), the middle level (19th

floor) and the top most level (38th floor) in the building were selected. Fire

characteristics are given in Appendix J. The location of fire was not varied

135

horizontally as this will not significantly reveal the impact of stack effect

and wind effect.

• Inadvertent opening of the door of the fire-affected unit to assume a

reasonably worse fire scenario. The apartment doors are normally of self-

closing type, however, in determining the risks, the door is assumed opened,

which may occur under pre-flashover condition, flashover condition or door

closure mechanism failure.

• The compartmentation failure of the fire-affected unit does not occur during

the fire simulation except inadvertent opening of the door as stated above.

All the units on the fire-affected floor are in compliant to BCA for the fire

resistance rating of wall, floor and ceiling.

• The doors leading to the protected lift lobby (or stairs) are assumed to be

closed. The doors are maintained through an effective and graded program

of fire prevention inspections. They will be opened only during the occupant

movement, letting smoke into the lift lobby (or stairs). Protected lift lobby

doors are modelled for the movement of entire floor population (door

opening/closing by 32 occupants), whereas stair doors are modeled for half

the floor population (door opening/closing by 16 occupants as there are two

stairs).

• No automatic or manual fire control or suppression system is considered, or

these systems are assumed to be non-operational. This assumption leads to a

worst credible scenario that prompted the need to evacuate the building.

Other fire protection measures such as smoke alarm in the fire-affected unit

and smoke detector in the public corridor are considered in the risk analysis.

The occupants may not be subjected to high risk in sprinkler protected apartment

buildings, but the possibility of building evacuation can not be ruled out. Sprinklers

do not eliminate the possibility of a fire producing large volume of smoke.

Sprinklers are also not entirely 100% reliable at all times. Therefore, sprinklers are

not considered. The evacuees in the lift cabins may be exposed to smoke, toxic gases

and heat. The FDS model does not consider the wall of lift cabin in the lift shafts, as

the lift cabin would be in motion (it can be modelled only in a stationary lift). This

model does not also consider stairs in order to avoid the specifications of

136

complicated three dimensional obstructions inside the stair and to reduce

computation time. Such a simplification is believed to err on the conservative side in

the risk assessment and the impact on the outcomes of the results is small.

5.3.2 Concept Designs

Methods to estimate the probability of failure and associated risks of the unprotected

lift lobby, protected lift lobby and double protected lift lobby are examined for the

hypothetical building (see Chapter 3, Figure 3-9 and Chapter 4, Section 4.5.1). The

pressurisation of the lift lobby and stairs is not considered in FDS modelling. Either

the smoke lobby or the pressurisation (without smoke lobby) is acceptable under the

provision of codes and regulations (see Chapter 3, Section 3.1.4). The concept design

C incorporates the smoke lobby. Stairs in all the concept designs incorporate the

smoke lobby.

5.3.3 FDS Model Boundary Conditions

The Building Code of Australia has estimated that peak Heat Release Rate (HRR) for

apartment building is 5 MW for un-sprinklered and 1.5 MW for sprinklered building

(ABCB, 2005). Therefore a peak HRR of 5 MW is used in the analysis for un-

sprinklered fire. The fire specified in the FDS model follows medium t-squared curve

fire Q= αt2, where the fire growth rate α is the fire growth coefficient (0. 01172

kW/s2)} to a constant peak value (depending on the ventilation conditions – with and

without wind). The fire was assumed to be wood, typically found in kitchen

(cupboards) and drawing rooms of apartment buildings. The size of the fire is 2.5 m

× 2.0 m × 0.6 m. The outside environment temperature and building temperature

were 20°C. The combustion yields, HRR output and temperature output are given in

Appendix J.

The limited number of random variables (varying location of fire floor, wind speed

and compartmentation) affecting the spread of smoke and fire in lift shafts are

subjected to uncertainty in fire scenarios. Therefore the building is analyzed with the

following effects:

137

• Stack effect – influence the spread of smoke and toxic gases due to

vertically varying location of fire

• Wind effect – influence the spread of smoke and toxic gases in the lift

evacuation route

Three heights were considered for varying stack effect and wind effect analysis viz.

fire at 2nd floor, wind at 19th floor and wind at 38th floor. Other layout parameters of

the building are the same. The fire floor is considered with a view to create

reasonably worse case fire scenarios in both the evacuation routes (lifts and stairs).

Reasonably worse case fire scenario is essentially a deteriorating fire condition

arising from inadvertent opening of unit doors and failure of sprinklers. The building

model is prepared for 38 storeys and therefore wind velocities are determined at

various building heights using the following formula (Dalgliesh and Boyd, 1962):

k

r

rhh

hVV

= 5.14

where

h is the height (m)

hr is the reference height (m)

Vh is the mean wind speed at height h above the ground (m/s)

Vr is the mean speed at the reference height hr above the ground (m/s)

k is the exponent for the best-fitting curve

A reference height of 10 m is internationally recommended as the standard, and

exponent for mean wind speed is taken as 0.5 for the urban areas. For an arbitrarily

selected gradient wind of 2.22 m/s (8 kmph), the mean wind speed for the selected

building is calculated from 1.4 m/s (5.05 kmph) to 7.43 m/s (26.77 kmph) depending

upon the height of the building (see Figure 5-4).

138

Figure 5-4 – Wind Speed Profile

The wind speed is given as a ramp function in the FDS model and the wind effect is

observed after breaking of window glass. The wind direction is perpendicular to the

building window. The wind speed could not increase beyond 7.43 m/s as the model

crashes frequently due to the use of different grid sizes (see Chapter 9, Section 9.3).

5.3.4 Fire Simulation Scenarios

Table 5-1 gives the concept designs and the values of variables (vertical floor

location and wind speed) used in FDS models for 24 fire simulation scenarios. In

each concept design, there are 6 fire scenarios under the variable conditions of

vertical location and wind speed. Limited number of fire scenarios are selected for

exploring the issues of fire hazard (stack effect and wind effect) on the lift

evacuation system (see Chapter 3, Section 3.1.1). However, the fire scenarios can be

extended to variable horizontal locations and functionality of fire protection systems.

4 m

55 m

112 m 5.21 m/s H

eight (Z

-axis

)

Wind Speed

1.4 m/s

7.43 m/s

Y-axis

Z-axis

Lift shaft

139

Table 5-1: Description of Fire Simulation Scenarios

Concept Design Fire Floor Level Fire Floor

Height (m)

Wind Speed

(m/s)

Fire Simulation

Scenario No.

0 1 Lower Level (2nd Floor)

3

1.4 2

0 3 Middle Level (19th Floor)

54

5.21 4

0 5

Concept Design ‘A’

Top level (38th Floor)

111

7.43 6

0 7 Lower Level (2nd Floor)

3

1.4 8

0 9 Middle Level (19th Floor)

54

5.21 10

0 11

Concept Design ‘B’

Top level (38th Floor)

111

7.43 12

0 13 Lower Level (2nd Floor)

3

1.4 14

0 15 Middle Level (19th Floor)

54

5.21 16

0 17

Concept Design ‘B’

(Evacuation for 25% population) Top level

(38th Floor) 111

7.43 18

0 19 Lower Level (2nd Floor)

3

1.4 20

0 21 Middle Level (19th Floor)

54

5.21 22

0 23

Concept Design ‘C’

Top level (38th Floor)

111

7.43 24

5.4 FDS Model Set Up

5.4.1 Conventional Domain and Grid System

A simplistic 38 storey structure incorporating relevant building features such as fire

compartment, corridor, lift shaft and stair has been considered since it is virtually

impossible to run a FDS model with a finer grid for the entire building. The

computational domain of the FDS for the building was set with main components of

the bounding walls and the roof top of the building. The computational domain of the

main building framework has a volumetric space of 57 m × 20 m × 114 m (see

Figure 5-5). The computational domain was extended beyond the physical boundary

of the building by 0.9 m (front and rear window sides). The space allows wind to

flow around the building. The front and rear boundaries were open to the outside.

140

Figure 5-5 – Computational Domain for FDS Model

FDS uses a Large Eddy Simulation (LES) model for computation and the grid size

allows the sub grid scale stress model to accurately compute the viscous stress of the

flow field. The characteristic length scale near the fire is mainly the characteristic

fire diameter D* (m) (Baum and McCaffery, 1989), as given below:

5/2*

*

=

∞∞ gTc

QD

pρ 5.15

where

Q* is the heat release rate (kW)

ρ∞ is the density at ambient temperature (kg/m3)

cp is the specific heat of gas (kJ/kg.K)

T∞ is the ambient temperature (K)

g is the acceleration due to gravity (m/s2)

141

Baum and McCaffery (1989) and Bounagui et al. (2004) recommended that grid

independence could be achieved at grid size of 0.1D* near the fire. For the fire size

of 5 MW, D* was calculated to be 1.11 m and so 0.1D* was about 0.11 m.

Therefore, the grid size was 0.1 m for the fire compartment. The grid size was 0.3 m

for the corridor, lift lobby and stair lobby, and 0.6 m for the lift shaft and stair shaft

(see Figure 5-6). The grid sizes of 0.3 m and 0.6 m were selected to reduce the

computational time.

Figure 5-6 – Three Grid Sizes used in the FDS Model (38th floor view)

5.4.2 Smoke Leakages/ Openings

Door opening and closing for lift lobbies and stair lobbies are incorporated in FDS

model for occupant’s movement as shown in Appendix K. The following openings

are considered in modelling:

• a permanent gap between the lift landing doors and the frames

• a temporary gap between the lift landing door and the frame at the time of lift

service (temporary gap operates at the time of lift service only)

• a temporary gap for the entire door width for the protected lift lobby at the

time of occupant movement

• a temporary gap for the entire door width for the protected stair lobby at the

time of occupant movement (first door for stair protection)

Stair shaft (0.6 m)

Fire compartment (0.1 m)

Lift shaft (0.6 m)

Public corridor (0.3 m)

Lift lobby

(0.3 m)

142

• a temporary gap for the entire door width for the protected stair shaft at the

time of occupant movement (second door for stair protection)

A maximum gap of 6.5 mm between lift landing door and frame is permitted (AS

1735.1, 2003) and the area calculated is 0.045 m2 per door. For two doors, the area

calculated is 0.09 m2. Therefore, permanent gap between the lift landing door and lift

frame is considered to be equivalent to a square opening of size 0.3 m × 0.3 m for

two lifts. Additional temporary gap between the lift landing door and the frame is

considered as a rectangular opening of size 0.3 m × 2.0 m since the lift cabin

provides an obstruction to smoke propagation from lobby to shaft. Hence the gap

between landing door and frame is considered to be 0.3 m of 2.0 m height (see

Figure 5-7).

Figure 5-7 – Smoke Leakage Openings in the Lift Shaft Wall

The doors protecting the lift and stair lobbies are considered to be 0.9 m. The timings

for door openings and closings were determined from ARENA simulation model,

which were further added with 3 seconds. Generally, it is observed that the duration

of door opening/closing (or door swing) is 5 seconds, whereas effective opening of

full door width is considered for 3 seconds.

5.5 FDS Output

The output points were placed in the lift lobby, lift shaft, LMR and stair shaft to

record the predicted extinction coefficient, species concentrations of CO, CO2, O2,

temperature and radiant heat flux (see Figure 5-8).

143

Figure 5-8 – Output points in the Fire Compartment, the Lift Lobby, the Lift Shaft

and the Stair

The simulation results from the output points located 1.8 m above the floor level in

the lift lobby and 3.0 m above the floor level in the lift/stair shafts are presented to

highlight the visibility, smoke and toxic gases and temperature hazards on the

evacuees. The output points are placed 3.0 m above the floor level in the stair shafts

as travellers from upper level would be coming down in the stair shaft. The output

was located 1.0 m above the floor level in LMR to highlight the temperature rise to

electronic components. Two detectors (one in fire compartment and another in

public corridor) were placed near the ceiling.

Snap shots of FDS output were taken for the 2nd

, 19th and 38

th floors. The snapshots

of smoke view and PLOT3D temperature contour for the 2nd

floor are shown in

Figure 5-9 (a) and (b). The smoke view indicates the conditions of lift shaft at 600

seconds, which is quickly filled up with smoke and hot gases.

144

(a) Smoke in the Building at 600 seconds (top view)

(b) Temperature Contour in the Corridor at 720 seconds (side view)

Figure 5-9 – Snapshots of Smoke View and Temperature Contour (Fire Scenario 1)

145

The snapshots of PLOT3D temperature contour for the 19th floor are shown in Figure

5-10 (a) and (b). Temperature in the corridor is closer to 60°C at 720 seconds at 1.8

m above the floor level while it is closer to 350°C at 1200 seconds in the lift shaft,

which arises due to the consumption of oxygen contents (yellowish-orange plume).

(a) Temperature Contour in the Corridor at 720 seconds

(b) Temperature Vector Slice in the Lift Lobby and Shaft at 1200 seconds

Figure 5-10 – Snapshots of Temperature Contour and Vector slice (Fire Scenario 3)

146

The snapshot of smoke view for the 38th floor of Fire Scenario 5 is shown in Figure

5-11. The snapshot indicates smoke spread in the unprotected lift lobby. Tenability

limits exceed quickly in the unprotected lift lobby. The visibility diminishes to 4 m in

the lift lobby at 241 seconds.

Figure 5-11 – Slice Snapshot of Visibility in the Lift Lobby (Fire Scenario 5)

147

The snapshot of PLOT3D temperature contour for the 38th floor is shown in Figure

5-12. Temperature is closer to 60°C at 1.8 m above the floor level in the corridor.

Figure 5-12 – Snapshot of Temperature Contour at 720 seconds (Fire Scenario 5)

The snapshots of temperature, radiant heat flux, CO2, CO, visibility, extinction

coefficient through lift landing door frame gap (square opening) for the 38th floor of

Fire Scenario 5 are shown in Figure 5-13 (a) to (h). Temperature and CO2 are closer

to the tenability limit criteria. Extinction coefficient is also more than the tenability

limit.

148

(a) Temperature Vector Slice through Lift Door (b) Temperature

(c) Vector Slice of Radiant Heat Flux (d) Radiant Heat Flux

(e) CO2 Concentration (f) CO Concentration

(g) Visibility (h) Extinction Coefficient

Figure 5-13 – Slice Snapshots in a Vertical Plane in the Lift Lobby at 600 seconds

(Fire Scenario 5)

149

5.6 FDS Results

The FDS model was used to calculate the times to exceed the tenability limits in the

lift and stair systems. The corridor may become untenable due to which evacuees

may not be in a position to use lifts or stairs. In such circumstances, the strategy ‘stay

in place’ may be adopted, which is another area for research. This research was

confined to explore the feasibility of using lifts on a comparative basis. It determined

that the smoke alarm operates at 90 seconds in the fire compartment while the smoke

detector operates at 140 seconds in the public corridor. FDS results are given for Fire

Scenarios 1 to 6 in Section 5.6.1 followed by analysis. The results for Fire Scenarios

7 to 24 are given in Appendix M.

5.6.1 Concept Design A (Unprotected Lift Lobby)

Figure 5-14 gives the results of time vs. extinction coefficient, concentrations of CO,

CO2 and O2, temperature and radiant heat flux in lift lobby and lift shaft for Fire

Scenarios 1 to 6. Temperature in the lift machine room (LMR) is also shown.

Fire Scenario 1 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

150

Fire Scenario 1 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

Fire Scenario 1 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

151

Fire Scenario 1 (Temperature and Radiant Heat Flux in Lift Shaft)

0

50

100

150

200

250

300

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux (kW/m2)

Temperature Radiant heat flux

Fire Scenario 1 (Temperature in LMR)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

Temperature

152

Fire Scenario 2 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 2 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

153

Fire Scenario 2 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scneario 2 (Temperature and Radiant Heat Flux in Lift Shaft)

0

50

100

150

200

250

300

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux (kW/m2)

Temperature Radiant heat flux

154

Fire Scenario 2 (Temperature in LMR)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

Temperature

Fire Scenario 3 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

155

Fire Scenario 3 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (c)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

Fire Scenario 3 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

156

Fire Scenario 3 (Temperature and Radiant Heat Flux in Lift

Shaft)

0

50

100

150

200

250

300

0360

720

1080

1440

1800

2160

2520

2880

3240

3600

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

Fire Scenario 3 (Temperature in LMR)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

Temperature in LMR

157

Fire Scenario 4 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 4 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

158

Fire Scenario 4 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 4 (Temperature and Radiant Heat Flux in Lift

Shaft)

0

50

100

150

200

250

300

0360

720

1080

1440

1800

2160

2520

2880

3240

3600

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

159

Fire Scenario 4 (Temperture in LMR)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

Temperature

Fire Scenario 5 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

160

Fire Scenario 5 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

Fire Scenario 5 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

161

Fire Scenario 5 (Temperature and Radiant Heat Flux in Lift Shaft)

0

50

100

150

200

250

300

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux

Temperature Radiant heat flux

Fire Scenario 5 (Temperature in LMR)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

Temperature

162

Fire Scenario 6 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800

Time (second)

Extinction coefficient (1/m

)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 6 (Temperature and Radiant Heat Flux in Lift Lobby)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800

Time (second)

Temperature (C)

0

1

2

3

4

5

6

7

8

9

10

Radiant heat flux (kW/m2)

Temperature Radiant heat flux

163

Fire Scenario 6 (Smoke and Gases in Lift Shaft)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800

Time (second)

Extinction coefficient (1/m)

and CO (hundred ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 6 (Temperature and Radiant Heat Flux in Lift

Shaft)

0

50

100

150

200

250

300

0 360 720 1080 1440 1800

Time (second)

Temperature (C)

0

5

10

15

20

25

30

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

164

Fire Scenario 6 (Temperature in LMR)

0

20

40

60

80

100

120

140

160

0 360 720 1080 1440 1800

Time (second)

Temperature (C)

Temperature

(Simulation crashed after 1800 seconds under the influence of wind)

Figure 5-14 – Smoke, Gases and Heat in the Lift Lobby, the Lift Shaft and the LMR

(Fire Scenarios 1 to 6)

165

5.6.2 FDS Results Analysis

Concept Design A: Figure 5-14 highlighted the danger of smoke and toxic gases in

the unprotected lift lobby under variable conditions of wind and vertical location.

The concentration of asphyxiant gas CO has increased to about 18 500 ppm while

that of CO2 has increased from small traces to about 16%. The concentration of

oxygen reduced from 20.72% to zero. The extinction coefficient was nearly 40 m-1

(tenability limit is 0.5 m-1 for a visibility distance of 4 m).

Unprotected Lift Lobby: During the fire occurrence on the 38th floor, wind increases

the temperature in the lift lobby, but dilutes the concentration of fire effluents (see

Figure 5-15 (a) and (b)). Time to exceed tenability limit for temperature is decreased

while time to exceed tenability limit for asphyxiant is increased.

38th Floor Fire (temperature with and without wind)

0

20

40

60

80

100

120

140

160

0 500 1000 1500

Time (second)

Temperature (C)

Fire Scenario 5 Fire Scenario 6

(a) Temperature on the 38th floor lift lobby

{Fire Scenario 5 (without wind) and Fire Scenario 6 (with wind)}

166

38th Floor Fire (CO with and without wind)

0

2

4

6

8

10

12

14

16

18

20

0 500 1000 1500

Time (second)

CO (hundred ppm)

Fire Scenario 5 Fire Scenario 6

6

6

(b) CO on the 38th floor lift lobby

{Fire Scenario 5 (without wind) and Fire Scenario 6 (with wind)}

Figure 5-15 – Temperature and CO on the 38th Floor for Unprotected Lift Lobby

Temperature in the lift lobby is the function of opening and closing of lift landing

door. The opening and closing of lift landing door caused a flow of hot gases to lift

shaft. Minor dips are seen in the temperature curve at the times of lift landing door

opening (see temperature and heat flux curves in Figure 5-14). Temperature

increases if lifts are not served.

Unprotected Lift Shaft: The maximum temperature in the lift shaft during the

simulation period is 280°C for the fire occurrence on the 2nd floor, 263°C for the fire

occurrence on the 19th floor and 167°C for the fire occurrence on the 38

th floor (see

Figure 5-16 a). Temperature and radiant heat flux curves show spikes in the lift shaft

due to turbulence (whereas temperature and radiant heat flux curves are smooth in

the lift lobby). The temperature is slightly diluted under wind conditions on the

upper levels (due to ventilation from window in LMR). The maximum temperature

in the lift shaft during the simulation period is 280°C during the fire occurrence on

the 2nd floor, 209°C during the fire occurrence on the 19th floor and 163°C during the

fire occurrence on the 38th floor (see Figure 5-16 b). The concentration of oxygen is

comparatively improved with wind on upper levels.

167

Lift Shaft Temperature (without wind)

0

50

100

150

200

250

300

0 1000 2000 3000 4000

Time (second)

Temperature (C)

2nd floor fire 19th floor fire 38th floor fire

(a) Temperature in the Lift Shaft (without wind)

Lift Shaft Temperature (wind)

0

50

100

150

200

250

300

0 1000 2000 3000 4000

Time (second)

Temperature (C)

2nd floor fire 19th floor fire 38th floor fire

(b) Temperature in the Lift Shaft (with wind)

Figure 5-16 – Temperature in the Lift Shaft (with and without wind)

Unprotected LMR: The maximum limit of temperature is reached quickly in the

LMR. The temperature in the long shafts is also diluted with height. The maximum

temperature is 59°C for the fire occurring on the 2nd floor, 140°C for the fire

occurring on the 19th floor and 140°C for the fire occurring on the 38

th floor in the

early stages (see Figure 5-17). The safe limit is 43°C. There is significant increase

168

in the LMR temperature due to the fire source being closer to LMR. Major dips (ups

and downs) in Fire Scenario 5 (38th floor fire) are caused by the lift landing door

opening near LMR (see Figure 5-17).

LMR Temperature (without wind)

0

20

40

60

80

100

120

140

160

0 1000 2000 3000 4000

Time (second)

Temperature (C)

2nd floor fire 19th floor fire 38th floor fire

Figure 5-17 – Temperature in the LMR (without wind)

Concept Design B – Protected Lift Lobby, Shaft and Stairs: Figure M1

(Appendix M) shows the concentrations of smoke, toxic gases and temperature in

protected lift lobby, lift shaft and stairs under variable conditions of wind and

vertical location (Fire Scenarios 7 to 12). The concentration of asphyxiant gas CO is

about 2000 ppm. The concentration of CO2 is less than 2%. The concentration of

oxygen is not much changed. However, extinction coefficient is nearly 4 m-1. Stack

effect contributes to the spread of smoke, toxic gases, temperature and radiant heat

flux to the shaft. Wind speed also contributes to the spread of smoke, toxic gases,

temperature and radiant heat flux to the lift lobby. However, the spread of fire

effluents is reduced considerably with the provision of protected lift lobby.

Concept Design B with 25% population – Protected Lift Lobby, Shaft and

Stairs: Figure M2 (Appendix M) shows the concentrations of fire effluents under

variable conditions (Fire Scenarios 13 to 18). The concentration of asphyxiant gas

CO is less than 100 ppm. The concentration of CO2 is less than 1%. The

concentration of oxygen is also not affected. Only the extinction coefficient for

169

visibility is more than the tenability limit. Temperature is not increased much in the

lift lobby and thus no appreciable temperature rise in the lift shaft.

Concept Design C – Double Protected Lift Lobby and Shaft: Figure M3

(Appendix M) shows that the concentrations of smoke, toxic gases and temperature

in the double protected lift lobby and lift shaft (Fire Scenarios 19 to 24). The

concentration of asphyxiant gas CO is less than 100 ppm. Extinction coefficient is

increased under wind speed. Due to double protection, there is a very little change in

the lift lobby temperature.

Summary

• Time to exceed tenability limit for visibility is influenced by wind speed.

• Smoke and toxic gases stay longer in the protected lift lobby (confined

location) and are dissipated only at the time of lift service.

• There is a significant temperature increase in the unprotected lift shaft due to

‘stack effect’.

• There is a significant temperature increase in the LMR due to ‘stack effect’.

• Temperature in the lift lobby is a function of lift landing door opening and

closing. Radiant heat flux is less than 2.5 kW/m2 in the protected lift lobby.

170

5.7 FED of Smoke, Asphyxiant and Heat

Fractional Effective Dose (FED) for smoke, asphyxiant and heat are determined from

Fire Scenarios 1 to 24. Equations 5.1 to 5.5 are used for calculating the FED in lift

lobby, lift shaft and stair shaft. The calculation is given for 40 minutes, during which

the entire building can be evacuated by lifts or stairs. FED results are given for Fire

Scenarios 1 to 6 in the next section followed by analysis. Appendix N gives the

calculation for Fire Scenario 1. The results for Fire Scenarios 7 to 24 are given in

Appendix M.

5.7.1 Concept Design A (Unprotected Lift Lobby)

Figure 5-18 shows the representations for FED smoke, asphyxiant gases and heat in

Fire Scenarios 1 to 6.

Fire Scenario 1 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

(The representation shows FED 5; incapacitation occurs at 1)

171

Fire Scenario 1 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

Fire Scenario 2 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant and

FED heat

Smoke Asphyxiant Heat

172

Fire Scenario 2 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant and

FED heat

Smoke Asphyxiant Heat

Fire Scenario 3 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant and

FED heat

Smoke Asphyxiant Heat

173

Fire Scenario 3 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant and

FED heat

Smoke Asphyxiant Heat

Fire Scenario 4 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

174

Fire Scenario 4 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

Fire Scenario 5 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

175

Fire Scenario 5 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

Fire Scenario 6 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30

Time (minute)

FED smoke, FED asphyxiant and

FED heat

Smoke Asphyxiant Heat

176

Fire Scenario 6 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30

Time (minute)

FED smoke, FED asphyxiant

and FED heat

Smoke Asphyxiant Heat

Figure 5-18 – FED in Fire Scenarios 1 to 6

177

5.7.2 Summary of FED Results and Analysis

Tables 5-2 and 5-3 give the times to exceed tenability limits in the lift lobby and lift

shaft for smoke obscuration, asphyxiant toxic gases and heat. These times based on

FED results are given in seconds. Table 5-3 also gives the times to exceed the

tenability limit for temperature in the LMR.

Table 5-2: Time to Exceed Tenability Limits in Lift Lobby

Time to Exceed Tenability Limits (second)

Asphyxiant Gases Heat

Building Fire

Scenario Smoke

Obscuration

Limit Tenability

Limit

Lethal

Dose

Tenability

Limit

Lethal

Exposure

1 241 600 900 720 1260

2 241 600 840 720 1320

3 241 600 900 720 1080

4 241 600 840 480 840

5 241 600 900 720 1140

6

Conce

pt D

esig

n A

241 660 900 480 720

7 403 2040 N.E. N.E. N.E.

8 400 2040 N.E. N.E. N.E.

9 403 2040 N.E. N.E. N.E.

10 353 1740 N.E. N.E. N.E.

11 404 1980 N.E. N.E. N.E.

12

Conce

pt D

esig

n B

367 2100 N.E. N.E. N.E.

13 508 N.E. N.E. N.E. N.E.

14 508 N.E. N.E. N.E. N.E.

15 511 N.E. N.E. N.E. N.E.

16 446 N.E. N.E. N.E. N.E.

17 529 N.E. N.E. N.E. N.E.

18 Conce

pt D

esig

n B

(25%

)

501 N.E. N.E. N.E. N.E.

19 N.E. N.E. N.E. N.E. N.E.

20 N.E. N.E. N.E. N.E. N.E.

21 N.E. N.E. N.E. N.E. N.E.

22 554 N.E. N.E. N.E. N.E.

23 N.E. N.E. N.E. N.E. N.E.

24

Conce

pt D

esig

n C

518 N.E. N.E. N.E. N.E.

178

‘N.E.’ denotes ‘Not Exceeded’ during FED calculation period. Table 5-1 may be

referred for the definitions of fire scenarios.

Table 5-3: Time to Exceed Tenability Limits in Lift Shaft and LMR

Time to Exceed (second)

Asphyxiant Gases Heat

Building

Fire

Scenario Smoke

Obscuration

Limit Tenability

Limit

Lethal

Dose

Tenability

Limit

Lethal

Exposure

Temperature

Limit in

LMR

1 317 1140 1500 1140 1260 2473

2 310 1020 1500 1080 1200 2485

3 327 1020 1560 1140 1320 1290

4 306 900 1440 900 1140 1022

5 303 840 1140 1080 1680 854

6

Conce

pt D

esig

n A

306 780 1080 780 1380 522

7 N.E. N.E. N.E. N.E. N.E. N.E.

8 N.E. N.E. N.E. N.E. N.E. N.E.

9 N.E. N.E. N.E. N.E. N.E. N.E.

10 475 N.E. N.E. N.E. N.E. N.E.

11 N.E. N.E. N.E. N.E. N.E. N.E.

12

Conce

pt D

esig

n B

493 N.E. N.E. N.E. N.E. N.E.

13 N.E. N.E. N.E. N.E. N.E. N.E.

14 N.E. N.E. N.E. N.E. N.E. N.E.

15 N.E. N.E. N.E. N.E. N.E. N.E.

16 N.E. N.E. N.E. N.E. N.E. N.E.

17 N.E. N.E. N.E. N.E. N.E. N.E.

18 Conce

pt D

esig

n B

(25%

)

N.E. N.E. N.E. N.E. N.E. N.E.

19 N.E. N.E. N.E. N.E. N.E. N.E.

20 N.E. N.E. N.E. N.E. N.E. N.E.

21 N.E. N.E. N.E. N.E. N.E. N.E.

22 N.E. N.E. N.E. N.E. N.E. N.E.

23 N.E. N.E. N.E. N.E. N.E. N.E.

24

Conce

pt D

esig

n C

N.E. N.E. N.E. N.E. N.E. N.E.

Time relating to LMR is used for determining the reliability of lift operational system

in the next chapter.

179

Table 5-4 gives the times to unsafe conditions in the stair shaft.

Table 5-4: Time to Exceed Tenability Limits in Stair Shaft

Time to Exceed Tenability Limits (second) Building Fire

Scenario Smoke Asphyxiant Gases Heat

1 or 7 554 N.E. N.E.

2 or 8 554 N.E. N.E.

3 or 9 532 N.E. N.E.

4 or 10 428 N.E. N.E.

5 or 11 530 N.E. N.E.

6 or 12

100%

popula

tion

497 N.E. N.E.

13 N.E. N.E. N.E.

14 558 N.E. N.E.

15 N.E. N.E. N.E.

16 508 N.E. N.E.

17 558 N.E. N.E.

18

75%

popula

tion

461 N.E. N.E.

Note: Fire Scenarios 1 and 7 are the same as they contain the protected stair lobby

and are considered for the entire population. (‘N.E.’ denotes Not Exceeded)

Summary of Results

• Tenability limits for smoke visibility, asphyxiant gases and heat are exceeded

in the unprotected lift lobby.

• Tenability limits for smoke visibility and asphyxiant gases are exceeded in

the protected lift lobby.

• Tenability limit for smoke visibility are exceeded in the protected lift lobby

with partial evacuation.

• Tenability limit for smoke visibility are exceeded in the double protected lift

lobby under the influence of wind.

• Tenability limit for smoke visibility are exceeded in stairs. However, smoke

visibility was improved during the limited number of occupants using stairs

(75% of the building population).

180

• Lift door is used by the entire population whereas stair door is used by half

the number of evacuees on the floor (as evacuees were evenly distributed at

two stairs). Therefore smoke spreads quickly in the lift lobby.

• Smoke logged in the protected lift lobby takes considerable time to dilute

(being in an isolated place with minor gap opening in lift shaft).

• Two level compartmentalization strategies in the double protected lift lobby

provide a buffer for smoke and hot gases leakage to lift lobby from the fire-

affected unit.

Analysis of Untenable Conditions: In the unprotected lift lobby, extinction

coefficient exceeds the tenability limit for visual obscuration and sensory irritation

during the fifth minute. The FED limit exceeds the combined effects of temperature

and heat during the period of eighth to twelfth minute. The FED limit exceeds the

combined effects of asphyxiant gases during the tenth minute. In protected lift

lobby, extinction coefficient exceeds the tenability limit for visual obscuration and

sensory irritation during the period of sixth and seventh minutes. The FED limit

exceeds the combined effects of temperature and heat under the influence of wind

during the period of twelfth to fourteen minutes. In the protected lift lobby for 25%

of the population, extinction coefficient exceeds the tenability limit for visual

obscuration and sensory irritation during the period of eighth and ninth minutes. In

the double protected lift lobby, extinction coefficient exceeds the tenability limit for

visual obscuration and sensory irritation under the influence of wind during the ninth

minute. In stairs, the extinction coefficient exceeds the tenability limit for visual

obscuration and sensory irritation during the period of eighth to tenth minutes.

Aged and disabled persons have a ‘high risk’ level than others (Miller, 2005), and

thus consideration needs to be taken into account. Therefore, the effects of FED will

be more in the ‘high risk’ group. Aged and disabled persons (16% of the population)

are exposed to fire effluents in the unprotected lift lobby (asphyxiant gases and

temperature/ heat) and protected lift lobby (asphyxiant gases). Aged and disabled

persons are not exposed to fire effluents in the protected lift lobby for 25% of the

population, double protected lift lobby and stairs.

181

5.8 Determination of Safety Index

The probability of time period for occupants’ evacuation must be greater than the

probability of time to exceed the tenability limits. Random variables were

determined from the stochastic evacuation modelling and fire hazard modelling. The

safety index β is used to determine the safety margin between the lift evacuation time

and onset of hazardous condition.

5.8.1 Strength Variables ASET

The mean and standard deviation of the strength variables (ASET) are obtained from

Tables 5-2 to 5-4. The period for which the tenability limits are not exceeded is the

safe period for evacuation, and the entire population will evacuate the building. The

means and standard deviations of the ASET for the locations of lift lobby, lift shaft

and stairs are given in Tables 5-5 to 5-7. These values are determined from the time

to exceed tenability limits for asphyxiant toxic gases or heat, whichever is reached

first. The ASET, for which tenable limits are not exceeded, is considered for the

FED calculation period (2400 seconds).

Table 5-5: Means and Standard Deviations of ASET (lift lobby)

Concept Design/

Evacuation strategy

Evacuation

Strategy Mean

(µs) Standard

Deviation (σs)

Concept Design A 100% population 610 24.5

100% population 1990 128 Concept Design B

25% population > 2400 -

Concept Design C 100% population > 2400 -

Table 5-6: Means and Standard Deviations of ASET (lift shaft)

Concept Design/

Evacuation strategy

Evacuation

Strategy

Mean

(µs) Standard

Deviation (σs)

Concept Design A 100% population 950 134

100% population > 2400 - Concept Design B

25% population > 2400 -

Concept Design C 100% population > 2400 -

182

Table 5-7: Means and Standard Deviations of ASET (stairs)

Concept Design/

Evacuation strategy

Evacuation

Strategy

Mean

(µs) Standard

Deviation (σs)

100% population > 2400 - Stairs

75% population > 2400 -

5.8.2 Load Variables RSET

The load variables RSET were obtained from the stochastic evacuation modelling

and results are reproduced from Chapter 4 (see Table 5-8). The time values are given

to the nearest second.

Table 5-8: Means and Standard Deviation of RSET

Output Variables Evacuation

Strategy Mean

(µL)

Standard

Deviation (σL)

100% population 1347 30 Lift Pre-Evacuation Time

(tLPM) – RSET in lift lobby

25% population 924 24

100% population 2288 490 Lift Evacuation Time

(tLE) – RSET in lift shaft 25% population 1624 303

100% population 1550 238 Stair Evacuation Time

(tSE)

75% population 1188 181

183

5.8.3 Safety Index

The margin of mean and standard deviation are calculated from Eq. 5.9 and the

safety indices are determined from Eq. 5.10 (see Table 5-9). The system is functional

if the margin is positive i.e. the strength is higher than the load. Where tenable limits

are not exceeded, safety indices are based on the maximum FED calculation period

(2400 seconds). The difference between ASET and RSET is more in the lift lobby

than in the lift shaft. Therefore, average safety indices are calculated to avoid

extreme values.

Table 5-9: Safety Indices for Lift and Stair Evacuation

Conceptual design

and evacuation

strategy

Location Mean

(µM) Standard

Deviation (σM)

Safety

Index

(M

M

σµ

β = )

Average

Safety

Index

(β)

Lift lobby -737 38.73 -19.02 Concept Design A

Lift shaft -1338 508 -2.63

-10.82

Lift lobby 643 131.46 4.89 Concept Design B

Lift shaft > 112 490 > 0.22

> 2.55

Lift lobby > 1476 24 > 61.5 Concept Design B (25% population)

Lift shaft > 776 303 > 2.5

> 32.0

Lift lobby > 1053 30 > 35 Concept Design C

Lift shaft > 112 490 > 0.22

> 17.61

Stair (100% population) > 850 238 > 3.57 > 3.57

Stair (75% population) > 1212 181.8 > 6.66 > 6.66

5.8.4 Safety Index Analysis

The safety index β was determined for the concept designs and evacuation strategies.

The safety index is positive in Concept Designs B and C whereas negative in

Concept Design A (the system is functional – the strength is higher than the load).

Protected lift lobby for one-fourth of the building population, double protected lift

lobby and stairs provided a life safety index of more than 3 to the evacuees. The

safety index is the maximum for partial building lift evacuation. Protected lift lobby

for the entire population had lower safety indices when compared to that of stairs.

Unsafe conditions arrive quickly in the unprotected lift lobby.

184

5.9 Conclusion

The safety index concept is based on probabilistic theory. On the other hand, the

computational fluid dynamics simulation using FDS is a deterministic approach. The

FDS predicted the time when the lift lobbies and stairs became unsafe under variable

conditions. The unsafe conditions within a hypothetical building were determined

with regard to temperature of the hot layer, concentration of toxic gases and

visibility. Two variables (vertical location and wind speed) were considered to

determine the impact on the three concept design options with additional evacuation

strategies (unprotected lift lobby, protected lift lobby for the entire population and

one-fourth population and double protected lift lobby)

Stack effect is significant in the unprotected lift lobby, but not in the protected lift

lobby for the spread of smoke and hot gases. An exponential increase in wind speed

with height demonstrated that the time to exceed tenability limits at upper levels

comparatively influenced more (compared to lower floors). Wind caused visibility

obscuration in protected and double protected lift lobbies. However, visibility could

be maintained for longer periods if the occupants use the doors less frequently (for

example, by 25% of the population). The temperature was significantly increased in

the unprotected LMR in the unprotected lobby scenarios due to stack effect and

wind, but was not much affected in protected lift lobby scenarios.

The load variables (RSET) and the strength variables (ASET) were used to calculate

the safety index. Unprotected lift lobby provided a negative safety index and exposed

to the evacuees to the maximum risks. Protected lift lobby, double protected lift

lobby and stairs provided a positive safety index to the evacuees. Time to exceed

tenability limits with regard to temperature and toxic gases did not arrive in the

protected lift lobby for one-fourth of the building population and in the double

protected lift lobby. The FED calculations showed that the times to exceed tenability

limits for visibility in lift lobbies were not significantly influenced by the protected

lift lobbies.

185

6. RELIABILITY OF LIFT OPERATIONAL MECHANISM

6.1 Introduction

During fire emergencies, many components may affect the reliability of lift

operational mechanism. While using lifts for emergency evacuation, the occupants

may inadvertently trapped in the lifts and be exposed to heat and smoke on the fire-

affected floor. The main issues that affect the reliability of lift systems are

temperature sensitive electronic components, electric power supply failure and close

proximity of water based fire fighting systems.

Lift operational mechanism makes use of temperature sensitive electronic devices.

The lift system can be affected by the malfunctioning of electronic devices due to

excessive temperature rise in fires. The electric power supply in the developed

countries is generally characterized as stable and reliable. However, this may not be

true for electric power supply in other countries of the world. The challenge of

providing a reliable electric power supply is made even more difficult in the presence

of fires, which makes electric systems more susceptible to power failures and shut

down the lift operation (Klote, 1982). Available statistics have shown that electrical

fault is one of the prevalent causes of fires in the buildings (NFPA, 2006). In fire

incidences that are initiated by electrical appliances, the electric power supply to lifts

will be a concern. The lift system may also be affected by water spread from fire

fighting system as it can damage the electrical and/or mechanical components of the

lift operating system (Klote, 1982).

This chapter addresses the issues related to reliability of lift operational mechanism

based on probabilistic risk assessment techniques. The techniques used include fault

tree and event tree analyses. The objectives of this chapter are to:

1. To determine the reliability of lift operational mechanism.

2. To determine the feasibility of reliability improvement.

186

6.2 Analysis of Lift Operational Mechanism

To determine the reliability of lift operational mechanism, the issues relating to

malfunctioning of lifts, electric power failure and water damage are analysed. The

objectives of analysis are to:

1. determine the probabilities of malfunctioning of lifts, electric power failure

and water damage; and

2. analyse the reliability of lift operational mechanism that can raise concerns of

human behavioural response (panic).

The reliability of lift operational mechanism is analyzed based on the following

assumptions:

• Water based fire extinguishment systems are not used on fire that occurred in

electrical systems.

• Apartment buildings are provided with secondary (alternate) sources of

power supply in the form of electric generators.

• Water based fire protective and fire fighting measures (for example, fire hose

reel, automatic sprinkler and hydrant system) are present. Fire extinguisher

constitutes only a tiny fraction of water spread and hence fire extinguishment

using fire extinguishers is not included in the water spread analysis. Manual

extinguishing facility is interpreted as fire hose reel.

For the high probability of operating fire protective and fire fighting measures, the

probability of impact from high risk is low (see Chapter 3, Section 3.1.5). However,

water based fire protection systems such as fire hose reel and sprinkler system were

not considered in Chapter 5 for evaluating the impact of high risk.

187

6.3 Methodology

Lift Malfunctioning due to Excessive Temperature Rise: @RISK software is used

to determine the probability of lift malfunctioning due to excessive temperature (see

Chapter 2, Section 2.10). The probability of temperature attainment in lift machine

room (LMR) can be given as:

( ) { }TTProbTF LMR <= 6.1

and TLMR is the temperature in LMR. F(T) is the probability distribution function and

it is the probability that the lift motor room temperature will not exceed any given

value T. Temperatures in LMR were determined from FDS modelling for Fire

Scenarios 1 to 6. Temperatures did not exceed the safe limit (43°C) in LMR for Fire

Scenarios 7 to 24. Therefore Fire Scenarios 7 to 24 are not considered for lift

malfunctioning (see Chapter 5, Table 5-3). The averages of the temperatures during

the simulation period for Fire Scenarios 1 to 6 are used to determine the probability

of lift malfunctioning.

Electric Power Failure: Fault tree analysis is used to determine the probability of

electric power failure. The impact of fire is considered on electrical system. A

typical electrical system of essential and non-essential electric supplies is illustrated

with its interrelationship with primary and secondary sources of power supply.

Boundary conditions are established during a fire scenario. Boundary conditions are

the physical boundaries of the system (i.e. which parts of the system are included in

the analysis and which parts are not?), the initial conditions (i.e. what is the

operational state of the system when the top event is occurring?) and external stresses

(impact from external events). Fault tree is constructed and outcomes are analysed

qualitatively and quantitatively. The effect of temperature in LMR is included in the

analysis.

Water Damage: Probabilistic analysis is conducted to determine the quantity of

water spread from various fire fighting measures. However, building evacuation may

or may not be required at the stage of fire extinguishment. The water spread is time

188

variant and can be described as a complex parallel and series system of water based

fire protective and fire fighting measures. This approach is based on a hybrid

combination of parallel and series system (Modarres and Billoch, 2002). The

probability of water spread can be derived from the individual probability of

occurrence POE of water spread from each fire protection system FPS. The

probability of water spread Pw can be derived as:

{ }∏=

=n

i

iiW POEFPSP1

)( 6.2

The water based fire protection systems are assigned a number from 1 to n. Each fire

protection system is assigned a probability of occurrence in order to arrive at a

probability of water spread. The estimated total quantity of water generated Qt from

the water based fire protection measures can be derived as follows:

∑=

=n

i

it tfQQ1

)( 6.3

>

<=

ttforQ

ttfortf

ii

i0)( 6.4

where

Qi is the water discharge of the individual fire protection system

t is the time of the fire protection system i to activate

ti is the total time of the fire protection system i to activate

The fire is extinguished with the fire protection and fire fighting measures and one of

the fire fighting system may be missing (or non-operational) during actual

emergencies. The scenario is considered for minimum and maximum water spread

cases. The complex parallel-series system is used in the analysis. The system is

divided into basic parallel and series modules and then the probability function for

each module is determined separately. The analysis determines the probability of

water spread and quantity of water spread on the floor.

189

6.4 Lift Malfunctioning due to Excessive Temperature Rise

Solid state electronic devices are used in the lift control systems and operate under

variable loads and demands. These devices are subjected to a regular stress during

the normal operating conditions. Internal heat is released from the losses in

machinery such as lift motor, control cabinet transformer, converter, invertors and

power supplies. External heat from ambient conditions and other electrical devices

such as lights is also added to the stress level. These stresses may reduce the

reliability of lift systems. Therefore, lift manufacturers have recommended a

temperature limit of 32°C for LMR (this value may vary from manufacturer to

manufacturer). This value is chosen by the lift manufacturers to ensure that the actual

temperature in the controller cabinets, which is typically 10°C to 15°C higher than

the ambient room temperature, is not above the design operating limits of the solid

state devices (Marchitto, 1991). Under the recommended limit, there is no effect on

the functionality of the lift system.

The Australian Standard AS 1735.1 (2003) specifies an ambient temperature not

more than 43°C in LMR. This is the maximum temperature under which electronic

devices can operate without a safety margin. Degraded performance and long term

reduced reliability may result between the recommended and specified temperatures

(see Table 6-1). Beyond the specified rating of electronic devices, the performance

is not assured. Electronic devices may malfunction under high stresses and may

recover during the reduced load conditions. Electronic devices for lift systems can be

commercial, industrial or military grades. The design temperature for commercial

grade is 70°C, for industrial grade is 85°C and for military grade is 125°C

(Robibero, 1991). Above the design temperature, electronic devices may not recover

and the lift system may go to a permanent failure mode (see Table 6-1).

190

Table 6-1: Temperatures and their Impact on Lift Systems

Temperature Impact

Less than 32°C No impact

Between 32°C and 43°C Long term reduced reliability

Between 43°C and 70°C Lift malfunctioning

More than 70°C Failure

The reliability of lift systems can be adjudged from the thermal environment that lift

devices may be subjected to. The field model (FDS) results indicated that lift

machine rooms via unprotected lift lobbies are susceptible to high temperatures.

However, protected lift lobbies provide adequate safety from high temperatures

occurring in LMR due to fires. The probability of lift malfunctioning depends on the

temperature rise beyond the temperature threshold of the electronic components (i.e.

43°C).

The computer package @RISK (Palisade Corp, 1996) is used to determine the

probability of lift malfunctioning. The cumulative graphs show the temperature to be

more than 43°C or 70°C in Fire Scenarios 1 to 6 (see Figure 6-1). They show the

probability (percentage) of malfunctioning due to excessive temperature. By

dragging the delimiters displayed on the cumulative graph, the probability

(percentage) of exceeding a temperature is calculated (for example, it is 100 – 88.4 =

11.6 for 43°C in Fire Scenario 1). The cumulative graphs indicate that the

temperature beyond the threshold value is a function of the vertical location of a fire.

In the cumulative graph, blue line indicates the actual values of temperature, whereas

red line shows the imaginary best fitting curve for determining the percentage of

malfunctioning due to excessive temperature.

191

Fire Scenario 1

43.00

88.4%20.00

5.0%

0

0.2

0.4

0.6

0.8

1

15 20 25 30 35 40 45 50 55 60

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

Fire Scenario 2

20.00

5.0%

43.00

88.2%

0

0.2

0.4

0.6

0.8

1

15 20 25 30 35 40 45 50 55 60

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

Fire Scenario 3

20.0

5.0%

70.0

84.2%

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

192

Fire Scenario 420.00

5.0%

43.00

92.3%

0

0.2

0.4

0.6

0.8

1

15 20 25 30 35 40 45 50 55 60

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

Fire Scenario 5

20.6

5.0%

70.0

94.2%

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140 160

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

Fire Scenario 6

20.6

5.0%70.0

93.7%

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

Temperature (C)

F(T)

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

@RISK Student Version

For Academic Use Only

Figure 6-1 – Probability of Excess Temperature Rise in LMR (Fire Scenarios 1 to 6)

193

The probability (percentage) of exceeding a temperature is used to calculate the

probability of reduced reliability or malfunctioning or failure (see Table 6-2). The

average probability (percentage) of exceeding the recommended temperature of 32°C

or more is 36.26% (0.36) for Fire Scenarios 1 to 6 whereas that for the specified

temperature of 43°C or more is 19.96% (0.2), and the designed temperature of 70°C

or more is 4.65% (0.04), for the lift machine room located on the top of unprotected

lift shaft.

Table 6-2: Probability of Excess Temperature Occurrence in LMR in Fire Scenarios

Probability (%) of excess temperature Fire Scenario

More than recommended

temperature 32°C

More than specified

temperature 43°C

More than designed

temperature 70°C

Fire Scenario 1 24.9 11.6 -

Fire Scenario 2 25.3 11.8 -

Fire Scenario 3 40.8 33.1 15.8

Fire Scenario 4 19.8 7.7 -

Fire Scenario 5 40 20.6 5.8

Fire Scenario 6 66.8 35 6.3

Average 36.26 19.96 4.6

The probability of lift malfunctioning (including failure) is conservatively considered

to be for temperatures more than 43°C. The average probability of reduced

reliability between temperature 32°C and 43°C is calculated as 0.16 (36% – 20% =

16%). Similarly, probability of malfunctioning between temperature 43°C and 70°C

is calculated as 0.16 (20% – 4% = 16%). The probability of complete failure, above

temperature 70°C, is 0.04 (4%). The probability of lift malfunctioning, above

temperature 43°C, is therefore equal to 0.2 (see Figure 6-2).

194

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100 120 140

Temperature (C)

Probability

32 7043

Complete

failure

(0.04)

Reduced

reliability

(0.16) Malfunction

(0.16)

No effect

(0.64)

Figure 6-2 – Probability Distribution and Consequences of Excess Temperature in

LMR

With the failure probability of 0.2, on average 41% of the building population may

not be able to use the alternative evacuation facility i.e. lifts (see Table 6-3). A

portion of this population may also be trapped in the lift cabins in the unprotected lift

shaft.

Table 6-3: Unavailability of Lifts to Building Population

Fire scenario Probability of excess temperature rise

(43°C)

Time to exceed limit for specified

temperature (second)

% of population remaining in the

building

Fire Scenario 1 0.11 2473 0

Fire Scenario 2 0.11 2485 0

Fire Scenario 3 0.33 1290 36

Fire Scenario 4 0.07 1022 54

Fire Scenario 5 0.2 854 67

Fire Scenario 6 0.35 522 88

Average 0.2 1441 40.8

Probability of lift malfunctioning increases as the fire is on the upper portion of the

building. A significant number of persons may face the problem of unavailability of

lifts (although remaining persons would have the option of using stairs).

195

6.5 Electric Power Failure

There are no general statistics on power failures in high rise building fires. With the

help of fault tree analysis, the probability of electrical power failure is determined

during the occurrence of a fire.

6.5.1 System Descriptions

The common provision is made for two sources of electric power supplies to a

common bus bar as shown in Figure 6-3.

Figure 6-3 – Typical Electrical System for Essential and Non-Essential Supplies

Hot gases

Non-essential electric

supply

Essential electric

supply

To LMR

Local fuse

To SOU

Circuit breaker

Circuit breaker

Circuit breaker

Transformer-2 (Secondary supply –

generator)

Transformer-1

Bush bar

Manual trip

by fire fighter

(Primary supply

– city mains)

Electric Fire

196

Electrical supply to the essential services is independent and uninterrupted at all time

even if the entire electrical supply in the premises is switched off. Essential services

include fire fighting pumps, smoke control system, emergency lighting and lifts

whereas non-essential services include apartment units. Secondary source of power

supply (generator set) is required to ensure the continued operation of essential

equipment during building evacuation.

6.5.2 System Boundary Conditions

Physical boundary of the system is defined by incorporating the primary and

secondary sources of power supplies, distribution to essential and non-essential

power supplies during the occurrence of a fire (see Figure 6-3). The initial condition

is assumed to be the occurrence of a fire in a sole occupancy unit (SOU) and then

spread to common area. The fire can result from an electrical short-circuit. Table 6-4

indicates the initial conditions and top event. Boundary conditions with respect to

external stresses (fire and hot gases) are included in the analysis.

Table 6-4: System Boundary Conditions

Condition Initial Conditions External Stresses

Condition 1 Electrical short circuit within the compartment

of fire origin

Nil

Condition 2 Fire spread beyond the origin of fire – Impact

on unprotected lift lobby

Lift malfunctioning due

to excessive temperature

Condition 3 Fire spread beyond the origin of fire – Impact

on protected lift lobby

Nil

6.5.3 Data and Statistics

The following data/ statistics were used in the fault tree analysis:

• NPFA recorded the causes of the fires in apartment buildings, which shows

that 10% of them are due to electrical fires. Hence 0.1 is considered as the

probability of electrical fire in apartment buildings.

• The failure rate of electrical surge is assigned 0.04 /D (WASH 1400, 1974).

This failure rate is an occurrence in which the circuit breaker ceases to

perform the required function on demand (D). Electrical surges on the mains

197

supply can come from a lightning or heavy current load through inductive

effects. Lightning may also induce currents into signal cables entering a

building. Lift motors can generate a significant number of surges.

• The failure rate of circuit breaker is assigned 1 × 10-3 /D (WASH 1400,

1974). This failure rate is an occurrence in which the circuit breaker ceases

to perform the required function on demand.

• The failure rate of electric fuse is assigned 1 × 10-5 /D (WASH 1400, 1974).

This failure rate is an occurrence in which the electric fuse ceases to perform

the required function on demand.

• The probability of fire spread beyond the origin of fire (fire compartment) is

0.06 (see Chapter 2, Table 2-2).

• For electric power failure, urban Australian locations are subjected to

approximately three outrages of 10 minutes duration per annum, i.e., a failure

rate of 5.70 × 10-5 per annum (Lacey, 2000). Failure rate is the frequency of

power outrage expressed in failure per annum and calculated as 0.5 hour / (24

× 365). The failure is based on day-to-day power supply.

• The failure rate of secondary power supply (generator set) is assigned 0.03 /D

(WASH 1400, 1974). This failure rate is an occurrence in which the

generator ceases to perform the required function on demand.

• The error rate of manual tripping by operator or fire fighter is assigned 0.003.

Operator or fire fighter may shut down the entire electrical system (including

essential supply) by selecting the wrong control in a group of identical and/or

labeled controls (NUREG/CR-1278, 1980).

• Teo (2001) highlighted that the lift breakdown rate is once in three months in

apartment buildings over a period of 10 years. Taking an average breakdown

for a period of 4 hours, a probability of failure is calculated as 1.82 × 10-3 per

annum. The breakdown rate is the frequency of lift maintenance during which

lifts are not available, expressed as breakdowns per annum and calculated as

4.0 hour × 4 breakdown services / (24 × 365).

The probability of lift malfunctioning due to excessive temperature is 0.2 for lifts

with unprotected lift lobby (see Section 6.4). The probability of lift malfunctioning is

considered nil for lifts with protected lobby.

198

6.5.4 Fault Tree Analysis

A fault tree is a “top-down” method of analysing the susceptibility of a product to

failure. Fault tree analysis contains the following modules i.e. Basic Event,

Undeveloped Event, ‘AND’ Gate and ‘OR’ Gate and Fault Event. Basic Event

contains a failure at the lowest level which has the capability of causing a fault to

occur. Undeveloped Event contains a failure at the lowest level which is not fully

developed due to lack of information. The ‘OR’ gate indicates that the output event

occurs if any of the input events occur. The ‘AND’ gate indicates that the output

event occurs only if all the input events occur at the same time. Fault Event contains

description of a lower level fault. External Event is the occurrence of fault externally.

Condition 1: Fault tree analysis to estimate the probability of electric failure due to

fire in electrical origin in an apartment is shown in Figure 6-4. The data used in the

fault tree analysis is obtained from section 6.6.3. The power is fed from non-essential

services, which is not connected to the lift system. This has little effect on lift electric

power supply.

Figure 6-4 – Fault Tree Analysis for Electric Fire in SOU

Excessive current in main circuit

“OR” Gate (add)

1.0E-6

Legend

“AND” Gate (m) Excessive current in local circuit

Local fuse fails to open

Primary wiring failure (short circuit)

0.104

Primary power supply failure (surge)

0.104

1.0E-05

1.0E-6

Circuit breaker fails to open

Main electric supply failure

1.0E-09

Basic Event

0.1 0.04

Fault Event/

Top Event

1.0E-03

Electric power failure in non-essential system

(though primary power supply is available)

1.0E-09 1.0 × 10-9

199

Condition 2: Fault tree analysis to estimate the probability of electric power failure

in unprotected lift lobby is shown in Figure 6-5. The probability of lift electric power

supply failure is determined from the impact of fire in a common area (beyond the

origin of fire – SOU), primary and secondary sources of electric power failure,

malfunctioning of lift due to excessive temperature, lift maintenance breakdown and

manual tripping. The probability of lift electric power failure in unprotected lift

lobby is 0.2648.

Figure 6-5 – Fault Tree Analysis for Electric Power Failure in Unprotected Lift

Lobby

Malfunctioning of lifts due to hot

gases

Secondary power supply failure (Generator)

Primary power

supply failure

Lift power supply failure

0.2

0.03

0.003

Essential power supply

failure Manual

tripping

5.7E-5

1.71E-6

1.71E-6

0.2648

Lift maintenance breakdown

Electric wire short circuit due to fire

or hot gases

1.82E-3

“OR” Gate

Legend

“AND” Gate

Undeveloped Event

Basic Event

Fault Event/

Top Event

External Event

0.06

200

Condition 3: Fault tree analysis to estimate the probability of electric power failure

in protected lift lobby is shown in Figure 6-6. The probability of lift electric power

supply failure is determined from the primary and secondary sources of electric

power failure, lift maintenance breakdown and manual tripping. The probability of

lift system failure due to electricity is 0.0048, which is during the reasonably worst

circumstances (manual breakdown and manual tripping).

Figure 6-6 – Fault Tree Analysis for Electric Power Failure in Protected Lift Lobby

Secondary power supply failure (Generator)

Primary power

supply failure

Lift power

supply failure

0.03

0.003

Essential power supply

failure Manual

tripping

5.7E-5

1.71E-6

1.71E-6

4.82E-03

Lift maintenance breakdown

1.82E-3

“OR” Gate

Legend

“AND” Gate

Undeveloped Event

Basic Event

Fault Event/

Top Event

201

6.5.5 Analysis of Results

Table 6-5 indicates the probability of lift failure in an event of fire.

Table 6-5: Impact on Lift System

Condition Initial Scenario Probability Impact on Lift System/

Reasons

Condition 1 Electrical short circuit within the compartment of fire origin

1.0 × 10-9 No impact on lift electric system; The probability is related to the failure of non-essential services.

Condition 2 Fire spread beyond the origin of fire – Impact on unprotected lift lobby

0.2648 High impact is on essential services; Hot gases in LMR are the main reasons for high probability of failure.

Condition 3 Fire spread beyond the origin of fire – Impact on protected lift lobby

0.0048 A little impact on lift electric system; The probability is related to the failure of essential services.

For condition 1, fire of electric origin in SOU does not contribute to the failure of lift

electric power system. Figure 6-3 illustrates that electric fire in SOU is confined to

the non-essential electric services only (not connected to the essential services). The

derived probability (1.0 × 10-9) is related to the failure of non-essential services.

For condition 2, fire spread from SOU to a common area (including unprotected lift

lobby) contributes significantly to lift failure. External stress from lift malfunctioning

due to excessive temperature in the building is the main reason for electrical failure.

The condition is assumed under which (a) fire spread beyond the origin of fire in

common area, (b) smoke and hot gases spread in unprotected lift lobbies, (c) manual

tripping of electrical essential services, and (d) failure of electric supply system. If

the wiring is laid in the common area, the damage from fire or hot gases may occur

due to direct exposure of wire. However, FDS results have demonstrated that the

maximum temperature during the period of fire simulation is less than 300°C in

unprotected lift shaft, which may not be capable of causing electric short circuit

(exposing/ melting of PVC wire) during a short time.

202

For condition 3, protected lift lobby provides a higher protection against the lift

failure. The probability of electrical failure is only 0.0048.

The probability of a fire outbreak and incoming electrical power failure is relatively

small, provided that the fire and electrical power failure do not have the same cause

in the building. However, the probability of electrical power failure during a fire is

relatively high in unprotected lift lobbies since the fire frequently damages electrical

systems in the building.

6.6 Probabilistic Analysis of Water Damage

The water requirement for fire fighting mainly depends on the size of fire. The water

based fire protective system includes fire hose reels (FPS1), sprinklers (FPS2) and

fire hydrants (FPS3) in the apartment buildings. Sprinklers cover the risk area

whereas fire hose reels and fire hydrants are generally positioned near the exit routes

such as locations near stairs or lift lobbies. Professional fire fighting operation is

generally conducted at a later stage, when fire brigade is called automatically or

verbally.

Fire suppression technology employs the life cycle of fire progress in different forms.

The incipient stage where heat is low to moderate and there is not much visible

smoke or flame but there is the threat of fire propagation. At this stage, the occupants

may undergo coping stage and they would like to use some handy form of fire

extinguishment such as fire hose reel (FHR). The water spread is minimal. If the fire

is not controlled and visible smoke or heat threatens the life, occupants have no

option but to leave the premises. At this stage or later, sprinkler heads will actuate to

control or extinguish the flame. The water spread is moderate. As the fire further

progresses and is not extinguished with sprinklers, fire-fighters intervene to

extinguish the flame. The water spread is the maximum since the quantity of heat, at

this stage, is at its peak and therefore fire fighters are using nozzles of bigger sizes. A

large quantity of water is accumulated by this time on the floor. With the application

of water, fire may be uncontrolled, controlled or suppressed, which may or may not

require building evacuation. Figure 6-7 shows the heat release rate (HRR) with time

203

during the three stages of water application. This graph is a modified form of graph

indicated in FCRC (1996) for HRR during sprinkler operation.

Figure 6-7 – HRR during Three Stages of Water Application

6.6.1 Water Spread from Fire Protection and Fire Fighting Measures

The probabilities of water spread from fire protective and fire fighting measures

determined from statistics are given below:

FPS 1 Fire Hose Reel – Nearly 10% of fires are extinguished with the help of

building manual extinguishing facility (NZFS incident statistics, 1993-1997).

Uncontrolled fire may include non-operational (non-functional) of manual fire

fighting facility and failure to extinguish the fire with the help of manual fire fighting

facility. The data on non-operational/ non-functional and failure to extinguish with

the help of manual fire fighting facility such as fire hose reel are not available. It is

assumed that the percentage of non-functionality of manual fire fighting facility is

12.4. This is considered at par with sprinkler system (see FPS2 sprinkler). Out of the

remaining 88.6%, 10% is considered as fire extinguishment with the help of fire hose

reel and 77.6% is considered as the failure to extinguish the fire. Therefore, the

probability of water spread due to fire hose reel for extinguishment is assigned as

Manual fire fighting

(FPS1)

Automatic fire

fighting (FPS2) H

RR

Time

FHR uncontrolled

Sprinkler extinguishment

Sprinkler uncontrolled

Sprinkler control

FHR extinguishment

Fire brigade extinguishment

Fire brigade extinguishment

Internal fire fighting

Fire brigade intervention +

others (FPS3)

FHR failure

204

0.1, the probability of water spread for uncontrolled fire is 0.776 and the probability

of non-water spread due to non-functionality is 0.124. Technically, a fire hose reel is

capable of producing 0.55 l/s (AS1221, 1997). However, when applied to fire, some

amount of water is used for extinguishing, some goes to vaporisation and the rest

goes on spreading. Building evacuation is required under uncontrolled fire conditions

arising in the preliminary stage.

The use of FDS model showed that the fire detector in the SOU operated at 90

seconds (see Chapter 5, Section 5.6). So, it is assumed that the occupants would

commence fire fighting operation after travelling to the corridor and fetch fire hose

reel for extinguishment. Fire hose reel is considered as first aid fire fighting

equipment and meant for the use of general public. The fire fighting operation starts

after 120 seconds and the tenability limits are exceeded at 187 seconds in the SOU.

Therefore the occupants presumably continue fire fighting for 67 seconds and are

forced to leave the premises due to untenable environmental conditions.

FPS2 Sprinklers – The NFPA statistics (Rohr, 2001) for the ten year reporting period

from 1989 to 1998 indicates that the operational reliability of automatic sprinkler

systems for apartment buildings is 87.6%. However, most of the sprinklers are

designed to control the fire but not necessarily to extinguish the fire. Statistics for the

same period showed that the percentage of fires, where sprinklers are present and

that are reported as being extinguished by an automatic suppression system, is 20%

in apartment buildings. Automatic suppression system is primarily the sprinkler

system in apartment buildings. Therefore the probability of fire extinguishment using

sprinklers is 0.2 and controlled fire is 0.676 while that of failure is 0.124.

Marryatt (1988) reports an average of 1.22 sprinklers in operation for the 33 recorded

fires in apartment buildings. As a result the fire is kept from spreading with a

minimal amount of water. Generally sprinklers are installed in apartment buildings

according to extra light hazards. Water spread from the sprinkler system depends on

sprinkler water spray rate, water distribution pattern, inter-spacing among the

sprinklers and design of the building. Heat is absorbed by the discharged water from

the sprinklers and a portion of discharged water vaporizes into steam. The excess

water flows and causes the water spread on the floor and subsequently may cause

205

damage to the lift system. The water application rate for apartment building is 2.25

mm/min, the area of sprinkler operation is 84 m2 and water flow rate is 225 l/m for

the duration of 30 min. Hence, the water discharge rate is calculated as 4.57 l/s in

the sprinkler operated area.

FPS3 Fire Hydrant – An analysis of residential fire incidents reveals that 65% of the

fires are extinguished by the fire service. Remaining 35% of the fires were

extinguished by passerby or other methods (Davis, 2000). About 90% of residential

fires are controlled in less than an hour (Beever and Davy, 1999). The emergency

response time of fire brigade is established to a maximum of 10 minutes in order to

arrive at the fire incidents. The Fire Brigade Intervention Model (FBIM) estimates

the time of arrival of the fire service in the enclosure of fire origin (AFAC, 1997).

Residential properties require less than 10 l/s of water to extinguish a fire (Davis,

2000) and typical hydrant flow rate is 30 l/s.

6.6.2 Complex Parallel and Series System for Water Spread

The probabilities arising from the water based fire protective measures and fire

brigade intervention, due to which water can spread, are shown in Figure 6-8. The

quantity of water spread is shown in Figure 6-9. The values are represented

individually against the outcomes.

Three levels of water spread are assumed for causing damage. Low level of water

spread is assumed from fire hose reel. Medium level of water spread is assumed from

sprinkler system alone (or including fire hose reel). High level of water spread is

assumed from fire brigade intervention alone (or including sprinkler system and fire

hose reel). The levels and their combined probabilities of occurrence are given in

Table 6-6. The level of water spread is virtually absent if none of the water based fire

protective or fire fighting measures is functioning.

206

Table 6-6: Probability of Water Spread at Three Levels

Level of Water

Spread

Fire Protective or Fire

Fighting

Combined Probability of Occurrence

Low Fire hose reel 0.1 + 0.034 + 0.005 = 0.139

Medium Sprinkler system alone (or including fire hose reel)

0.025 + 0.155 + 0.029 + 0.184 = 0.393

High Fire brigade intervention alone (or including sprinkler and fire hose reel)

0.01 + 0.34 + 0.055 + 0.063 = 0.468

The combined probability of occurrence for water spread at the low level is 0.139,

the medium level is 0.393 and the high level is 0.468. Water gets moving at a slope

of approximately 1 cm per metre. Water may flow toward the lift shaft during the

medium or high level of water spread (sprinkler or fire brigade intervention) and

damage the lift system. Water spread from fire hose reel may not cause lift damage.

Further, unprotected lift lobbies are more likely to be subjected to water damage as

lift lobbies are directly connected to the public corridor or risk area. Protected lift

lobbies provided with suitable form of compartmentation from risk area and adhering

with norms for water control measures can restrict water spread. However, this is a

design aspect (see Chapter 2, Section 2.7.4).

207

Figure 6-8 – Complex Parallel-Series System for Probability of Water Spread

208

Figure 6-9 – Complex Parallel-Series System for Quantity of Water Spread

209

6.6.3 Water Spread Result Analysis

The probability of water spread due to sprinkler and fire fighting operation is

considerably high. The combined probability of water spread (0.468) includes fire

brigade intervention, during which fire is controlled and extinguished. The maximum

quantity of water spread occurs from fire brigade intervention.

The probability of maximum water spread as a single outcome was 0.34 for the

quantity of water accumulated from fire hose reel, sprinklers and fire brigade

intervention (see Figure 6-8). The maximum quantity of water spread was 28 000 ltrs

approximately (see Figure 6-9). The maximum quantity of water spread is plotted

against the population remaining in the building (see Figure 6-10). The remaining

population was determined from the stochastic evacuation model (see Chapter 4).

0

5000

10000

15000

20000

25000

0100

150

187

200

250

300

350

370

600

900

1200

1500

1800

2100

2400

Time (second)

Water (litre)

0

10

20

30

40

50

60

70

80

90

100

Population (%)

Water Quantity

Population remaining in building

FHR

Phase

Sprinkler Phase Fire Brigade

Intervention

Figure 6-10 – Quantity of Water Spread and Building Evacuation

The water damage may occur if the building is not provided with adequate drainage

facilities. About 80% of the population remains in the building at the time of fire

brigade intervention (see Figure 6-10). By this time, about 1800 litres of water has

already accumulated from the fire hose reel and sprinkler system. In such

circumstances, fire brigade at the scene can envisage if their fire fighting operation is

causing enough water damage to the lift system and assist lift evacuation.

210

6.7 Outcomes

Following outcomes are obtained from the analysis:

• The protected lift shafts provide adequate safety to LMR and do not require

additional provisions of redundancy measures. The unprotected lift shafts do

not provide adequate safety to LMR.

• The protected lift shafts containing protected electric supply do not require

additional provisions of redundancy measures. Dual electrical supply in

protected lift shaft ensures continuity of lift operations. The electrical

installation needs to conform to the relevant codes and standards (see Chapter

2, Section 2.7.4). The unprotected lift shafts do not provide adequate safety to

electric power supply.

• Water spread is a design aspect, which must follow the provisions of relevant

codes and standards (see Chapter 2, Section 2.7.4).

6.8 Influence on Human Behavioural Response

Lifts with unprotected lobby do not provide adequate safety against lift

malfunctioning, electrical power failure and water spread. The probability of lift

malfunctioning is 0.2. A combined probability of lift malfunctioning, electric power

failure and maintenance breakdown is 0.26. During the probability of maximum

water spread (0.43), about 80% of the population are remaining in the building.

Irrational human behavioural response would be the maximum due to unavailability

of evacuation route.

Lifts with protected lobby provide adequate safety against electrical malfunctioning,

electrical failure and water spread. The probability of lift malfunctioning due to

excessive temperature is little as there is no rise in LMR temperature due to

compartmentation. The probability of electrical power failure is also reduced to a

great extent (0.0048). Water spread can be restricted by adhering the norms such as

compartmentation, floor slope and drainage facilities. The reliability of lift

operational mechanism has increased considerably and irrational human behavioural

response would be the minimal.

211

6.9 Conclusion

This chapter has described an investigation into the reliability of lift operational

mechanism. The results demonstrated that the lifts protected with lobby are

significantly more reliable in comparison to the lifts without lobby. The spread of

hot gases to the LMR via unprotected lift lobby increased the probability of lift

malfunctioning. In unprotected lift lobbies, temperature rise in LMR depends upon

the location of fire. Protected lift lobbies provided adequate safety against the

temperature rise in LMR. The probability of electric power failure due to fire in a

residential unit was found to be small. Electrical system adhering to relevant codes

and standards in protected lift lobbies provides adequate safety for electrical

installation.

Another cause of lift malfunctioning is the water used in fire fighting. A complex

event tree analysis was conducted for determining the quantity and probability of

water spread from fire protective and fire fighting measures. It was found that a

copious amount of water generated from fire fighting measures could damage the lift

components. The maximum quantity of water spread is predicted to occur from fire

brigade intervention. The probability of maximum water spread was higher with all

fire fighting measures. Unprotected lift lobbies are more susceptible to water damage

in the absence of any protection measures (as getting water flow in the absence of

barrier). Protected lift lobbies with protection measures (for example floor slope and

water drainage facility) could provide a suitable barrier to water spread.

The reliability of lift operational mechanism contributes to the satisfactory use of lift

evacuation system. Evacuees’ irrational behavioural can not be ruled out in

unprotected lift lobbies. The probability of evacuees’ irrational behavioural have

considerably reduced in protected lift lobbies. The reliability is used in the

calculation of overall risks of using lift evacuation systems in the next chapter.

212

7. RISK ASSESSMENT OF EVACUATION ROUTES

7.1 Introduction

Fire safety strategies are often developed based on uncertain conditions and sparse

data (Watts, 1995). Modelling fire risk is extremely complex process and involves a

network of interacting components. Analysing the risks in lift evacuation routes

require systematic and practicable approaches. The risks in lift evacuation system

range from asphyxiant toxic gases, movement of smoke and hot gases in the

evacuation routes, behaviour of people in fire condition to reliability of lift

operations. The consequences can vary from psychological to physiological impact

with known (or unknown) probabilities. Therefore, an integrated risk assessment

methodology is required for a consequence based analysis. The consequence based

analysis gives the overall probability of psychological and physiological impacts on

evacuees arising in the evacuation routes.

The ‘Multi-Objectives Decision Analysis’ (MODA) is a consequence based analysis

method and provides an evaluation of influencing, dependent and interacting issues.

The MODA does not provide solutions, but is rather an information source,

providing insight into the situation, uncertainty, objectives and trade offs (Pfeiffer,

1997). The MODA is defined as an approach to decision making under conditions of

complexity, with inherent uncertainty, multiple objectives and different perspectives

towards the decision problem (Clemen, 1996). It is a systematic procedure for

transforming opaque decision problems into transparent decision problems on the

basis of a sequence of transparent steps (Howard, 1988). Opaque means ‘hard to

understand, solve or explain’ and transparent means ‘readily understood, clear or

obvious’. This approach offers a plausible solution and is used for the first time in

the field of fire safety engineering in this research project. It has been successfully

used earlier in other fields such as economics, nuclear energy and resources, policy

analysis, scientific research management, industrial management, manpower

planning and medical diagnosis and defence.

213

The MODA approach demonstrates the impact on evacuees on the same scale for all

the concept design options considered. The parameters (variables) are given

appropriate weights based on data generated from statistics and survey reports. The

priorities of conflicting key issues are assigned with the help of Analytical Hierarchy

Process (see Chapter 3, Section 3.1.2). The analysis is prescriptive and the approach

can assist decision makers to understand all aspects of the building evacuation

process and can reveal insights into design options.

7.2 Risk Analysis of Building Evacuation System

Important aspects of lift evacuation system include evacuation time periods and

tenability limits of fire, smoke and toxic gases in evacuation routes. Researchers

have addressed these areas in isolation (Kuligowski, 2003 and Klote, 2003). A

comprehensive approach for addressing the risks in relation to these aspects is

discussed here. An integrated risk evaluation model is proposed. The objectives of

this chapter are to:

1. develop a model for analyzing the risks involving ‘uncertainty’, ‘panic’ and

‘injuries (nonfatal and fatal)’ in building evacuation system; and

2. develop a decision model applicable to risk assessment of building

evacuation system; and

3. analyse the risks associated with the lift evacuation system.

7.2.1 Assumptions

The risks are compared between lift and stair systems based on the following

assumptions:

• low risk (or decision uncertainty) occurs in the lift and stair systems. This

may be caused by longer lift waiting time or longer travelling time in the

stairs.

• medium risk (or panic) occurs in the lift and stair systems. This may be

caused by visual threat (pre-life threatening condition) or unavailability of

evacuation route (lifts or stairs).

214

• high risk (or injuries) occurs in the lift and stair systems. This may be caused

if the evacuees are exposed to life threatening conditions in the lift system or

stairs. Life threatening conditions can arise from smoke, toxic gases, fire/

temperature and precipitated risk from pre-existing health conditions.

• evacuees are not considered safe inside the building, although they may be in

the protected stairs or protected lift lobbies or lift cabins. Hence evacuation is

considered complete, if they exit the building.

• lift evacuation strategy for one-fourth of the building population includes

mainly aged and disabled persons.

7.2.2 Methodology – Multi-Objectives Decision Analysis

The Multi-Objectives Decision Analysis (MODA) technique used here is based on

the Simple Multi-Attribute Rating Technique (SMART) of Edwards (1977) and

further illustrated by Donegan (2002) as explained in Figure 7-1.

Figure 7-1 – Multi-Objectives Decision Analysis Methodology

wn …… w2 w1

p2 p1 …… pn

Parameters

(Step 2)

Concept Design Option 1 Concept Design Option 2 …

Concept Design Option…

Values

(Step 4)

Parameters (Step 2)

Model Options (Step 1)

v1 v2 ….... vn

p2 p1 …… pn

Weights (Step 3)

w1 v1 w2 v2 …… wn vn + + +

Sensitivity Analysis (Step 5)

Result Evaluation (Step 4)

( )∑=

=n

i

iii pvwR1

215

The MODA involves the following steps:

Step 1: Identify concept design options

Step 2: Specify evaluation considerations and evaluation measures

Step 3: Specify weights to each parameter

Step 4: Determine value functions

Step 5: Analyse the results (sensitivity analysis)

The first step involves identification of potential concept design options for risk

assessment. It takes into account relevant events and risks in a system. This is carried

out with the help of influence diagram (see Figure 7-2). Influence diagram is a

structure display of decisions, uncertain events and outcomes, and provides a

snapshot of the decision environment at a single point in time. Evaluation measures

are the quantitative weights (importance) assigned to the evaluation considerations as

the contributing parameters may not have equal weights. A formal process is adapted

to award weights using AHP risk priorities and data from various sources. Values are

measures of the intensity, level or degree of hazard afforded by the parameters in a

particular design. Building evacuation models, fire hazard models and reliability

assessment models are used to determine the values. A scaling technique is used to

capture the essential meaning of quantitative values on a ratio scale. The ratio scale is

an interval scale with absolute zero on one end so that the values on it are absolute

rather than relative. The ratio scale identifies each individual parameter so that

reliable difference among the design models can be represented. Thus models can be

rated on a quantitative basis for results. The model is constructed in terms of n

evaluation parametric measures p1, p2, . . . , pn, and the overall value of the model is

given by:

( )∑=

=n

i

iii pvwR1

7.1

where the wi are weights and the vi(pi) are non-dimensional value functions that

normalise dimensional parametric measures pi (i=1, 2, …, n) into non-dimensional

values between 0 and 1 (or between 0 and 100). When the MODA approach is used

for risk assessment, the risk, as defined by Eq. 7.1, is interpreted as a total of

216

weighted multiple non-dimensional risk attributes. The impact of varying the

relative weight for the evaluation measures can be studied by sensitivity analysis.

7.3 Risk Assessment

7.3.1 Identify Concept Design Options and Evacuation Strategies

The use of lifts can be associated with evacuation strategies. Two evacuation

strategies, namely use of lifts by the entire building occupant population or a 75-25%

stair-lift split, were considered. In total five concept design options along with

alternative evacuation strategies are considered as listed and labelled in Table 7-1.

Table 7-1: Concept Design Options and their illustrations

Concept Design Illustration

A Lifts with unprotected lobby for use by 100% population

B Lifts with protected lobby for 100% population

C Lifts with double protected lobby for 100% population

D Stairs for 100% population (stairs only)

E Stairs used by 75% population (E-75) and protected lifts by 25% (E-25)

The concept design option D is an option for which the associated risk is deemed

acceptable. This option was included as a reference for the comparative study.

7.3.2 Evaluation Considerations and Evaluation Measures

Figure 7-2 indicates three risk levels in the influence diagram of building evacuation

risk model. Based on perceptual information, evacuees’ behavioral response

determines the actions he or she would perform. Perception is defined as the

awareness of the human being of environment through physical sensation. Evacuee

may simply respond in his/her environment or may cope (or interact) with this

environment or even communicate with other evacuees. Based on the level of

perception, two kinds of behavioural autonomy may be generated: either to evacuate

by stairs or lifts. The autonomous behaviour concerns the capability of acting

independently exhibiting control over their internal state.

217

Figure 7-2 – Influence Diagram of Building Evacuation Risk Model

Main issues of human behavioural response and life safety can cause decision

uncertainty, panic and injuries (nonfatal or fatal) in the lift and stair systems. The

evaluation considerations are kept common for risks in both systems (lifts and stairs).

Evaluation criteria (or parameters), units, symbols are specified in Table 7-2. The

parameters of similar nature are selected for both systems. The parameters may or

may not be directly related to the risk involved in the lift and stair systems. The

parameters which are not directly related to the risk can be referred as proxy-

parameters. The parameters of each issue are identified for assigning a corresponding

parametric weight.

Three main risks have seven parameters, which need to be given weights and values.

These parameters may be argued. The parameters p1, p2 and p4 are considered as

proxy-parameters as these parameters are not directly related to the risk. The

parameters are considered after a deliberation of thoughts, judgments and availability

of statistical data. For example, three parameters are considered for decision

uncertainty i.e. the time for building evacuation, the number of evacuees in queue

and the percentage of aged and disabled persons. Decision uncertainty may also

depend on building features, namely, the number of lifts, the fire safety and fire

protection systems and the level of the fire-affected floor. These parameters are

DECISION FOR ROUTE CHOICE HIGH RISK MEDIUM RISK LOW RISK

OR

Overcrowd

Decision

uncertainty

Untenable condition in

lift

Long

waiting

Use of

lifts

Use of

stairs

Perceptual

information

Action

Autonomy

decision

Untenable condition in

stair

Long

travelling

Lift un-

availability

Panic

Lift stuck during

transport

Injuries (nonfatal

or fatal)

218

lacking their importance and statistical resources. Evacuees may not be concerned

about the number of lifts (or stairs) and the types of fire safety and fire protection

measures in the building. The level of fire-affected floor is also involved in multiple

floor analyses (keeping in view of fire at each floor level). Hence these parameters

are not considered. In MODA, large number of parameters can be reduced to smaller

number of parameters or appropriate sub-sets (Watt, 1991).

Table 7-2: Risk related Parameters

Risk Category Parameter Symbol Unit

Lift waiting time (tLW) and transportation time (tLT ) or stair travelling time (tST)

p1 Second

Number of evacuees in queue in lift or stair lobbies

p2 Person

Decision uncertainty

Proportion of aged and disabled evacuees p3 %

Non-availability of evacuation route p4 % Panic

Time to exceed tenability limit for visual threat

p5 Second

Safety index p6 − Injuries (nonfatal and fatal) Presence of fire effluents in evacuation route

(e.g., temperature, concentrations of smoke and asphyxiant toxic gases.)

p7 −

7.3.3 Specify Weights

Each parameter is given a degree of importance (weight). The weights are estimated

from the survey reports and statistics keeping in view the maximum risk in the

evacuation routes. Using these weights, all the concept designs are analysed.

Decision Uncertainty: The parameters related to the decision uncertainty are lift

waiting time tLW and lift transportation time tLT (or stair travelling time tST), number

of evacuees in queue and percentage of aged and disabled persons in evacuation

route. The aged and disabled persons are more prone to the risk of decision

uncertainty (see Chapter 1, Section 1.2). A survey report by Sekizawa et al. (1996)

indicated that 47% residents used lifts and 42% residents used stairs and 7%

residents used both during a fire. A small 4% were shown as others. Splitting this

proportion and the proportion that used both lift and stair, an estimate of 52.5% was

given to the proportion of lift use and the rest to stair use (see Chapter 2, Section

219

2.4). Considering the average number of evacuees for lift evacuation, the required

timings and number of evacuees in queue are determined from ARENA stochastic

evacuation model (see Chapter 4). Literature review also indicated that the

percentage of aged and disabled was 16% (ABS, 2004 and Pauls, 1977). The lift time

period, number of evacuees in queue and percentage of aged and disabled persons

are given as:

� Average tLW and tLT (p1) = 449 seconds

� Average number of evacuees in queue (p2) = 3.51 persons/ floor

� Percentage of aged and disabled persons (p3) = 16% population

Panic: Panic may arise due to unavailability of the evacuation route (lift or stair) and

pre-life threatening condition (visual threat). Unavailability of lift depends upon the

lift malfunctioning due to excessive temperature or electric power failure or lift

maintenance breakdown. However, the evacuees are mainly concerned of their

evacuation safety and are not concerned about the reasons of unavailability of

evacuation routes. A survey was conducted in high-rise apartments with the

objectives of determining comforts of high-rise living (Mori and UHK, 2002) (see

Chapter 4, Section 4.1). In response to a question relating to the disadvantages of

high-rise living, 36% population reported fire escape, 20% reported lift breakdown,

2% reported strong wind, 2% reported heat and 4% reported lack of play areas (see

Figure 7-3). The figure shows the importance of parameters p4 and p5. Other

parameters such as lack of play areas, heat and strong wind are not significant and

are therefore not considered. The 20% population was concerned with the non-

availability of lifts in the apartment building. The 36% population was concerned

with the fire escape. The concern of fire escape occurs due to the presence of visual

threat in the evacuation route (which is ultimately time to exceed the tenable limit for

visual smoke). The evacuees’ concern of non-availability of lift and fire escape is

given in the form of population as:

� Evacuees’ concern for unavailability of lifts (p4) = 20%

� Evacuees’ concern for fire escape (p5) = 36%

220

Disadvantages of High Rise Living

Figure 7-3 – Disadvantages of High Rise Living (Mori and UHK, 2002)

Injuries: The injuries are related directly to three general areas – consequences of

inhalation of toxic products of combustion (smoke, CO, CO2, and other poisonous

gases, hypoxia and asphyxia), exposure to fire (burns, thermal injuries to airways and

incineration), shocks from injuries that precipitate deaths from pre-existing health

conditions (cardiac failure and respiratory diseases) during the exposure of toxic

gases and/or fire (see Chapter 2, Section 2.3, Table 2.6). The fatal injuries are

caused as the strength variables are lower than the load variables. The system is non-

functional and can be determined from safety index (see Chapter 5, Section 5.2.2).

The causes of fatal injuries in residential fires are interpolated as the fatalities caused

by hazardous exposure in evacuation routes. Miller (2005) showed that 28.5% of

evacuees were found dead while attempting to evacuate the buildings, but did not

give the locations of victims. The findings of residential fire deaths are also assumed

for evacuation routes although there may be fewer victims of burns/ incinerations

and more victims of toxic gases in evacuation routes. The following data is taken

from Table 2.6 (Miller, 2005):

� Causes of deaths due to smoke asphyxiant toxic gases, fire and temperature/

incineration (p6) = 174 cases

� Causes of deaths from injuries that precipitate deaths from pre-existing health

conditions (p7) = 11 cases

221

The safety index is used for determining the fatal injuries caused by smoke

asphyxiant toxic gases, fire and temperature/ incineration (the probability of time

period for occupants’ evacuation is less than the probability of time to exceed the

tenability limits).

Parametric Global Weights: Table 7-3 gives the global weights of parameters. The

parametric global weights were obtained from the individual values of the parameters

{for example p1, 449 seconds is 58.84% of 763 seconds – the total waiting and

transportation time for the entire population, similarly for p2, 3.51 persons is 10.96%

of 32 persons – the floor population}. However, the values of parameters from p4 to

p7 have different basis (units) to derive the global weights. The risk priorities along

with group weights for p1 to p7 are indicated in the value tree (see Figure 7-4). A

value tree represents the structural hierarchical position of all the parameters. The

risk priorities were obtained in Chapter 3 from the Analytical Hierarchical Process

(AHP). The weights for parameters p1 to p7 did not share the same basis as the values

are based on evacuation periods, percentage of evacuees, evacuees’ response and

number of injuries (nonfatal and fatal) obtained under different conditions. Under

such conditions, multi-criteria decision approach is the most appropriate method to

determine the group weights and global weights.

Table 7-3: Parametric Values and Weights relating to Building Evacuation

Risk

Category

Parameter

(p)

Individual

Value

Individual

Percent

(%)

Group

Percent

(%)

Group

Weight

Global

Weight

(w)

p1 449 seconds 58.84 68.58 0.6858 0.0051

p2 3.51 persons 10.96 12.77 0.1277 0.0009

Decision uncertainty

p3 16 percent 16 18.65 0.1865 0.0014

p4 20 percent 20 35.7 0.3570 0.0370 Panic

p5 36 percent 36 64.3 0.6430 0.0668

p6 174 cases 94.05 94.05 0.9405 0.8359 Injuries

p7 11 cases 5.95 5.95 0.0595 0.0529

Total 1.0000

.

222

Figure 7-4 – A Value Tree for the Parametric Global Weights

7.3.4 Value Functions

The purpose of introducing value functions is to normalize dimensional risk

parameters with given lower and upper bounds into non-dimensional parameters with

0 to 1 or 0 to 100% scale. The values of the value functions are obtained from

various models and analysis for all the concept design options. The parameters are

multi-dimensional and require deriving a value function. The value functions for p1

to p7 are given below:

Lift Waiting and Transportation Time, and Stair Travelling Time (p1): The

building evacuation times {(tLW + tLT) or tST} were determined from ARENA

stochastic models (see Chapter 4). The parameter strength varies from 338 seconds

Group Weights from statistics, survey, stochastic modelling,

FDS modelling

Priorities from AHP

Matrix

Decision uncertainty

0.0073 Panic

0.1039

Building Evacuation Time

p1 = 0.6858

Evacuees in queue

p2 = 0.1277

Unavailability of the evacuation route

p4 = 0.3570

Injuries (nonfatal/ fatal)

0.8888

Exposure to asphyxiant toxic gases, fire and temperature

p6 = 0.9405

Building evacuation

Visual threat (pre-life threatening condition)

p5 = 0.6430

Aged / disabled persons

p3 = 0.1865

Parametric Global Weights:

p1: 0.0073 × 0.6858 = 0.0051

p2: 0.0073 × 0.1277 = 0.0009

p3: 0.0073 × 0.1865 = 0.0014

p4: 0.1039 × 0.3570 = 0.0370

p5: 0.1039 × 0.6430 = 0.0668

p6: 0.8888 × 0.9405 = 0.8359

p7: 0.8888 × 0.0595 = 0.0529

Total = 1.0000

Pre-existing health condition

p7 = 0.0595

223

to 763 seconds (see Table 7-5). The strength needs to be translated into a value. The

longest time is given a value function of 100 while the minimum time is given a

value function of 0. In mathematical notation, it can be written as:

• v (763) = 100 (A, B and C)

• v (338) = 0 (E-25)

where v(763) means the ‘value function of 763 seconds’, which is given as 100.

Similarly, the shortest time is 338 seconds and is given a value function of 0. This

enables the estimation of the value of (tLW + tLT) or tST of any model between the

most and the least times (100 and 0) on a ratio scale. However, the values are not

necessarily corresponding to arithmetically derived values between the two extremes.

This is further explained with the help of bisection curve (bisection divides the

parametric value equally according to its strength) (see Figure 7-5). There is no

significant decrease in stair travelling time with the reduction of population size. If

the population via stair is reduced to 50% or 25%, the stair travelling time will not be

significantly changed although the population size is varying considerably. The lift

evacuation time for 50% of the population is determined as about 450 seconds

[v(450)=50]. Similarly, the lift evacuation time for 75% of the population is

determined as about 619 seconds [v(619)=75]. The time gives the value function in

Y-axis. Accordingly the graph is skewed upward. The extreme values along with the

midpoint are plotted to determine the value for 367 seconds and 372 seconds. After

earmarking two value functions at the extremes, two value functions are obtained

from the graph. The value functions for 367 seconds and 372 seconds are calculated

as:

• v (367) = 15 (E-75)

• v (372) = 20 (D)

224

0

10

20

30

40

50

60

70

80

90

100

338 363 388 413 438 463 488 513 538 563 588 613 638 663 688 713 738 763

Building Evacuation Time (second)

Value

Bisection value curve

Arthmetic value curve

Figure 7-5 – Value Functions for Building Evacuation Times

Number of Evacuees in Queue (p2): The number of evacuees in queues was

determined from ARENA stochastic evacuation models (see Chapter 4). The

parameter strength varies from 0 to 8.16 persons/ floor (see Table 7-5). The longest

queue containing 8.16 persons/ floor is given a value function of 100. The stairs for

75% of the population has no queue and a value function of 0 is given. In

mathematical notation, it can be written as:

• v (8.16) = 100 (A, B and C)

• v (0) = 0 (E-75)

Lifts for 25% of the population has 1.2 persons/ floor. Stairs for the entire population

has only 0.05 person/ floor. The following value functions for these values are

determined proportionately:

• v (1.2) = 14.7 (E-25)

• v (0.05) = 0.6 (D)

225

Percentage of Aged and Disabled Persons in the Evacuation Route (p3): The

percentage of aged and disabled persons in the apartment building was determined

from the pilot survey (see Chapter 4). The parameter strength varies from 0 to

15.75% (see Table 7-5). Decision uncertainty may vary with the type of evacuation

routes. For example, aged and disabled persons require travelling several steps in

stairs, whereas there are no physical efforts in lift evacuation. Decision uncertainty

would be more in stairs than in lifts. It varies with the types of evacuation routes

such as lifts for aged and disabled persons, lifts for all persons and stairs for all

persons. If lifts are permitted for 25% of the population, lifts may contain most or all

of the aged and disabled persons and almost none in the stairs. Likewise, most or all

of the aged and disabled persons will be using stairs, if lifts are not permitted. The

parameter of aged and disabled persons in evacuation routes is given a risk factor

(see Table 7-4). The correlations among the sub-parameters are based on the Saaty’ 9

point scale (Saaty, 1980), as discussed in Chapter 3. The point scores are based on

judgment.

Table 7-4: Matrix (3 × 3) for Priorities Risk Factor (p3)

Category p3a p3b p3c Priority Risk Factor

Aged and disabled persons in

dedicated lifts (p3a)

1 1/3 1/9 0.077

Aged and disabled persons in lifts

– along with general public (p3b)

3 1 1/3 0.231

Aged and disabled persons in stairs

(p3c)

9 3 1 0.692

The aged and disabled persons can promptly evacuate the building using lifts and

therefore a risk factor of 0.077 (low risk for p3a) is given. The lift system for all

evacuees is given a risk factor of 0.231 (medium risk for p3b) and the stair system is

given a risk factor of 0.692 (maximum risk for p3c). Considering 15.75% as the

percentage of aged and disabled population, the values are assigned 1.21 (15.75 ×

0.077) in dedicated lifts, 3.63 (15.75 × 0.231) in lifts (with general public) and 10.9

(15.75 × 0.692) in stairs. A value of 10.9 for the stair route is given a value function

of 100. No one in the stair (75% population evacuation) is given a value function of

0. Therefore, the following value functions are proportionately given:

226

• v (10.9) = 100 (D)

• v (0) = 0 (E-75)

• v (3.63) = 33.3 (A, B and C)

• v (1.21) = 11.1 (E-25)

Unavailability of Evacuation Route (p4): The lifts may not be available due to

malfunctioning of lift components due to excessive temperature, electric power

failure or maintenance breakdown or manual tripping. The unavailability of lifts was

found to be 26.48% for unprotected lift lobby and 0.48% for single protected lift

lobby and double protected lift lobby (see Chapter 6). The unavailability of stairs

depends upon the crowd density (if the evacuees’ density is more than 3.5 P/m2, the

evacuees can not move) or locking of stair door. Data are not available for the

unavailability of stairs and it is conservatively assumed that the stairs would always

be available in the buildings. The parameter strength varies from 0 to 26.48 (see

Table 7-5). The following value functions are given:

• v (26.48) = 100 (A)

• v (0) = 0 (D and E-75)

• v (0.48) = 1 (B, E-25 and C)

Smoke Visual Threat (p5): The time to exceed the visual smoke obscuration in the

evacuation route was determined by fire hazard modelling (see Chapter 5). The

parameter strengths vary from 241 to 1778 seconds (see Table 7-5). The following

value functions are given:

• v (241) = 100 (A)

• v (1778) = 0 (C)

The maximum time is given a value function of 0 as there is the minimum risk of

visual smoke threat. The minimum time is given a value function of 100 as there is

the maximum risk of visual smoke threat. The following value functions are derived

proportionately:

227

• v (1147) = 42 (E-75)

• v (515) = 82 (D)

• v (500) = 83 (E-25)

• v (388) = 90 (B)

Safety Index (p6): The safety indices were determined from stochastic evacuation

and FDS models (see Chapter 5). The parameter strengths are between -10.82 and 32

(see Table 7-5). The following value functions are given:

• v (-10.82) = 100 (A)

• v (32) = 0 (E-25)

The following value functions are derived proportionately from the extreme values:

• v (3.57) = 66 (D)

• v (6.66) = 59 (E-75)

• v (2.55) = 68.77 (B)

• v (17.61) = 33.6 (C)

Number of Fire Effluents Causing Deaths from Pre-existing Health Conditions

(p7): The deaths from the pre-existing health conditions (mainly among the aged and

disabled persons) were influenced by the presence of fire effluents such as visual

smoke, asphyxiant toxic gases and fire/ temperature. The parameter strengths were

obtained from Chapter 5 (see Tables 5-2 to 5-4) and are given in Table 7-5. The

parameter strengths vary from 0 to 3. The following value functions are given based

on the extreme values:

• v (3) = 100 (A)

• v (0) = 0 (C and E-75)

The following value functions are derived proportionately based on the extreme

values:

228

• v (2) = 66 (B)

• v (1) = 33 (E-25 and D)

Parameter Values: The parameter strengths of p1 to p7 are given in Table 7-5 while

Table 7-6 gives the summary of assigned values.

Table 7-5: Parameter Strengths for Concept Design Options

Concept Design Option

E

Parameter

A B C D

E-25 E-75

p1 (second) 763 763 763 372 338 367

p2 (person) 8.16 8.16 8.16 0.05 1.2 0

p3 (%) 15.75 15.75 15.75 15.75 15.75 0

p4 (%) 26.48 0.48 0.48 0 0.48 0

p5 (second) 241 388 1778 515 500 1147

p6 (number) -10.82 2.55 17.61 3.57 32 6.66

p7 (number) 3 2 0 1 1 0

Table 7-6: Summary of Assigned Values

Concept Design Option

E

Parameter

A B C D

E-25 E-75

p1 100.0 100.0 100.0 20.0 0.0 15.0

p2 100.0 100.0 100.0 0.6 14.7 0.0

p3 33.3 33.3 33.3 100.0 11.1 0.0

p4 100.0 1.0 1.0 0.0 1.0 0.0

p5 100.0 90.0 0.0 82.0 83.0 42.0

p6 100.0 68.8 33.6 66.0 0.0 59.0

p7 100.0 66.0 0.0 33.0 33.0 0.0

Results: The value functions are multiplied by the weights and all the weighted

grades are added to give a final grade of risk. Table 7-7 shows the total risk obtained

for the five models (Options A, B, C, D and E). The risk grade reflects the risk

associated with a particular model.

229

Table 7-7: Total Risk Values for Concept Design Options

Concept Design Option

E = (E-25) + (E-75)

Parameter Global

Weight

(w) A B C D

E-25 E-75 E

p1 0.0051 0.510 0.510 0.510 0.102 0.000 0.077 0.077

p2 0.0009 0.090 0.090 0.090 0.001 0.013 0.000 0.013

p3 0.0014 0.047 0.047 0.047 0.140 0.016 0.000 0.016

p4 0.0370 3.700 0.037 0.037 0.000 0.037 0.000 0.037

p5 0.0668 6.680 6.012 0.000 5.478 5.544 2.806 8.350

p6 0.8359 83.590 57.510 28.086 55.169 0.000 49.318 49.318

p7 0.0529 5.290 3.491 0.000 1.746 1.746 0.000 1.746

Total Risk 1.0000 99.907 67.697 28.770 62.635 7.356 52.200 59.556

7.3.5 Sensitivity Analysis

The results of total risk values given in Table 7-7 show that the proposed use of

protected lift system to evacuate 25% of the population and the remaining population

by stairs (Option E) and lifts with double protected lift lobby (C) involve less risk

when compared with the use of stairs (D). On the basis of the overall risk values

reported in Table 7-7, suitable evacuation schemes can be recommended. To

determine the robustness of results, a sensitivity analysis is conducted against the

weightings used in the risk analysis for panic (w4 and w5). Table 7-8 and Table 7-9

show the value of total risks for the different design options as a function of changes

to the weights used in the risk analysis for panic {w (P4 and P5) = 0.0370, 0.0668}. If

panic is given zero weighting, two parameters (p4 and p5) would have zero

weighting. On the other hand if they are given 100% weighting, the remaining five

parameters (p1, p2, p3, p6, and p7) would have zero weighting.

230

Table 7-8: Total Risk Values from Analyses based on a Zero Weight for Panic

Concept Design Option Parameter Global Weight

(w) A B C D E

p1 0.0051 0.510 0.510 0.510 0.102 0.077

p2 0.0009 0.090 0.090 0.090 0.001 0.013

p3 0.0014 0.047 0.047 0.047 0.140 0.016

p4 0 0.000 0.000 0.000 0.000 0.000

p5 0 0.000 0.000 0.000 0.000 0.000

p6 0.8359 83.590 57.510 28.086 55.169 49.318

p7 0.0529 5.290 3.491 0.000 1.746 1.746

Total Risk 0.8962 89.527 61.648 28.733 57.158 51.170

Table 7-9: Total Risk Values from Analyses based on 100% Weight for Panic

Concept Design Option Parameter Global Weight

(w) A B C D E

p1 0 0.000 0.000 0.000 0.000 0.000

p2 0 0.000 0.000 0.000 0.000 0.000

p3 0 0.000 0.000 0.000 0.000 0.000

p4 0.037 3.7 0.037 0.037 0 0.037

p5 0.0668 6.68 6.012 0.000 5.478 8.350

p6 0 0.000 0.000 0.000 0.000 0.000

p7 0 0.000 0.000 0.000 0.000 0.000

Total Risk 0.1038 10.380 6.049 0.037 5.478 8.387

Figure 7-6 shows how the total risk values for the different design options vary with

changes in the weighting placed on panic. The use of protected lift system to

evacuate 25% of the population and the remaining population by stairs (Option E)

would have total risk values of 51.17 and 8.38 when the value of weight given for

panic was varied from 0 to 100%. Similarly, the concept design option B would have

risk values of 61.64 and 6.04 and the concept design option D would have risk values

of 57.15 and 5.47 at the two extremes. Comparing these risk values, it can be seen

that the safe design option E has changed to option D at 62 (equivalent risk) for the

case of the highest weight given to panic (100%). The safe design option E is

sensitive to the slight variation in the proportion of building population. The most

safe evacuation concept design option (Option C) is not affected by the change. The

risk value for unprotected lift lobby (Option A) remains relatively higher than those

for stairs (Option D) despite the change of weight values for panic.

231

Concept Design Option Weight Given

for Panic Resultant Weight w

(%) A B C D E

0 89.62 89.527 61.648 28.733 57.158 51.170

100 10.38 10.380 6.049 0.037 5.478 8.387

Figure 7-6 – Sensitivity Analysis for Different Weights Placed on Panic

7.4 Analysis of Results

The overall risk values calculated for the five design options are shown in Figure 7-7.

If the stair alone evacuation (Option D) is considered as an acceptable evacuation

design, then the lift with simply protected lobby for one-fourth of the building

population (Option E) and double protected lift lobby (Option C) provide acceptable

alternatives since the latter two have lower associated risk than the former. The

acceptable risk RA is shown in Figure 7.7 (RA = 62.635). Lifts with unprotected

lobby (Option A) has the highest risk, followed by lifts with simply protected lobby

232

for entire population evacuation (Option B). The results also show that the risks are

considerably reduced by providing protected lift lobby. A combination of lifts and

stairs provide a lower level of risks in comparison to stair alone evacuation (Option

D) or lift alone evacuation with the same lobby protection (Option B).

0 20 40 60 80 100

A

B

D

E

C

Risk Value

Concept Design Option

RA

Figure 7-7 – Risk Values for Concept Design Options

7.5 Conclusion

This chapter has demonstrated that the MODA method can provide a rational basis for

determining the risks. Multiple risk parameters in relation to human behavioural

response, fire hazard exposure and reliability of lift operational mechanism were

included in the assessment. This method was applied to all design options considered

and the out come of risk values were meaningful only on relative terms. The

imperfection in the selection of parameters might influence the absolute risk values

but would not significantly alter the relative scales of risk associated with alternative

design options. The results of the assessment are summarized as follows:

• Lifts with unprotected lobby contribute to the maximum for risks.

233

• Lift with protected lobby does not provide adequate safety in comparison to

stair for the entire building population. Overall risks of lifts with protected

lobby were slightly higher than the risks of stair system.

• A combined system of lifts with protected lobby for one-fourth of population

and stair for three-fourth of population can provide adequate safety in

comparison to individual system alone (lifts or stairs).

• Lifts with double protected lobby provides a better performance.

The MODA method provides a versatile means for risk assessments. It is based on a

comparative study of multiple options and incorporates multiple risk attributes into

the evaluation. The MODA method involves the ranking of level of importance for

multiple risk attributes. This ranking is still by and large empirical and/or subjective.

In its application to fire risk assessment, the method is linked to other means of

evaluations such as the ASET/RSET analysis, stochastic modeling and safety index

analysis. This chapter did not include the cost-effectiveness analysis in the risk

assessment, although in theory such an inclusion is achievable.

234

8. FEASIBILITY AND DESIGN CONSIDERATIONS

8.1 Introduction

The traditional prescriptive approach to building design embodied in the building

codes is based on various norms and guidelines. Experience based traditional fire

regulations may not be always adequate and therefore alternative solutions are put

forward. Such alternative solutions must be adequately investigated if uncertainties

exist. The performance-based fire safety designs are used in many parts of the world,

which are also acceptable under the provisions of the Building Code of Australia

(BCA). The objective of BCA is to safeguard occupants from illness, injury or death

during building evacuations. A building is to be provided with safeguards so that the

occupants have sufficient time to safely evacuate before the environment in any

evacuation route becomes hazardous due to fire. At one of the fire safety

conferences, a NSW Fire Brigade officer stated that the NSW Fire Brigade is

supportive of innovative designs that provide equal evacuation opportunities for all

the occupants (Honeybrook, 2002). The use of lifts for building evacuation during

fire emergencies goes against the “norm”, and therefore requires consideration of a

range of factors, and confirmation and demonstration that all the concerns and safety

factors are addressed.

The design of lift systems plays a vital role in the safety of lift evacuation system.

This research project was undertaken by considering the uncertainties of parameters

relating to evacuation procedure (human movement and behavioural response) and

variable conditions in fire development (wind and stack effects) in order to determine

feasible, safe and effective design solutions. The design alternatives were

investigated for a limited number of hypothetical fire scenarios. The results were

evaluated and the following two options were determined for lift evacuation as

discussed in Chapters 4 to 7:

• Lifts with protected lobby to evacuate one-fourth (25%) of the building

population and stairs for the rest of the population (75%)

• Lifts with double protected lobby to evacuate the entire population

235

This chapter shows the feasibility of these options as alternative evacuation systems

and proposed various redundancy measures to enhance the occupant safety during

evacuation.

8.2 Feasibility Options

8.2.1 Lifts with Protected Lobby to evacuate 25% of the Building Population

Protected lift lobbies can be managed for emergency evacuation efficiently and

effectively for one-fourth of the building population. Able-bodied people can use the

stairs for evacuation. The statistics showed that the percentage of aged and disabled

persons in apartment buildings is 16%, which can be managed using the lift system

(see Figure 8-1).

Figure 8-1 – Evacuation Option 1: One-Fourth of the Building Population using

Protected Lifts and the rest using Stairs

236

The Occupant Emergency Evacuation Plans are prepared for high-rise apartment

buildings, which must incorporate predefined means of evacuation facilities for the

building occupants. It should clearly indicate that the lifts are to be used by a fraction

of population such as the aged and disabled persons. The sign posted in front of lifts

“DO NOT USE LIFT IN CASE OF FIRE” can be replaced with “DO NOT USE

LIFT IN CASE OF FIRE, LIFTS ARE PERMITTED ONLY FOR ASSISTED

EVACUATION [AGED AND FRAIL PERSONS]”. There will be a necessity to

develop and evaluate a strategy for public education to ensure the effectiveness of the

proposed lift evacuation system. Without proper preparation and training (fire drills),

evacuees may become fearful of the dangerous conditions as uncertainties may arise

in relation to the proportion of population to be evacuated by lifts.

8.2.2 Double Protected Lift Lobby for the Entire Building Population

As shown in previous chapters, evacuation via double protected lift lobby provides is

safer than evacuation using stair systems. Details of the provision of double protected

lift lobby (conceptual design option C) are shown in Figure 8-2.

Figure 8-2 – Evacuation Option 2: Double Protected Lift Lobby

The BCA includes DTS provisions relating to travel via fire-isolated exits. The

requirements for fire-isolated exits typically require a smoke lobby, if more than two

FD

FD LIFT LOBBY

EX

IT

SD

LIFT

EX

IT

SD4

FD – Fire Door SD – Smoke Door

237

access doorways are open to a required fire-isolated exit in the same storey. In the

absence of pressurisation of lift shaft, a protected lift lobby can not be used as smoke

lobby as it permits smoke leakage to the lift shaft. For creating a smoke lobby in the

fire-isolated exit (lift system), one set of doors can be self-closing fire door while the

other can be a magnetically held open door or a normal smoke door. Smoke and hot

gases can be restricted to the fire-affected apartment unit (or corridor), i.e. away from

the lift shafts. The disadvantages of providing a double protected lift lobby are the

additional space requirement for smoke lobby between the two doors (additional

space is approximately 1.5 times more than that for the stair lobby). Evacuees may

also feel unease due to the double door opening and closing during their regular use

of lifts. Further, additional space is required in the lift lobby for accommodating the

entire population.

8.3 Redundancy Measures

This research has demonstrated the feasibility of the proposed two options for lift

evacuation. In this section, a number of redundancy measures are proposed that are

likely to enhance the reliability of proposed lift evacuation systems.

8.3.1 Common Lift and Stair Lobby

There are advantages in high-rise apartment buildings to provide access within the

same protected lobby to both stairs and lifts designed for operation during fire

emergencies. Building codes require fire compartmentation by providing a lift lobby

between the lift shaft and risk area and such lifts are often used as emergency or fire

fighter lifts. The conventional rectangular arrangement of lifts may optimize the

space requirement and allows common space to be utilised for both lift and stair

systems (see Figure 8-3).

238

Figure 8-3 – Rectangular Arrangement for a Common Lift Lobby

The scissor type stairs also satisfy the BCA DTS requirements as there is double

door protection (smoke lobby). With the provision of double door, pressurisation is

not a necessity (ABCB, 2005) and it will not cause any concern to lift door opening

and closing. The provision can be useful for lift evacuation. Stair is also

approachable as a redundancy measure for lifts. The provision can be used as fire

brigade staging area (as it provides an alternative evacuation facility). Other safety

features may also be incorporated into the system such as:

(a) a monitor in each lift lobby to indicate where the lifts are at any one time;

the time when the next lift will come; whether it is still safe to wait for the

lift and if one should move into the stairs; and

(b) the provision of an expanded landing to allow disabled persons to wait

there if the lifts cannot be used.

8.3.2 Refuge Area

Presently two types of systems are used for the aged and disabled persons, i.e.

staging area and horizontal separation. The staging areas (refuge area) are intended

as spaces in which people with disabilities can safely wait during a fire. Horizontal

separation consists of one or more barriers which divide a floor into separate areas

with the intent of restricting smoke and fire spread (compartments or landings in the

FD

FD LIFT LOBBY

EX

IT

EX

IT

SD

LIFT

LIFT

SCISSOR STAIRS

FD – Fire Door SD – Smoke Door

SD

239

stairwell). The refuge area provided in the building can be utilized as a redundancy

measure (see Figure 8-4). In this case, evacuees will have a choice to exit the

building using stairs or lifts or move to refuge area.

Figure 8-4 – Lifts located in the Refuge Area

8.3.3 Scattered Design of Lift System

During heavy traffic conditions, there are many hall calls to serve and the lifts have a

tendency to move side by side, which is called bunching of lifts. Lift stops at the

nearest floor. However, this phenomenon can be removed by giving priorities for the

long or timeout hall calls using electronic controls (Siikonen, 1997).

Lifts can be used effectively by scatter design lifts. This strategy involves

distribution of evacuees to different locations in the building for evacuation so that

the evacuees would not wait longer for lifts at one strategic location. Further, there

would be less door opening and closing due to which unsafe conditions do not arrive

quickly. This design provides a longer duration of tenable environment in the lift

lobby.

FD

FD LIFT LOBBY

EX

IT

EXIT

SD

FD

SD

REFUGE FLOOR

LIFT

FD – Fire Door SD – Smoke Door

240

8.3.4 Pressurization

The BCA does not mandate lift shaft pressurization for apartment buildings.

Sufficient pressure difference across the lift lobby and corridor can be maintained to

restrict the smoke spread in the unprotected and protected lift shafts. Pressure can be

maintained in the strategic location in the lift shaft or in the lift lobby of the fire-

affected floor as the occupants will be waiting there. The basic theory of this smoke

control method system may be similar to that of stair pressurisation which increases

the pressure inside by supplying sufficient amount of air to it. However, the

maximum pressure should not hinder the lift door opening and closing. The

pressurization system is generally interlinked with the smoke alarm system. Lift

landing door frequently opens and closes, hence the total leakage area is the leakage

area of both the lift landing door opening and closing (at the time of service) and the

leakage area between the lift landing door and lift landing wall frame at all the floors.

BS 5588, Part 4, Chapter 5 gives the required air flow for pressurised space as:

2/1827.0 EEE PAQ =

where

QE is the air supply to the pressurized space (m3/s)

AE is the total leakage area out of the space (m2)

PE is the pressurisation level in the pressurized space (Pa)

8.3.5 Smoke Seal in Lift Landing Door

If the leakage area of the lift doors is reduced with the help of smoke seal (or gasket)

then tenable conditions can be maintained in the lift shaft and upper floors. The

smoke seal can fit in between the lift landing door and lift landing wall frame (see

Figure 8-5). Warnock Hersey International Inc. (WHI, 1987) conducted a test to

determine the resistance to air leakage of brush type gasketing for lift landing doors.

The test was conducted after imparting 100 000 cycles to the gasket and the air

leakage was measured at several differential pressures at room temperature and at an

elevated temperature of over 400°F (204°C). The application of gasket showed at

241

least a 95% decrease in leakage rates. However, this option is applicable if the lifts

are provided in the unprotected lift shaft and the lifts are not served at the fire-

affected floor. Lift operating software is to be programmed with fire detection

system, however, this aspect was not analysed in this research.

Figure 8-5 – Smoke Seal in the Lift Landing Door and Wall Frame

8.4 Fire Protection Measures for Lift System

The fire protection systems in the lift system are provided to combat fires in lift

machine room and lift shafts involving lift equipment. During the fires in this system,

lift mechanism will not be operative and therefore it can not be used by the building

occupants. However, if the lift mechanism is still operative, safety provisions must

be made independently to cut off the lift system from regular operations after

returning the lift car to the designated landing (ground floor). A fire occurrence in

the lift machine room at the top may not necessarily require a building evacuation. A

fire occurrence in the lift shaft would give smoke and hot gases plume in the lift shaft

with leakages to the floors, which may require building evacuation. Primary route of

building evacuation (stair) is available in this case.

242

Automatic fire sprinkler protection is required in the lift machine room, and in the

top and bottom of lift shaft (NFPA 13, 1996). However, automatic fire sprinkler

protection is exempted in the machine rooms of traction-type lifts when the rooms

are located on the top of the lift shaft and are separated from other areas of the

building, other than the shaft, by not less than one-hour fire resistive occupancy

separation. The exception of the top and bottom of lift shaft is also provided, if the

lift shaft is of non-combustible construction and does not contain combustible

hydraulic fluid (NFPA 13, 1996). The system is provided with at least one smoke

detector and is designed in a separate and independent sprinkler branch.

8.5 Strategic Planning

The successful emergency evacuation of a building requires comprehensive

management procedures. Groner (2002) quoted - “Engineering the hardware to

prevent intrusions of heat, smoke and water is not the primary obstacle. The greater

challenge is in real world use of these systems-in the strategic planning, interface

design and operator training that will enable the safe, effective and efficient use of

elevators during emergencies”. If lifts are considered for emergency evacuation,

regular drills are required to be conducted to inculcate a sense of confidence among

the occupants. Building Fire Safety Regulation may incorporate the provisions of

conducting regular fire drills for the residents of apartment buildings. For a strategic

planning of partial lift evacuation, permanent residents can be taught the proper use

of stairs and lifts to avoid ambiguity in the evacuation procedure.

8.6 Conclusion

This chapter has demonstrated the feasibility of the two options of lift evacuation

systems. It has been shown that the lift system can be provided for evacuating part of

the population (25%) without reducing the level of fire safety. For the evacuation of

the entire population, lift systems need special design considerations such as double

door protection. The lifts require adherence of rules and norms relating to regular

electrical supply and precautionary measures for water spread. However, redundancy

measures can provide additional safety to the lift system. If lifts become a total or

243

partial means of alternative evacuation facility, the stair capacity should not be

reduced. Lifts are the mechanical means of transportation and require regular

maintenance. Lifts can not be a substitute for a non-mechanical and a non-

maintainable means of stair system. The stairs are required to maintain through an

effective and graded program of fire prevention inspections. The main thrust can be

diversified to discover and eliminate the fire hazards in the evacuation routes such as

stair and lift systems in the buildings.

Current building codes and regulations provide limited aspects of lift installation.

They are mainly related to the general requirements for normal lifts and certain

provisions for people with disabilities. If lifts are used as evacuation systems during

fire emergencies as proposed in this research, their design and installation in

apartment buildings deserve a special consideration as part of the overall design

process.

244

9. CONCLUSIONS

9.1 Summary

The main objective of this research was to study if lifts provide an acceptable means

for evacuation in apartment buildings during fire emergencies. Building evacuation

models, fire hazard models and risk assessment models have been developed for

systematic and in-depth analysis of lift operations. They were applied to a number of

hypothetical case studies for uncertain and complex factors. A brief summary of

their features, applications and evaluations are given as follows:

• A research methodology was developed. Risks associated with the lift

evacuation system were identified and risk cause-effect relationship was

established. The issues of human behavioural response, fire hazards and lift

operational mechanism were identified. Risk priorities were identified with

the help of Analytical Hierarchical Process. Considering that panic may be a

rare event and the outcome of the study of this phenomenon is inclusive, a

relatively low risk priority was give to this parameter in the risk analysis. A

research strategy was developed by reducing the level of consequences to an

acceptable level. Design options and evacuation strategies for lifts were

proposed for comparison with stairs.

• A pilot survey predicted the residents’ awareness level of emergency

evacuation procedure and their inclination toward the use of lifts. It also

demonstrated the necessity of alternative evacuation facility in high-rise

apartment buildings. Interviews also showed the concerns and controversy of

the unresolved issue (use of lifts for emergency evacuation).

• A stochastic simulation model was developed for the assessment of

parameters relating to lift evacuation system. The parameters included

relevant lift time periods (waiting, transportation, pre-evacuation and

evacuation times) and the number of evacuees in queues at various floor

levels. The model was developed for a 38 storey hypothetical building.

245

Results of hypothetical scenarios for the evacuation of the entire and one-

fourth (25%) of the building population using lifts under uncertainties were

generated. The output variables for one-fourth of the population were

determined for evaluating a safe and efficient lift evacuation strategy. Output

variables were plotted for obtaining the mean and standard deviation. The

simulation model was useful to comprehensively evaluate the risks relating to

human behavioural response and life safety in the lift system. It incorporated

effects of different conditions (social, physiological and psychological

characteristics, temporal and a priori heuristics of the lift domain), which may

vary during the emergency evacuations. The modeling results have provided

a base for evaluating relevant lift parameters for an acceptable level of safety

(for example, minimal human behavioural or life safety concerns during lift

evacuation by a smaller population).

• A stochastic model for stair system (similar to lift stochastic model) was also

developed. The parameters included stair time periods and number of

evacuees in queues at various floor levels. The stair simulation model was

useful for comparing the results with lift system. The output variables were

determined for the evacuation of the entire and three-fourth of the population

(75%) using stairs. The results from the stochastic models of lift and stair

systems revealed that the lift evacuation time for one-fourth of the building

population was within the acceptable limit of stair evacuation time. The

number of evacuees in queue for lifts was less than 2 persons. This showed

that the human behavioural response such as decision uncertainty would be

minimal during the evacuation of one-fourth of the population using the lift

system.

• Models of risks associated with fire and smoke hazards in evacuation routes

were proposed for estimating the smoke, asphyxiant toxic and heat exposure

hazards. The time, concentration and toxicity of the fire effluents were

considered in combination for predicting the probable time to incapacitation

or death. A 38-storey hypothetical building (used for stochastic evacuation

modelling) was analysed for fire hazards. Twenty four fire scenarios were

analysed after incorporating stochastic uncertainties relating to wind speed

246

and vertical location. Three compartmentation strategies along with

additional evacuation strategies were considered. The fire profile was

characterised in terms of smoke extinction coefficient (visibility), asphyxiant

gas CO concentration, CO2 concentration, O2 concentration (low oxygen

causing hypoxia), temperature and radiant heat flux. The output of the design

fires provided the initial conditions for a smoke movement analysis, including

the time when the lifts and stairs became unsafe and unusable by the

evacuees. These models calculated the incapacitating dose of fire effluents in

the lift and stair systems. The concept of fractional effective doses of safe

criterion (one-tenth of incapacitation) was applied for calculating the safety

index. Critical times for unsafe conditions after which evacuees could not use

lifts for emergency evacuation were determined. The influences of stack

effect and wind speed were also analysed. This study provided useful

information in the identification of safe lift evacuation systems. The use of

lifts with a protected or double protected lobby gave positive safety indices

whereas those with an unprotected lobby did not give a positive safety index,

indicating the greater safety of the first two options.

• Probabilistic risk assessment models were proposed for evaluating risks

associated with water spread. The development of these models was based on

the traditional probabilistic event tree method (a complex parallel and series

combination). After quantification of water spread, probabilities of water

spread were derived. It was determined that a copious amount of water

generated from fire fighting measures could damage the lift components. The

maximum quantity of water spread occurred from fire brigade intervention.

Unprotected lift lobbies are more likely to suffer from water damage whereas

protected lift lobbies can provide a barrier to water spread.

The spread of hot gases to the lift machine room (LMR) via unprotected lift

lobby increases the probability of lift malfunctioning. In unprotected lift

lobbies, temperature rise in LMR depends upon the location of fire. Protected

lift lobbies provide adequate safety against temperature rise in LMR.

247

Probabilistic risk assessment models were also proposed for evaluating risks

associated with electric power systems. The development of this method was

based on a traditional fault tree analysis. After quantification of electric

power failure issues for the hypothetical building, probabilities of risks were

derived. The reliabilities of lift design options were obtained. It was

determined that the probability of electric power failure due to fire in a

residential unit is small. Fire can disrupt electrical system if fire and hot gases

spread to unprotected lift lobbies. Electrical system adhering with relevant

codes and standards in protected lift lobbies can provide adequate safety.

• This research has introduced the concept of Decision Analysis for risk

assessment in the lift evacuation system. An integrated risk assessment

approach was developed to include the uncertainties associated with issues

relating to human behavioural response, fire hazards and lift operational

mechanism. This development was based on the ‘Multi-Objectives Decision

Analysis’ approach for the assessment of results achieved from stochastic

evacuation models, deterministic fire models, probabilistic event tree and

fault tree analyses. This study provided useful information for identifying the

overall safe evacuation system.

• The research provided a good background to the feasibility and design

considerations and evacuation strategies for safe and efficient lift operations.

Design and technical redundancy measures such as refuge area, scatter design

of lifts, pressurisation, gasket between lift landing door and wall frame were

proposed. An overview of fire protection systems for the lift system was also

presented.

9.2 Research Findings

This research has concluded that there are two options if lifts are considered for

emergency evacuation in apartment buildings. Lifts with protected lobby can be used

in the evacuation of one-fourth of the building population without reducing the level

of fire safety. For considering the evacuation of the entire population, lift system

needs special design considerations such as double door protection (or lifts protected

248

with double lobby). The performance based lift evacuation system is achievable. A

brief of advancements and contributions made to the field of research are given next.

9.2.1 Advancements in Systematic and In-Depth Risk Analysis

This research work presents a distinctively different approach over traditional

methods to risk assessment by:

(i) Incorporating related social, physical, temporal and a priori heuristics of

the lift domain for getting insight into lift evacuation system (with

downward, upward and inter-floor movement); and

(ii) Integration of issues relating to human behavioural response, fire hazards

and lift operational mechanism for risk assessment and reliability; and

(iii) Effective quantification of system uncertainties using statistical and

probabilistic techniques. Specifically, the proposed methods (including

MODA and AHP) could reasonably advance methodologies of risk

analysis and assessment for effectively addressing critical issues of

emergency lift operations.

Animation models based on the proposed methods are developed for resolving

obstacles before any lift evacuation plan becomes reality.

9.2.2 Contribution to Building Evacuation Strategy

This research work is an extension of earlier research work contributed by Klote

(1982). A framework for resolving the issues of lift system has been developed. The

concept of MODA is used for the first time in the field of fire safety in this research

project.

This research has resolved that the risk of protected lift lobby is within the acceptable

limit for evacuating a reduced population. Therefore the protected lift system is

recommended with the intent of providing an alternative evacuation facility for the

aged and disabled persons. It was determined that partial evacuation is possible with

the existing infrastructure of protected lift systems. With the current deemed-to-

249

satisfy infrastructure in the apartment buildings, the lift system can be used as a

solution alternative to stairs.

9.3 Scope for Future Work

Many variables have been considered to evaluate the performance of the lift

evacuation system during fire emergencies. These variables were subjected to

uncertainties. To obtain the best results, data were collected to reduce the uncertainty

in relevant variables (for example, reliability of fire protective measures, walking

speed of occupants, coping and response of occupants). These data were simplified

into statistical frequency distribution for the variables. Some factors are still not

possible to describe numerically and, therefore, have to be treated by other means for

example by judgment (such as relationship between the human behavioural response

issues – decision uncertainty and panic in Chapter 3). Therefore the results presented

in this thesis showed the general trend in the safety level. Future work should be

focused to establish a relationship of the human behavioural issues. More surveys are

required to be conducted to evaluate the response of residents toward the lift

evacuation system. The parameters relating to building features of fire safety and fire

protection systems and the level of the fire-affected floor can also be added in the

risk assessment using the Multi-Objectives Decision Analysis approach.

There are certain limitations in the computer model (FDS) used in this research. The

wind speed could not be increased above 7.43 m/s (25 kmph). This is because the

building size is large, which requires a multi-grid system. The information from one

mesh is not properly interpreted at the exterior boundary of a given mesh. This

causes a numerical instability during wind conditions and therefore wind speeds

higher than 7.43 m/s could not be modelled. A rational approach of single-grid

system is needed to study the parameter of higher wind speeds in detail.

This research has attempted to develop a rational procedure for the evaluation of lift

systems as an emergency evacuation facility in high-rise apartment buildings. The

efficacy of the proposed lifts-stairs evacuation strategy (such as 25% of the

population using lifts and for the rest using stairs) has to be evaluated in apartment

buildings during fire drills. Lifts protected with double lobby can be used for the

250

entire population. Double protection of lift shaft can be compared jointly with single

protection of lift shaft (lift protected with a lobby) and lift shaft pressurisation. In

regard to lift shaft pressurisation, smoke does not filter in the FDS model due to

positive pressure in the lift shaft. Therefore, tests relating to the efficacy of lift shaft

pressurisation and leakage areas should be conducted in apartment buildings in the

future.

The apartment buildings are provided with the simple lift group controller system.

The research project begins with the simple lift group controller system and can be

extended to the advanced lift group controller system to achieve a better performance

of the lift system. Further, the cost effectiveness analysis can also be conducted for

various designs of lift evacuation strategies.

The work relating to the use of lifts during fire emergencies is still in its initial stage.

This research concentrated on the use of lifts as an alternative evacuation facility in

apartment buildings. It can be extended in the future to other types of high-rise

buildings such as hotel, hospital and office buildings.

“It can be envisaged that the use of lifts will be inevitable for super

high-rise structures”

251

REFERENCES

ABCB (2005), “The Building Code of Australia – Volume 1”, 2005, Australian

Building Codes Board, Canberra, Australia

ABCB (2005b), “International Fire Engineering Guidelines”, Edition 2005,

Australian Building Codes Board, Canberra, Australia

ABS (2004), “Housing and Lifestyle: High Rise Living”, Australian Bureau of

Statistics, Report No. 4102.0

ABS (2005), “Australian Social Trends - Supply of Housing”, Australian Bureau of

Statistics, Report No. 4102.0

AFAC (1997), “Fire Brigade Intervention Model”, Australasian Fire Authorities

Council, Version 2.1, November, Australia

Aherns, M. (1998), “Batteries Not Included”, NFPA Journal, May/ June, pp. 99-109

Alexandris, N.A. (1977), “Statistical Models in Lift System”, Ph.D. Thesis,

University of Manchester Institute of Science and Technology, pp. 252

Allen, T. H. (1995), “Elevator Evacuation”, Building Performance Research

Institute, USA

Aloi, S. and Rogers, J. (2002), “Reach of the Sky”, Fire Prevention & Fire Engineers

Journal, Volume 62, No. 219, FPA, London, UK

Andersson, C., and Wadensten, K. (2000), “Elevator Evacuation of a High-rise

Building”, Ove Arup and Partners

AS 1735.1 (2003), “Lifts, Escalators and Moving Walks – General requirements”,

Australian Standards, Sydney, Australia

AS 1221 (1997), “Fire Hose Reels”, Australian Standards, Sydney, Australia

ASME A17.1 (2000), “Safety Code for Elevators & Escalators Handbook”, ASME,

USA

Barnett, C., Sfintesco, D., Charles Scawthorn, C. and Zicherman, J. (1992), “Fire

Safety in Tall Building”, Council on Tall Buildings and Urban Habitat

Committee 8A, pp. 3-8

Baum, H.R., and MaCaffery, B.J. (1989), “Fire Induced Flow Field – Theory and

Experiment”, Fire Safety Science – Proceedings of the Second International

Symposium, pp. 129-148

252

Beever, P., and Davy, B. (1999), “A Rational Approach to Firefighting Water

Supplies”, AFAC Conference, Australia

Bell, J. R. (1983), “Five Die in Rockefeller Park Towers Fire, Cleveland Ohio”, Fire

Journal, 77, March, pp. 63-65

Bennetts, I.D., Moinuddin, K.A.M., Goh, C.C., and Thomas, I.R. (1999), “Testing

and Factors Relevant to the Evaluation of the Structural Adequacy of Steel

Members within Fire-resistant Elevator Shafts”, Centre for Environmental

Safety and Risk Engineering, Victoria University of Technology, Melbourne

City MC, Australia

Beyler, C.L. (2004), “Overview of Tenability Criteria and Analysis Methods”, NFPA

Fall Technical Meeting, NFPA, USA

Bill, M. (2002), “World Trade Center Disaster: Initial Response”, Fire Engineering,

UK, September

Bounagui, A., Benichou, N., McCartney, C. and Kashef, A. (2004), “Grid

Optimization for the Full-Scale Test Facility to Evaluate the Fire Performance

fo Houses – Part I – Basement Fires”, Institute for Research in Construction,

National Research Council Canada, Ottawa, Canada

Brennan, P. (1999), “Successful Evacuation in Smoke: Good Luck, Good Health or

Good Management?” Interflam 1999, Proceedings of the 8th International Fire

Science and Engineering Conference, pp. 697-706

Brennan, P., and Thomas, I. (2001), “Victims of Fire, Predicting Outcomes in

Residential Fires”, In Human Behaviour in Fire: Proceedings of the 2nd

International Symposium, Boston. London: Inter Science Communications

Bryan, J.L. (1982), “Human behavior in the MGM Grand Hotel fire”, Fire Journal,

Vol. 76, 2 March, pp. 37-48

Bryan, J.L. (1983), “An Examination and Analysis of the Dynamics of the Human

Behaviour in the Westchase Hilton Hotel Fire”, National Fire Protection

Association, Quincy, USA

Buchanan, A.H. (2001), “Fire Engineering Design Guide”, 2nd

Edition, University of

Canterbury, New Zealand

Building Fire Safety Regulation (1991), “Building Act 1975, Fire and Rescue

Service Act 1990”, Queensland, Australia

Bukowski, R. W., Budnick, E. K. and Schemel, C. F. (1999), “Estimates of the

Operational Reliability of Fire Protection Systems”, International Conference

253

on Fire Research and Engineering (ICFRE3), 3rd Proceedings, SFPE, NIST

and IAFSS, October 4-8, 1999, Chicago, pp. 87-89

Butchen, K. (1995), “Relationship for Smoke Control Calculations: Technical

Memoranda”, TM 19, The Chartered Institution of Building Services

Engineers, USA

Canter, D., (1990), “Fire & Human Behaviour”, 2nd Edition, David Fulton Publishers

Ltd, UK, pp. 66

Caldwell, J. (1996), “Standby Power and Elevator Safety in Fire Emergencies”,

Elevator World Journal, on web

(http://www.elevator-world.com/magazine/archive01/9602-002.htm)

CEN (2003), “Safety Rules for the construction and Installation of Lifts - Part 72:

Fire-fighter Lifts”, CEN/TC10, Committee for European Standardization,

Brussels

Clemen, R.T. (1996), “Making Hard Decisions: An Introduction to Decision

Analysis”, 2nd Edition, Belmond, California – Duxbury Press

Cornell, C.A. (1969), “A Probability-Based Structural Code”, ACI Journal, Volume

66(12), pp. 974-985

Cortes, P., Larraneta, J., and Onieva, L. (2004), “Genetic Algorithm for Controllers

in Elevator Groups: Analysis and Simulation during Lunch Peak Traffic”,

Science Direct, Applied Soft Computing 4, pp. 159-174

Dalgliesh, W.A. and Boyd, D.W. (1962), “Wind on Buildings”, CBD-28, Canadian

Building Digest, Institute of Research in Construction, National Research

Council, Canada

Davis, S. (2000), “Fire Fighting Water: A Review of Fire Fighting Requirements – A

New Zealand Perspective”, Fire Engineering Research Report, ISSN 1173-

5996, University of Canterbury, New Zealand

Degenkolb, J. (1991), “Elevator Use During a Building Fire”, ASME Symposium on

Fire and Elevators, Baltimore, MD, pp. 76-79

Donegan, H.A. (2002), “Decision Analysis”, The SFPE Handbook of Fire Protection

Engineering, 3rd Edition, pp. 5:8-23

Duncanson, M., Woodward, A., Reid, P., and Langley, J. (2000), “Unintentional

House Fire Deaths in New Zealand 1991-1998”, New Zealand Public Health

Report, Vol. 7 No. 7/8, July/August 2000, pp. 31-34

254

Dunlop, K.E., Shields, T.J. and Silcock, G.W.H. (1996), “Toward the Quantification

of Emergency Egress Capabilities for Disabled People”, Fire Engineering and

Emergency Planning, E & FN Spon Publisher, pp. 154-161

Dunstan, K. and Thomson, N. (2006), “Demographic Aspects of New Zealand’s

Ageing Population”, Statistics New Zealand, Ministry of Social Development,

New Zealand pp. 1-44

Edwards, W. (1971), “Social Utilities”, Decision and Risk Analysis Powerful New

Tools for Management, The Engineering Economist – Summer Symposium

Series VI, American Society for Engineering Education, pp.119-129

Emporis (2005), “Emporis Buildings”, on web

(http://www.emporis.com/en/wm/bu/?id=101921)

FCRC (1996), “Fire Engineering Guidelines”, First Edition, Fire Code Reform

Centre, 1st Edition, Australia

Frantzich, H. (1997a), “Fire Safety Risk Analysis of a Health Care Facility”, Report

3085, Department of Fire Safety Engineering, Lund Institute of Technology,

Lund University, pp. 1-68

Frantzich, H. (1997b), “Fire Safety Risk Analysis of a Hotel”, Department of Fire

Safety Engineering, Lund Institute of Technology, Lund University, pp. 1-28

Frantzich, H. (1997c), “Design Based on Calculated Risk”, on web

(http://www.brand./th.se/englsh/research/risk.pdf)

Frantzich, H., Magnusson, S.E., Holmquist, B., and Ryden, J. (1997), “Derivation of

Partial Safety Factors for Fire Safety Evaluation Using the Reliability Index β

Method”, Fire Safety Science – Proceeding of the Fifth International

Symposium, Melbourne, Australia, pp. 667-678

Groner, N. (2002), “A compelling case for emergency elevator system”, Fire

Engineering, 155:10, October, pp. 126-128

Hakonen, H. (2003), “Simulation of Building Traffic and Evacuation by Elevators”,

Licentiate Thesis, Helsinki University of Technology, Department of

Engineering Physic and Mathematics, System Analysis Laboratory, Finland

Hall, J.R. (2001), “High-Rise Building Fires”, NFPA Fire Analysis and Research

Division, USA

Hartzell, G.E., Priest, D.N., Switzer, W.G. (1985), “Modeling of Toxicological

Effects of Fire Gases: II. Mathematical modeling of intoxication of rats by

carbon monoxide and hydrogen cyanide”, Fire Science, 3: pp. 115-28

255

Hartzell, G.E. (2001), “Engineering Analysis of Hazards to Life Safety in Fires: the

Fire Effluent Toxicity Component”, Safety Science, 38: pp. 147-155

Hasofer, A.M. and Lind, N.C. (1974), “An Exact and Invariant First Order

Reliability Format”, ASCE, J. Eng. Mech. Division

He, Y., Horason, M., Taylor, P., and Ramsay, C. and Lai, D. (2001), “A Probabilistic

Fire Safety Engineering Assessment of a Refurbished High Rise Office

Building”, Proceedings of the International Conference on Engineered Fire

Protection Design, Society of Fire Protection Engineers, San Francisco, pp.

211-228

He, Y., Horason, M., Taylor, P., and Ramsay, C. (2003), “Stochastic Modelling for

Risk Assessment”, Fire Safety Science – Proceedings of the Seventh

International Symposium, International Association for Fire Safety Science

(IAFSS), Worcester, MA, USA, Intl. Assoc. for Fire Safety Science, Boston,

MA, Evans, D. D., Editor, pp. 333-344

Honeybrook, J. (2002), Superintendent, NSWFB – slide presentation, on web

(http://sfs.au.com/Documents/NSW%20FB%20LIfts%20PP.pdf)

Howard, R.A. (1988), “Decision Analysis: Practice and Promise”, Management

Science, pp. 679-695

HMSO (1993), “Manual of Firemanship – Book 12”, Practical Firemanship II,

HMSO, London, pp. 25-50

Hoeksema, R.J. & Kitanidis, P.K. (1985), “Analysis of the spatial structure of

properties of selected aquifers”, Water Resources Research, 21(4), pp. 563-572

Hostikka, S. and Rahkonen O. K. (2003), “Probabilistic Simulation of Fire

Scenarios”, Science Direct, Nuclear Engineering and Design 224, pp. 301-311

Hultquist, H., and Karlsson, B. (2000), “Evaluation of a Fire Risk Index Method for

Multi-storey Apartment Buildings”, Department of Fire Safety Engineering

Lund University, Sweden

ISO TR 16765 (2003), “Comparison of Worldwide lifts (elevators) safety standards –

Fire fighters lifts (elevators)”, International Standardisation Organisation

ISO TS 13571 (2002), “Life-threatening components of fire — Guidelines for the

estimation of time available for escape using fire data”, International

Standardisation Organisation

Jin, T. and Yamada, T. (1985), “Irritating Effects of Fire Smoke on Visibility”, Fire

Science & Technology, Vol.5, No.1, pp. 79-89

256

Jin, T. and Yamada, T. (1989), “Experimental Study of Human Behaviour in Smoke

Filled Corridors”, Fire Safety Science – Proceeding of the Second International

Symposium, Hemisphere Publishing Corp., Washington, pp. 511-519

Jin, T. (2002), “Visibility and Human Behaviour in Fire Smoke”, The SFPE

Handbook of Fire Protection Engineering, Third Edition, pp. 2-42

Juillet, E. (1993), “Evacuating People with Disabilities”, Fire Engineering,

November, Vol. 126, No. 12, pp. 100-103

Karlsson, B. (2000), “Fire Risk Index Method for Multi-storey Apartment

Buildings”, Trätek AB, Report I0009025, Stockholm, Sweden

Keefer, D. L., Kirkwood, C.W. and Corner, J.L. (2002), “Summary of Decision

Analysis Applications in the Operations Research Literature, 1990-2001”,

Technical Report, Department of Supply Chain Management, Arizona State

University, USA

King, M. (1997), “Australian National Fire Statistics 1993-1994”, Technical Report

TR 97/2, CSIRO, Australia

Kisko, T.M, Francis, R.L. and Nobel, C.R. (1998), “EVACNET 4 User’s Guide”,

University of Florida, USA

Klote, J.H. (1982), “Elevators as a Means of Fire Escape”, American Society of

Heating Refrigerating and Air Conditioning Engineers Transactions, NBSIR

82-2507; pp. 1-40

Klote, J.H., and Tamura, G.T. (1986), “Elevator Piston Effect and the Smoke

Problem’’, Fire Safety Journal, Vol. 11, No. 3, pp. 227-233

Klote, J.H., Alvord, D.M., and Deal, S. (1991), “Routine for Analysis of the People

Movement Time for Elevator Evacuation - ELVAC”, NISTIR 4730, USA

Klote, J.H. and Milke, J.A. (1992), “Design of Smoke Management System”,

ASHRAE and SFPE

Klote, J. H., Deal, S. P., Levin, B. M., Groner, N. E., and Donoghue, E. A. (1993a),

“Workshop on Elevator use During Fires”, NIST, NISTR 4993

Klote, J.H., Deal, S.P., Donoghue, E.A., Levin, B.M., and Groner, N.E. (1993b),

“Fire Evacuation by Elevators”, Elevator World, June

Klote, J. H., Levin, B. M. and Groner, N. E. (1995), “Emergency Elevator

Evacuation Systems”, Proc. of ASME 2nd Symposium on Elevators, Fire and

Accessibility, Baltimore, MD, Feb 19-21, 1995, ASME, New York, NY, pp.

131-149

257

Klote, J. H. (2002), “Smoke Control”, The SFPE Handbook of Fire Protection

Engineering, 3rd Edition, pp. 274-291

Klote, J.H. (2003), “Analysis of the Life Safety Consequences of Smoke Migration

through Elevator Shafts”, American Society of Mechanical Engineers, ASME,

pp. 1-20

Koshak, J. (2003), “Elevator Evacuation in Emergency Situations”, ASME, pp. 1-6

Kuligowski, E. (2003), “Elevators for Occupant Evacuation and Fire Department

Access”, Proceedings of the CIB-CTBUH International Conference on Tall

Building, Malaysia

Kuligowski, E. and Bukowski, R.W. (2004), “Design of Occupants Egress Systems

for Tall Buildings”, National Institute of Standards and Technology, CIB, USA,

pp. 1-10

Lacey, R. (2000), “Lift Management in Emergencies”, Elevator World, May 2000,

pp. 126

Larsson, D. (2000), “Developing the Structure of a Fire Risk Index Method for

Multi-storey Apartment Buildings”, Department of Fire Safety Engineering”,

Lund University, Report 5062, Lund, Sweden

Lee, W. B., Lau, H., Liu, Z, and Tam, S. (2001), “A Fuzzy Analytical Hierarchy

Process Approach in Modular Product Design”, Expert System, Vol.18 No.1,

pp. 32-42

Levin, B.C., Paabo, M., Gurman, J.J., and Harris, S.E. (1987), “Effects of Exposure

to Single or Multiple Combinations of the Predominant Toxic Gases and Low

Oxygen Atmospheres Produced in Fires”, Fundamental and Applied

Toxicology, Vol. 9, pp. 236-250

Magnusson, S.E. (1997), “Risk Assessment”, Fire Safety Science – Proceeding of the

Fifth International Symposium, Melbourne, Australia, pp. 41-58

Magnusson, S. E. (1996), “How to Derive Safety Factors”, International Conference

on Performance-Based Codes and Fire Safety Design Methods, Ottawa, Canada

Marchitto, N. (1991), “High Temperature Operation of Elevators”, Elevators and

Fire, pp. 25-31

Marryatt, H.W. (1988), “Fire - A Century of Automatic Sprinkler Protection

Australia & New Zealand 1886-1986”, Australian Fire Protection Association,

Melbourne, Australia

258

Martin, M. (2003), “Challenges to Using Elevators in a Fire”, American Society of

Mechanical Engineers, USA, pp. 1-5

McGrattan, K.B., Forney, G.P., Floyd, J.E., Hostikka, S. and Prasad, K. (2004), “Fire

Dynamics Simulator - Version 4.07”, Fire Research Division, Building and Fire

Research laboratory, National Institute of Standards and Technology, USA

Meacham, B.J. (2002), “Building Fire Risk Analysis”, The SFPE Handbook of Fire

Protection Engineering, 3rd Edition, pp. 5:162

Melinek, S.J. and Booth, S. (1975), “An Analysis of Evacuation Times and the

Movement of Crowds in Buildings”, Borehamwood, Building Research

Establishment, Fire Research Station, UK

Miller, I. (2005), “Human Behaviour Contributing to Unintentional Residential Fire

Deaths 1997 -2003”, New Zealand Fire Service Commission Research Report

Number 47, New Zealand Fire Service Commission

Modarres, M., and Billoch, F.J. (2002), “Reliability”, The SFPE Handbook of Fire

Protection Engineering, 3rd Edition, pp. 5-34

Mori, B.C., and UHK (2002), “A Survey on Residents’ Responses to High-Rise

Living in Hong Kong”, Mori Building Co. Ltd and University of Hong Kong,

Hong Kong

NBS (2000), “The Building Regulations 2000, Approved Document B”, ISBN-

101859462014, Office of the Deputy Prime Minister, NBS Publications, UK

Nelson, H.E., and MacLennan, H.A. (1995), “Emergency Movement”, The SFPE

Handbook of Fire Protection Engineering, 2nd Edition, pp. 3:286-295

NFPA (1982), “Fire at the MGM Grand”, Fire Journal, National Fire Protection

Association (NFPA), pp. 19-37

NFPA (1998), “Fire Protection Handbook”, 18th Edition, Table 8-1P, pp. 8-17

NFPA 13 (1996), “Standard for the Installation of Sprinkler Systems”, NFPA,

Quincy, MA

NFPA (1999), “Special Data Information Package, High-rise Fatal Fires – Apartment

Buildings”, NFPA, Quincy, MA

NFPA 101 (2000), “Life Safety Code”, Volume 5, NFPA, Quincy, MA

NFPA 72 (2002), “National Fire Alarm Code”, NFPA, Quincy, MA

NFPA (2006), “Leading Causes of Structure Fires in Homes, 1999-2002 Annual

Averages”, NFPA, Fire Statistics – Causes, USA

259

NIST Special Publication 983 (2003), “Collected Publications Related to the Use of

Elevators during Fires”, Building and Fire Research Laboratory, NIST, USA

NUREG/CR-1278 (1980), “Handbook of Human Reliability Analysis with Emphasis

on Nuclear Power Plant Applications”, U.S. Atomic Energy Commission,

Washington, D C, USA

NZFS incident statistics, 1993-1997, “Table 8: All Fires – How They were

Extinguished”, New Zealand Fire Service, New Zealand

O’Donoghue, J.J. and O’Donnel J.K. (2003), “Why the Fire Service Must Maintain

Control of Elevators during a High-Rise Fire/ Emergency”, ASME, USA

Palisade Corp (1996), “@RISK 4.5 Software”, USA

Pauls, J.L. (1977), “Human Response to Tall Building”, Dowden, Hutchinson and

Ross, Stroudsburg, PA, USA

Pauls, J.L., Gatfield, A. J. and Juillet, E. (1991), “Elevator Use and for Egress: The

Human Factors Problems and Prospects”, ASME Symposium on Fire and

Elevators, Baltimore, MD, Feb 19-20, ASME, New York, USA, pp. 63-75

Pauls, J. (1995), “Movement of People”, The SFPE Handbook of Fire Protection

Engineering, Second Edition, pp. 3-278

Peakcock, R.D., Jones, W.W., Bukowski, R. W., and Forney, C.L. (1991),

“Technical Reference Guide for the HAZARD I Fire Hazard Assessment

Method”, Version 1.1, NIST Handbook 146, Volume II, NIST, MD, USA

Peakcock, R.D., Forney, G.P., Reneke, P., Portier, R., and Jones, W.W. (1993),

“CFAST, the Consolidated Model of Fire and Smoke Transport”, NIST

Technical Note 1299, NIST, MD, USA

Pfeiffer, D.U. (1997), “Decision Analysis and Risk Analysis”, Risk Analysis and

Animal Health Course Manual, International Training Course, Diibendorf,

Switzerland, pp. 861-877

Phillips, W.G.B. (1995), “Computer Simulation for Fire Protection Engineering”,

The SFPE Handbook of Fire Protection Engineering, Second Edition, pp. 5-7

Proulx, G. (1993), “A Stress Model for People Facing a Fire”, Journal of

Environmental Psychology, 13, pp. 137-147

Proulx, G. (1995), “Evacuation Timing in Apartment Buildings”, Fire Safety Journal,

Vol. 24, pp. 229-246

Proulx, G. (2000), “Is it Wise to Evacuate during a High-Rise Fire?”, Presentation,

NFPA Conference, Denver, USA.

260

Proulx, G. (2003), “Human Behaviour Study: Cook County Administration Building

Fire October 17, 2003, Chicago”, Research Report 181, Institute for Research

in Construction, National Research Council, Ottawa, Canada.

Purser, D.A. (2002), “Toxicity Assessment of Combustion Products”, The SFPE

Handbook of Fire Protection Engineering, 3rd Edition, pp. 2:83-171

Ramachandran, G. (1991), “Human Behaviour in Fires-A Review of Research in the

United Kingdom”, Fire Technology, 47, 1, pp. 66-81

Ramachandran, G. (2002), “Stochastic Models of Fire Growth”, The SFPE

Handbook of Fire Protection Engineering, 3rd Edition, pp. 381-401

Reber, A.S. (1995), “The Penguin Dictionary of Psychology”, Second Edition,

Penguin Books Ltd, London, UK

Robibero, V.P. (1991), “High Temperature Operation of Elevators”, Elevators and

Fire, pp. 37-43

Rockwell (2000), “ARENA 5.0 Version”, Rockwell Software Inc., USA

Rohr, K. (2001), “U.S. Experience with Sprinklers”, NFPA, September, on web

(http://www.afsaflorida.org/os-sprinkler01.pdf)

Roytman, M.Y. (1969), “Principles of Fire Safety Standards for Building

Construction”, Construction Literature Publishing House, Moscow, English

translation (TT 71-580002) from National Technical Information Service

Saaty, T.L. (1977), “A Scaling Method for Priorities in Hierarchical Structures”,

Journal of Mathematical Psychology, Vol. 15, No.3, pp. 234-281

Saaty, T.L. (1980), “The Analytic Hierarchy Process”, McGraw-Hill, New York,

USA.

Schroeder, J. (1986), “The Energy Consumption of Elevators”, Elevator Technology,

(Editor Dr. G. Barney), Ellis Horwood

Schultz, D.P. (1968), “Contract Report NR 170-274”, University of North Carolina,

Charlotte, USA

Sekizawa, A. I. (1991), “Statistical Analyses on Fatalities Characteristics of

Residential Fires”, Fire Safety Science – Proceedings of the 3rd International

Symposium (Editors: Cox, G. and Langford), B., Edinburgh, UK, pp. 475-484

Sekizawa, A., Nakahama, S., Notake, H., Ebihara, M., and Ikehata Y. (1996), “Study

on Feasibility of Evacuation using Elevators in a High-rise Building”,

University of Tokyo, Taisei Corporation and Shimizu Corporation, Japan

261

Shestopal, V.O. (1993), “Computer Program for an Uninhibited Smoke Plume and

Associated Computer Software”, Fire Technology, Vol. 29, No. 3, pp. 246-267

Shields, T.J., and Silcock, G.W.H. (1987), “Means of Escape”, Buildings and Fire,

Longman Scientific & Technical, 397

Siikonen, M.L. (1989), “Hissiliikenteen Ja Ohjaukesen Mallintaminen

Tietokoneella”, Licentiate Thesis, Helsinki University of Technology,

Department of Engineering Physic and Mathematics, System Analysis

Laboratory, Finland

Siikonen, M.L. (1997), “Planning and Control Models for Elevators in High-Rise

Buildings”, Research Report (KONE Corporation), Helsinki University of

Technology, System Analysis Laboratory, Finland

Siikonen, M.L., and Hakonen, H. (2003), “Efficient Evacuation Methods in Tall

Buildings”, Elevator World, July, pp. 78-83

Sime, J.D. (1983), “Affiliative Behaviour during Escape to Building Exits”, Journal

of Environmental Psychology, Vol. 3 No. 1, pp. 21-41

Smith, D. (2003), “Lift could be used for Evacuation during Emergencies! Fact or

Fiction?”, ASME, pp. 1-20

So, A., and Yu, J. (2003), “Lift Logic”, Fire Prevention – Fire Engineers Journal,

August Issue, Gloucestershire, UK, pp. 26-28

Straskosch, G.R. (1998), “The Vertical Transportation Handbook”, Third Edition,

ISBN: 978-0471162919, John Viley & Sons, Inc, NJ, USA

Tamura, G.T. and Shaw, C.Y. (1976), “Air Leakage Data for the Design of Elevator

and Stair Shaft Pressurisation System”, ASHRAE Transactions, Vol. 82 (Part

2), pp. 179-190

Tamura, G. (1994), “Smoke Movement and Control in High-Rise Buildings”, NFPA,

USA

Taylor, R. E. (1975) “The Carlyle Apartment Fire: Study of a Pressurised Corridor”,

ASHRAE Journal, Vol. 17, pp. 52-55

Teo, H.P. (2001), Key Note Address, The International Association of Elevator

Engineers in Elevator Technology 11, Proceedings of Elevcon 2001, the 11th

International Congress on Vertical Transportation Technologies, Singapore

Proceedings of ELEVCON 2001, Singapore, July 2001

262

Tyni, T. and Ylinen, J. (2006), “Evolutionary Bi-Objective Optimisation in the

Elevator Car Routing Problem”, European Journal of Operational Research

169(3), pp. 960-977

Vistnes, J., Grubits, S. J., and He, Y. (2005), “A Stochastic Approach to Occupant

Pre-Movement in Fires”, Proceedings of the 8th International Symposium on

Fire Safety Science, pp. 531-542

Vrijling, J.K., Van Hengel, W., and Houben, R.J. (1995), “A Framework for Risk

Evaluation”, Journal of Hazardous Materials, 43, pp. 245–261

Walton, G. N. (1993) “CONTAMW 1.0 User Manual”, Multi-zone Airflow and

Contaminant Transport Analysis Software, NISTIR 6476, NIST

WASH 1400 (1974), “Reactor Safety Study, An Assessment of Accident Risks in

U.S. Commercial Nuclear Power Plants, Appendix III, Failure Data”, U.S.

Atomic Energy Commission, Washington, D.C.

Watts, J.M. Jr. (1991), “Criteria for Fire Risk Ranking”, Fire Safety Science,

Proceedings of the Third International Symposium, pp. 457-466

Watts, J.M. Jr. (2002), “Fire Risk Ranking”, The SFPE Handbook of Fire Protection

Engineering, 3rd Edition, pp. 5:125-142

Watts, J.M. Jr. and Hall, J. R. (2002), “Introduction to Fire Risk Analysis”, The

SFPE Handbook of Fire Protection Engineering, 3rd Edition, pp. 5:1-7

WHI (1987), “Test of Gasketing for Resistance to Air Flow at Elevated Temperature

to Determine Effectiveness as a Smoke Control Door Seal”, WHI Test Report

# 630-611300, Warnock Hersey International, Inc.

Willey, A. (1973), “Fire Baptist Towers Housing For the Elderly”, Fire Journal, May

1973, NFPA, pp. 28-36

Zhao, L. (1998), “Technical Report FCRC-TR 98-05”, Project 4, Part A

263

Appendix A

International Listing of Major Fires, where Lifts were used

The National Fire Protection Association (NFPA) has maintained an international

listing of high-rise building fires since 1911 (a few incidents reproduced in book

‘Fire Safety in Tall Building’ published by Council on Tall Buildings and Urban

Habitat Committee 8A). Some of the major incidents, where occupants tried to use

lift for evacuation, are given in Table A1:

Table A1: Use of Lifts during Major Building Fires

Date Building Occupancy Number of

storeys

(Origin of

fire)

Number

of

deaths

A brief of lift use

24.01.1969 Hawthorne House, Chicago

Apartment 39 (36) 4 Use of lifts by residents for building evacuation, which causes delay in fire fighting operation

5.08.1970 One New York Plaza

Office 50 (33) 2 Deaths from lift stopping at fire floor

4.12.1970 Office building NY

Office 47 (5) 3 Deaths from lift stopping at fire floor

2.07.1971 Motor hotel complex NY

Hotel 17 (12) 6 Victims mistakenly tried lifts versus stairs

1.02.1974 Joelana building Brazil

25 (12) 179 Lifts used heavily for rescue

12.11.1974 Century City LA

Office 15 (8) - Lifts used for evacuation

23.06.1980 Westvaco building NY

Office 42 (20) - Elevator malfunction (a severe problem)

21.11.1980 MGM Grand Hotel LA

Hotel 23 (1) 85 Smoke spread through air system, stairs and lift hoist way. Death from smoke in rooms, corridor and lifts.

10.02.1981 Hilton Hotel LA

Hotel 30 (8) 8 Fire rated lift vestibules. Fire started in lift lobby (arson) and spread to exterior

31.12.1986 Dupont Plaza Hotel Puerto Rico

Hotel 20 (1) 97 Smoke spread through lifts, air system, utility passage ways, stairs and building exterior.

24.12.1989 John Sevier Retirement Centre

Old age residential

11 (1) 16 Typifies evacuation problems in elderly housing

Note: In spite of notices not to use lifts, lifts were used for evacuation. Some of them

lead to a disastrous end.

264

Appendix B

Risk Priorities and Matrix Consistency Ratio

Calculation of Risk Priorities (3 × 3 Matrix)

Decision Uncertainty Panic Nonfatal Injuries Average Risk

Priorities

Decision

Uncertainty 1

0667.0951

1=

++

1/5

0625.021

51

51

=

++

1/9

0683.01

21

91

91

=

++

3

0683.00625.00667.0 ++ 0.066

Panic 5

3333.0951

5=

++

1

3125.021

51

1=

++

½

3105.01

21

91

21

=

++

3

3105.03125.03333.0 ++ 0.319

Nonfatal

Injuries 9

6.0951

9=

++

2

625.021

51

2=

++

1

6211.01

21

91

1=

++

3

6211.0625.06.0 ++

0.615

265

Calculation of Eigenvalue, Consistency Index and Consistency Ratio

Decision Uncertainty Panic Nonfatal Injuries

Decision Uncertainty 1 – λ 1/5 1/9

Panic 5 1 – λ ½

Nonfatal Injuries 9 2 1 – λ

To determine Eigenvalue λ, cubical equation is resolved as shown next:

[ ] ( ) ( ) ( ) ( ) ( ) ( )[ ]1112/129/19(5/1511123 −−−×+×+×−++− λλλ( ) ( ) ( ) ( ) ( ) ( )[ ] 019/192/15/1929/15155/122/11111 =××−××+××+××−××−××+

( ) ( ) ( )[ ] 010/99/1023 23 =++−−− λλ

( )[ ] 090/13 23 =−− λλ

10.3=λ

The Consistency Index (CI) is calculated as given below:

05.013

310.3

1

max =−−

=−

−=n

nCI

λ

The Consistency Ratio (CR) is calculated as given below (RI is 0.58):

086.058.0

05.0===

RI

CICR

266

Appendix C

Survey Questionnaire

1. Your age ……………

2. Floor level of your flat in the multi-storey building…………….level

3. Do you use the stairs for normal building access?

Always Sometimes Rarely Never

4. Do you use the lift for normal building access?

Always Sometimes Rarely Never

5. Do you use the stairs as a normal building exit?

Always Sometimes Rarely Never

6. Do you use the lift as a normal building exit?

Always Sometimes Rarely Never

7. How would you evacuate the building during a fire emergency?

Stairs only

Lift only

Both stairs and lift

Remain in the building

By fire brigade appliance

8. Have you ever been trained in fire emergency evacuation procedure?

Yes No

9. If you have used the stairs in the fire drill, have you had any difficulty in

using the stairs?

Yes No

10. Any specific information on difficulty encountered during fire drills/ any

other comments

……………………………….………………………….……...

‘THANK YOU FOR YOUR ACTIVE PARTICIPATION IN THE INTEREST

OF PUBLIC SAFETY’

267

Questionnaire Results:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

1 Age 26 52 58 47 35 52 28 58 43 55 54 65 25 34 42 25 36 54 42 40 36 56 60 40 57

2 Floor level 1 2 2 3 4 5 6 6 6 10 10 11 11 11 12 13 13 15 15 16 17 18 18 19 19

3 Stair for normal access

Always

Sometimes 1 1 1

Rarely 1 1

Never 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

4 Lift for normal access

Always 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Sometimes 1 1 1

Rarely

Never

5 Stair for exit

Always 1

Sometimes 1 1 1

Rarely 1 1

Never 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

6 Lift for exit

Always 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Sometimes 1 1

Rarely

Never

7 How would you evacuate

Stair only 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Lift only

Both stair and lift 1

Stay in place

By fire brigade

8 Are you trained in EEP

Yes 1 1 1 1 1 1 1

268

No 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

9 Difficulty in stair in EEP

Yes 1 1 1 1

No 1 1 1 1 1 1 1 1 1 1 1 1

10 Any specific information

(stair locked) 1 1

Other details are given on next page

--2--

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

1 Age 30 19 42 30 45 50 35 20 35 69 31 73 26 32 57 66 24 50 44 59 24 58 44 61 24

2 Floor level 20 21 21 21 21 22 23 25 26 26 27 27 28 28 29 29 30 31 31 32 33 33 34 37 38

3 Stair for normal access

Always

Sometimes

Rarely 1 1

Never 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

4 Lift for normal access

Always 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Sometimes

Rarely

Never

5 Stair for exit

Always

Sometimes

Rarely 1 1

Never 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

6 Lift for exit

Always 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Sometimes

Rarely

Never

7 How would you evacuate

269

Stair only 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Lift only

Both stair and lift 1

Stay in place 1

By fire brigade

8 Are you trained in EEP

Yes 1 1 1 1 1 1 1 1 1

No 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

9 Difficulty in stair in EEP

Yes 1 1 1 1 1

No 1 1 1 1 1 1 1 1 1 1 1 1 1

10 Any specific information

(Stair locked) 1 1

Comments from residents are given below:

Fire drills are tiring

Some had knee trouble

Problem with weaker sex for long traveling

More people downward as one travels

Long evacuation time

Lots of stairs to walk

Some have to stay in the building

High level of concern for evacuation due to high altitude

Never had a fire drill

Doors are generally locked – 8% indicated that stair doors were always locked.

270

Calculation of Confidence Interval:

For calculating the confidence interval in Excel, the Beta Distribution is used to

study variation in the percentage of sample across the data. The details for

calculating the confidence interval with the help of BETAINV are given below:

Syntax: BETAINV(p, α, β, A, B)

where

p is a probability associated with the beta distribution

α is a parameter of the distribution

β is a parameter of the distribution

A is an optional lower bound to the interval of x =0

B is an optional upper bound to the interval of x = 1

α = y + 1 and β = n – y + 1

where y is the number of successes in n trials

To obtain the 95% Confidence Interval for p, let x = 0.9275 and 0.0725

There are mean number of 46.375 occupants intend to use lifts in 50 occupants, then:

α = 47.375 and β = 4.625

Using Excel,

BETAINV(0.9275, 47.375, 4.625, 0, 1) = 0.961

BETAINV(0.0725, 47.375,4.625, 0, 1) = 0.849

Hence, a 95% Confidence Interval for the true proportion is 0.849 to 0.961.

271

Appendix D

Interview

Interviews were conducted with the Fire Engineer from NSW Fire Brigade, Sydney

(Person 1), two Community Safety Officers from QFRS, Brisbane (Persons 2 and 3).

These officers have rendered mixed services of regular fire fighting and fire-

prevention. The following questions and answers were covered although these may

not be the same exact words used by the fire personnel:

Questions and Replies – Person 1:

1. Q: Do you notice any problem with the use of stairs during fire emergencies

or drills?

A: As a fire-fighter in the U.S., I often noticed that space was an issue even

when the stairways were only used for fire fighting operations (not enough

space for equipment and hose). The problem is exacerbated due to the

conflicting needs of the evacuees and the fire fighting personnel.

2. Q: How would you recommend means of evacuation for disabled, aged

people, ladies and children from high-rise building during a fire emergency?

A: Refuge areas are separated from the building, which can be used for

temporary staging then use lifts and/or stairs with assistance as needed.

3. Q: During fire emergencies, how is the occupant behaviour – irrational or

rational? What about their priority on upper levels – stairs or lift? What about

visitors in office building?

A: It depends on the scenario and the people involved. I assume that people

would usually avoid lifts if there was evidence of smoke and/or fire.

Anything different in evacuation strategy would need a change in

culture/training/acceptance. If there was minimal evidence of fire and/or

smoke then they might use the lifts, but my concern is that it’s the “big ones”

that result in deaths and/or injury upon which a lot of our building codes are

based.

4. Q: Do you feel that lift should be considered as an option (secondary) for

means of evacuation?

A: Only if the detailed design and commissioning and maintenance and use

can be proven to support such a function. This would also require a change in

occupant acceptance and attitude, since they would need to feel safe in doing

so (in addition to the system working under the conditions it would need to).

One big advantage of stairs is that even if people are waiting, they are

flowing, whereas with lifts they will egress in increments and have not

control over their destiny between lift arrivals (as per your rational behaviour

query it’s my understanding that when people feel they have control over

272

their outcome during tense emergency situations then they will tend to act

rationally but when they feel that the situation is deteriorating more rapidly

than they can react appropriately then they become irrational (like when they

might be waiting for the lift to arrive with rapidly deteriorating environmental

conditions).

5. Q: Do you feel that “fire emergency evacuation drills” are regularly held in

NSW?

A: No idea, I believe they might be in offices but I know when I was

consulting in a high-rise downtown I never did one.

6. Q: In your opinion, what is acceptable limit for comfortable travelling in

building? Whether one can come down in stairs – let say 10 storey or 20

storey or 30 storey or more.

A: If I felt safe in the stair then it wouldn’t matter.

7. Q: How the fire fighters approach to the upper levels? Do they use stairs or

lift?

A: In a working fire environment I would expect them to use the stairs until

they felt the fire floor was safe (that’s my understanding of how they do

things here).

8. Q: How far fire fighter can climb in building with fire fighting gears?

A: I believe as far as they would need to, the time needed might be more of

the issue.

9. Q: Are power supply is reliable in buildings?

A: My opinion would be that unless it is specifically designed as “Emergency

Power” with on-site emergency back up generation then it cannot be assumed

to be. I’ve seen too many blocks in the CBD out of power for hours on end

even though the power has been “deemed” to be reliable enough for back-up

for emergencies (based on girded supply). This is because there are still

single point failure modes and the assessment is not done “in-depth” as would

be needed to address all single point failure modes and effects and providing

solutions to the same.

10. Q: Any specific information, you would like to give.

A: Keep up the good work!

273

Questions and Replies – Person 2:

1. Q: Do you notice any problem with the use of stairs during fire emergencies

or drills?

A: Bottleneck movement, fire fighters have entrance problem in the building

2. Q: How would you recommend means of evacuation for disabled, aged

people, ladies and children from high-rise building during a fire emergency?

A: Such occupants should go next level downward or go to top terrace or stair

landing

3. Q: During fire emergencies, how is the occupant behaviour – irrational or

rational? What about their priority on upper levels – stairs or lift? What about

visitors in office building?

A: Nearly 50% are occupants are rational and remaining 50% are irrational.

Normally occupants try to use lifts in office buildings, whereas occupants try

to use stairs in shopping centre. Visitors would like to go in car park

straightway from the upper levels.

4. Q: Do you feel that lift should be considered as an option (secondary) for

means of evacuation?

A: It can be considered for the disabled, aged and fire fighters. Normally fire

hydrants locations are near to lift lobby, which can be made as a staging area.

5. Q: Do you feel that “fire emergency evacuation drills” are regularly held in

QLD?

A: It is must in commercial buildings, but not in residential buildings.

Building Fire Safety Regulation permits for workers only.

6. Q: In your opinion, what is acceptable limit for comfortable travelling in

building? Whether one can come down in stairs – let say 10 storey or 20

storey or 30 storey or more.

A: At every 4th level, provision is made for re-entry. So occupants could go at

every alternate 4th level.

7. Q: How the fire fighters approach to the upper levels? Do they use stairs or

lift?

A: Whatever is convenient to us.

8. Q: How far fire fighter can climb in building with fire fighting gears?

A: Generally they can climb up to 8th floor.

9. Q: Are power supply is reliable in buildings?

A: It is normally reliable in Brisbane.

10. Q: Any specific information, you would like to give.

A: The provision such as pressurisation (zone smoke management) can be

considered for lift system.

274

Questions and Replies – Person 3:

1. Q: Do you notice any problem with the use of stairs during fire emergencies

or drills?

A: Overcrowding, problematic for disabled occupants, difficulties for staging

area during drills

2. Q: How would you recommend means of evacuation for disabled, aged

people, ladies and children from high-rise building during a fire emergency?

A: Lifts can be explored as an option in high-rise buildings.

3. Q: During fire emergencies, how is the occupant behaviour – irrational or

rational? What about their priority on upper levels – stairs or lift? What about

visitors in office building?

A: Experienced people acts rationally. Generally occupants use stairs for

emergency evacuation. Occupants are given instructions through public

address system, what they have to use for evacuation.

4. Q: Do you feel that lift should be considered as an option (secondary) for

means of evacuation?

A: It must be considered as an option for elderly and disabled occupants

5. Q: Do you feel that “fire emergency evacuation drills” are regularly held in

QLD?

A: Periodically fire drills are conducted in building normally once a year.

6. Q: In your opinion, what is acceptable limit for comfortable travelling in

building? Whether one can come down in stairs – let say 10 storey or 20

storey or 30 storey or more.

A: It depends on the individual strength.

7. Q: How the fire fighters approach to the upper levels? Do they use stairs or

lift?

A: Fire fighters should use lifts for upper level access.

8. Q: How far fire fighter can climb in building with fire fighting gears?

A: Again it depends on the individual strength.

9. Q: Are power supply is reliable in buildings?

A: Yes, they are reliable in Australia.

10. Q: Any specific information, you would like to give.

A: Stack arrangements for emergency evacuation can be researched for

evacuation purpose.

275

Appendix E

Occupant Response and Coping Times

The following response and coping scores for residential occupants are based on

Table 7.9 – Chapter 7 – Fire Engineering Guidelines:

Table E1 – Weighting Factor in Occupant Response Time

Factor Weighting

Alertness 2

Role 1

Commitment 4

Focal Point 1

Mobility 3

Social Affiliation 1

Position 2

Familiarity 4

Primary factors are shown in bold

RESPONSE TIME CALCULATION (a sum of primary factors is multiplied by 2

and secondary factors by 0.4)

((10 x 2) + (8 x 0.4)) / 8 = 2.9;

6 – 2.9 = 3.1 minutes (186 seconds)

Table E2 – Weighting Factor in Occupant Coping Time

Factor Weighting

Role 1

Mobility 2

Social Affiliation 1

Alertness 3

Position 4

Commitment 2

Focal Point 2

Familiarity 4

Primary factors are shown in bold

COPING TIME CALCULATION

((4 x 2) + (15 x 0.4)) / 8 = 1.75;

6 – 1.75 = 4.25 minutes (255 seconds)

• Residential occupant response score = 186 seconds

• Residential occupant coping score = 255 seconds

276

Appendix F

SIMAN Language

SIMAN Language Used in Lift Simulation Model (only a small portion)

Simulation model has generated SIMAN output files. Various modules indicating

their attributes, variables and distribution functions are given below:

; Model statements for module: Create 1 214$ CREATE, 4,SecondstoBaseTime(0.0),Entity 1:SecondstoBaseTime(EXPO(1)),1:NEXT(215$); 215$ ASSIGN: Elevator Arival.NumberOut=Elevator Arival.NumberOut + 1:NEXT(1$); ; Model statements for module: Assign 1 1$ ASSIGN: Elevator_Number=0: Elevator_Load=0: FloorNumber=38:NEXT(0$); ; Model statements for module: Station 1 0$ STATION, Floors; 218$ ASSIGN: WhatFloor=MEMIDX(Floors,M); 220$ DELAY: 0.0,,VA:NEXT(2$); ; Model statements for module: Decide 1 2$ BRANCH, 1: If,Elevator_Number>=1,221$,Yes: Else,222$,Yes; 221$ ASSIGN: Decide 1.NumberOut True=Decide 1.NumberOut True + 1:NEXT(15$); 222$ ASSIGN: Decide 1.NumberOut False=Decide 1.NumberOut False + 1:NEXT(30$); ; Model statements for module: Decide 2 15$ BRANCH, 1: If,ReleasedAtFloor==WhatFloor,16$,Yes: If,FloorNumber==WhatFloor,60$,Yes: If,(Elevator_Load < Elevator_Capacity) * (WhatFloor > FloorNumber),60$,Yes: Else,32$,Yes; ; Model statements for module: Assign 8 32$ ASSIGN: direction=WhatFloor < FloorNumber: Accl=( ABS( FloorNumber - WhatFloor ) == 1) * 0.95: WhatFloor= ((WhatFloor < FloorNumber) * (WhatFloor + 1)) + ((WhatFloor > FloorNumber) * (WhatFloor - 1)): ReleasedAtFloor=DISC(0.01, WhatFloor, 1.0, FloorNumber):NEXT(29$); ; Model statements for module: Route 5 29$ ROUTE: TravelTime,MEMBER(Floors,WhatFloor); ; Model statements for module: Assign 2 16$ ASSIGN: ReleasedAtFloor=EWhatFloor(2) * Elevator_Load + EWhatFloor(1) * ANINT(Elevator_Load * 7 / 100): Elevator_Load=Elevator_Load - ReleasedAtFloor:NEXT(22$); ; Model statements for module: Dropoff 1 22$ DROPOFF, 1,ReleasedAtFloor:18$,

277

FloorNumber:NEXT(60$); ; Model statements for module: Delay 10 60$ DELAY: (EXPO(10) + (DISC(0.25,0,1.0,1) * EXPO(10))),,Other:NEXT(17$); ; Model statements for module: Decide 3 17$ BRANCH, 1: If,Elevator_Load + NQ(Queues_2(WhatFloor)) <= Elevator_Capacity,225$,Yes: Else,226$,Yes; 225$ ASSIGN: Can_Fit_1_Passenger.NumberOut True=Can_Fit_1_Passenger.NumberOut True + 1:NEXT(19$); 226$ ASSIGN: Can_Fit_1_Passenger.NumberOut False=Can_Fit_1_Passenger.NumberOut False + 1:NEXT(20$); ; Model statements for module: Assign 3 19$ ASSIGN: Elevator_RemainingCap=NQ(Queues_2(WhatFloor)): Elevator_Load=Elevator_Load + Elevator_RemainingCap:NEXT(31$); 31$ PICKUP: Queues_2(WhatFloor),1,Elevator_RemainingCap; 58$ WHILE: TWhatFloor(1) || (NQ(Queues_2(FloorNumber)) < 1) * (Dispose 1.Numberout<1210) * EWhatFloor(2) :NEXT(21$); ; Model statements for module: Assign 5 21$ ASSIGN: FloorNumber= TWhatFloor(1)+(EWhatFloor(2)*DISC(0.027,38,0.054,37,0.081,36,0.108,35,0.135,34,0.162,33,0.189,32,0.216,31,0.243,30,0.27,29,0.297,28,0.324,27,0.351,26,0.378,25,0.405,24,0.432,23,0.459,22,0.486,21,0.513,20,0.54,19,0.567,18,0.594,17,0.621,16,0.648,15,0.675,14,0.702,13,0.729,12,0.756,11,0.783,10,0.81,9,0.837,8,0.864,7,0.891,6,0.918,5,0.945,4,0.972,3,1,2))+(TWhatFloor(2)*FloorNumber): ReleasedAtFloor=FloorNumber:NEXT(61$); 61$ SCAN: (QE(1) + QE(2) + QE(3) + QE(4) + QE(4) + QE(5) + QE(6) + QE(7)) + Elevator_Load >= 1; 59$ ENDWHILE:NEXT(213$); ; Model statements for module: Assign 11 213$ ASSIGN: FloorNumber= ((TWhatFloor(1)+(EWhatFloor(2)*DISC(0.027,38,0.054,37,0.081,36,0.108,35,0.135,34,0.162,33,0.189,32,0.216,31,0.243,30,0.27,29,0.297,28,0.324,27,0.351,26,0.378,25,0.405,24,0.432,23,0.459,22,0.486,21,0.513,20,0.54,19,0.567,18,0.594,17,0.621,16,0.648,15,0.675,14,0.702,13,0.729,12,0.756,11,0.783,10,0.81,9,0.837,8,0.864,7,0.891,6,0.918,5,0.945,4,0.972,3,1,2))+(TWhatFloor(2)*FloorNumber))*(WhatFloor == FloorNumber)) + (FloorNumber * (WhatFloor <> FloorNumber)): ReleasedAtFloor=FloorNumber:NEXT(32$); ; Model statements for module: Assign 4 20$ ASSIGN: Elevator_RemainingCap=Elevator_Capacity - Elevator_Load: Elevator_Load=Elevator_Capacity:NEXT(31$); ; Model statements for module: Route 2 18$ ROUTE: EXPO( 20 ),Apartments(FloorNumber); ; Model statements for module: Assign 6 30$ ASSIGN: Elevator_Number=LastElevatorNumber + 1: LastElevatorNumber=Elevator_Number:NEXT(32$); ; Model statements for module: Create 2 227$ CREATE, 1,SecondstoBaseTime(ResponseTime),Person:SecondstoBaseTime(POIS( 23.5 )),22:NEXT(228$); 228$ ASSIGN: Passenger Arrival Floor 01.NumberOut=Passenger Arrival Floor 01.NumberOut + 1:NEXT(14$); 14$ QUEUE, Floor 01.Queue:DETACH;

278

; Model statements for module: Create 3 231$ CREATE, 1,SecondstoBaseTime(ResponseTime),Person:SecondstoBaseTime(POIS( 6.81 )),32:NEXT(232$); 232$ ASSIGN: Passenger Arrival Floor 02.NumberOut=Passenger Arrival Floor 02.NumberOut + 1:NEXT(13$); 13$ QUEUE, Floor 02.Queue:DETACH; ; Model statements for module: Create 4 207$ ASSIGN: StartTime=TNOW:NEXT(210$); ; Model statements for module: Hold 1 210$ SCAN: Dispose 1.Numberout == 1206:NEXT(208$); ; Model statements for module: Record 1 208$ TALLY: Total System Time,INT(StartTime),1:NEXT(212$); ; Model statements for module: Assign 10 212$ ASSIGN: FinishReplication=1:NEXT(211$); ; Model statements for module: Dispose 2 211$ ASSIGN: Dispose 2.NumberOut=Dispose 2.NumberOut + 1; 609$ DISPOSE: Yes;

279

A r iv a lE le v a t o r

S t a t io n 1A s s ig n 1

F lo o r 0 1A r r iv a l

P a s s e n g e r

F lo o r 0 2A r r iv a l

P a s s e n g e r

F lo o r 0 3A r r iv a l

P a s s e n g e r

F lo o r 0 4A r r iv a l

P a s s e n g e r

T r u e

F a ls e

D e c id e 1

2S t a t io n

D is p o s e 1

0 2

A p a r t m e n t

0 3

A p a r t m e n t

0 4

A p a r t m e n t

D e la y 1 R o u t e 1

6

S t a t io n

F lo o r 0 4 . Q u e u e

Q u e u e

F lo o r 0 3 . Q u e u e

Q u e u e

F lo o r 0 2 . Q u e u e

Q u e u e

F lo o r 0 1 . Q u e u e

Q u e u e

R e le a s e d A t F lo o r = = W h a t F lo o r

F lo o r N u m b e r = = W h a t F lo o r

( E le v a t o r _ L o a d < E le v a t o r _ C a p a c it y ) * ( W h a t F lo o r > F lo o r N u m b e r )E ls e

D e c id e 2

A s s ig n 2

C a n _ F it _ 1 _ P a s s e n g e r

T r u e

F a ls e

R o u t e 2

A s s ig n 3

A s s ig n 4

A s s ig n 5

O r ig in a l

M e m b e r s

1

D r o p o f f

7

S t a t io n

8

S t a t io n

D e la y 2 R o u t e 3

D e la y 3 R o u t e 4

Passenger Zone

Controlle r and L ift Zone

Passenger Re-entry Zone

F ina l Exit

R o u t e 5A s s ig n 6

E le v a t o r _ R e m a in in g C a pQ u e u e s _ 2 ( W h a t F lo o r )

P ic k u p

A s s ig n 8

F lo o r 0 5A r r iv a l

P a s s e n g e r

F lo o r 0 6A r r iv a l

P a s s e n g e r

F lo o r 0 6 . Q u e u e

Q u e u e

F lo o r 0 5 . Q u e u e

Q u e u e

9S t a t io n

1 0

S t a t io n

F lo o r 0 7A r r iv a l

P a s s e n g e r

F lo o r 0 8A r r iv a l

P a s s e n g e r

F lo o r 0 8 . Q u e u e

Q u e u e

F lo o r 0 7 . Q u e u e

Q u e u e

1 1

S t a t io n

1 2

S t a t io n

F lo o r 0 9A r r iv a l

P a s s e n g e r

F lo o r 0 9 . Q u e u e

Q u e u e

1 3S t a t io n

0 6

A p a r t m e n t

0 5

A p a r t m e n t

D e la y 5 R o u t e 6

D e la y 6 R o u t e 7

0 8

A p a r t m e n t

0 7

A p a r t m e n t

D e la y 7 R o u t e 8

D e la y 8 R o u t e 9

0 9A p a r t m e n t

D e la y 91 0

R o u t e

Animation

T W h a t F lo o r ( 1 ) | | ( N Q ( Q u e u e s _ 2 ( F lo o r N u m b e r ) ) < 1 ) * ( D is p o s e 1 . N u m b e r o u t < 1 2 1 0 ) * E W h a t F lo o r ( 2 )

W h ile E n d W h ile

F lo o r D r o p P ic k D e la y

( Q E ( 1 ) + Q E ( 2 ) + Q E ( 3 ) + Q E ( 4 ) + Q E ( 4 ) + Q E ( 5 ) + Q E ( 6 ) + Q E ( 7 ) ) + E le v a t o r _ L o a d > = 1

S c a n

F lo o r 1 0A r r iv a l

P a s s e n g e r

F lo o r 1 1A r r iv a l

P a s s e n g e r

F lo o r 1 1 . Q u e u e

Q u e u e

F lo o r 1 0 . Q u e u e

Q u e u e

2 2

S t a t io n

2 3

S t a t io n

F lo o r 1 2A r r iv a l

P a s s e n g e r

F lo o r 1 2 . Q u e u e

Q u e u e

2 4S t a t io n

1 1

A p a r t m e n t

1 0

A p a r t m e n t

D e la y 1 11 1

R o u t e

D e la y 1 21 2

R o u t e

1 2

A p a r t m e n tD e la y 1 3

1 3

R o u t e

F lo o r 1 3A r r iv a l

P a s s e n g e r

F lo o r 1 4A r r iv a l

P a s s e n g e r

F lo o r 1 5A r r iv a l

P a s s e n g e r

2 8

S t a t io n

F lo o r 1 5 . Q u e u e

Q u e u e

F lo o r 1 4 . Q u e u e

Q u e u e

F lo o r 1 3 . Q u e u e

Q u e u e

2 9

S t a t io n

3 0

S t a t io n

F lo o r 1 6A r r iv a l

P a s s e n g e r

F lo o r 1 7A r r iv a l

P a s s e n g e r

F lo o r 1 7 . Q u e u e

Q u e u e

F lo o r 1 6 . Q u e u e

Q u e u e

3 1

S t a t io n

3 2

S t a t io n

F lo o r 1 8A r r iv a l

P a s s e n g e r

F lo o r 1 9A r r iv a l

P a s s e n g e r

F lo o r 1 9 . Q u e u e

Q u e u e

F lo o r 1 8 . Q u e u e

Q u e u e

3 3S t a t io n

3 4

S t a t io n

F lo o r 2 0A r r iv a l

P a s s e n g e r

F lo o r 2 0 . Q u e u e

Q u e u e

3 5

S t a t io n

F lo o r 2 1A r r iv a l

P a s s e n g e r

F lo o r 2 2A r r iv a l

P a s s e n g e r

F lo o r 2 2 . Q u e u e

Q u e u e

F lo o r 2 1 . Q u e u e

Q u e u e

3 6

S t a t io n

3 7

S t a t io n

F lo o r 2 3A r r iv a l

P a s s e n g e r

F lo o r 2 3 . Q u e u e

Q u e u e

3 8

S t a t io n

F lo o r 2 4A r r iv a l

P a s s e n g e r

F lo o r 2 4 . Q u e u e

Q u e u e

3 9S t a t io n

F lo o r 2 5A r r iv a l

P a s s e n g e r

F lo o r 2 5 . Q u e u e

Q u e u e

4 0

S t a t io n

1 3

A p a r t m e n t

1 4

A p a r t m e n t

1 5

A p a r t m e n t

D e la y 1 41 4

R o u t e

D e la y 1 51 5

R o u t e

D e la y 1 61 6

R o u t e

1 7A p a r t m e n t

1 6

A p a r t m e n t

D e la y 1 71 7

R o u t e

D e la y 1 81 8

R o u t e

1 9

A p a r t m e n t

1 8

A p a r t m e n t

D e la y 1 91 9

R o u t e

D e la y 2 02 0

R o u t e

2 0

A p a r t m e n tD e la y 2 1

2 1

R o u t e

2 2

A p a r t m e n t

2 1

A p a r t m e n t

D e la y 2 22 2

R o u t e

D e la y 2 32 3

R o u t e

2 3

A p a r t m e n tD e la y 2 4

2 4

R o u t e

2 4

A p a r t m e n tD e la y 2 5

2 5

R o u t e

2 5

A p a r t m e n tD e la y 2 6

2 6

R o u t e

F lo o r 2 6A r r iv a l

P a s s e n g e r

F lo o r 2 7A r r iv a l

P a s s e n g e r

5 4

S t a t io n

F lo o r 2 7 . Q u e u e

Q u e u e

F lo o r 2 6 . Q u e u e

Q u e u e

5 5

S t a t io n

F lo o r 2 8A r r iv a l

P a s s e n g e r

F lo o r 2 9A r r iv a l

P a s s e n g e r

F lo o r 2 9 . Q u e u e

Q u e u e

F lo o r 2 8 . Q u e u e

Q u e u e

5 6

S t a t io n

5 7

S t a t io n

F lo o r 3 0A r r iv a l

P a s s e n g e r

F lo o r 3 1A r r iv a l

P a s s e n g e r

F lo o r 3 1 . Q u e u e

Q u e u e

F lo o r 3 0 . Q u e u e

Q u e u e

5 8S t a t io n

5 9

S t a t io n

F lo o r 3 2A r r iv a l

P a s s e n g e r

F lo o r 3 2 . Q u e u e

Q u e u e

6 0

S t a t io n

F lo o r 3 3A r r iv a l

P a s s e n g e r

F lo o r 3 4A r r iv a l

P a s s e n g e r

F lo o r 3 4 . Q u e u e

Q u e u e

F lo o r 3 3 . Q u e u e

Q u e u e

6 1S t a t io n

6 2

S t a t io n

F lo o r 3 5A r r iv a l

P a s s e n g e r

F lo o r 3 5 . Q u e u e

Q u e u e

6 3

S t a t io n

F lo o r 3 6A r r iv a l

P a s s e n g e r

F lo o r 3 6 . Q u e u e

Q u e u e

6 4

S t a t io n

F lo o r 3 7A r r iv a l

P a s s e n g e r

F lo o r 3 7 . Q u e u e

Q u e u e

6 5

S t a t io n

F lo o r 3 8A r r iv a l

P a s s e n g e r

F lo o r 3 8 . Q u e u e

Q u e u e

6 6

S t a t io n

2 6

A p a r t m e n t

2 7

A p a r t m e n t

2 8

A p a r t m e n t

D e la y 2 72 7

R o u t e

D e la y 2 82 8

R o u t e

D e la y 2 92 9

R o u t e

3 0

A p a r t m e n t

2 9

A p a r t m e n t

D e la y 3 03 0

R o u t e

D e la y 3 13 1

R o u t e

3 2

A p a r t m e n t

3 1

A p a r t m e n t

D e la y 3 23 2

R o u t e

D e la y 3 33 3

R o u t e

3 3A p a r t m e n t

D e la y 3 43 4

R o u t e

3 5

A p a r t m e n t

3 4

A p a r t m e n t

D e la y 3 53 5

R o u t e

D e la y 3 63 6

R o u t e

3 6

A p a r t m e n tD e la y 3 7

3 7

R o u t e

3 7

A p a r t m e n tD e la y 3 8

3 8

R o u t e

3 8

A p a r t m e n tD e la y 3 9

3 9

R o u t e

C r e a t e 4 3 A s s ig n 9 R e c o r d 1H o ld 1

D is p o s e 2

A s s ig n 1 0

A s s ig n 1 1

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0 0

Figure F1 – Overall Model used for Lift Simulation

280

SIMAN Language Used in Stair Evacuation Model (only a small portion) ; Model statements for module: Create 3 931$ CREATE, 1,SecondstoBaseTime(Response1),Person:SecondstoBaseTime(POIS( 34 )),16:NEXT(932$); 932$ ASSIGN: Passenger Arrival Floor 02.NumberOut=Passenger Arrival Floor 02.NumberOut + 1:NEXT(878$); ; Model statements for module: Assign 73 878$ ASSIGN: Floor_entered=02: AllTWay=DISC(0.99, 1, 1.0, 0): WantedFloor= (AllTWay == 1) + (AllTWay <> 1) * DISC(0.028,38,0.056,37,0.084,36,0.112,35,0.14,34,0.168,33,0.196,32,0.224,31,0.252,30,0.28,29,0.308,28,0.336,27,0.364,26,0.392,25,0.42,24,0.448,23,0.476,22,0.504,21,0.532,20,0.56,19,0.588,18,0.616,17,0.644,16,0.672,15,0.7,14,0.728,13,0.756,12,0.784,11,0.812,10,0.84,9,0.868,8,0.896,7,0.924,6,0.95,5,0.98,4,1,3) :NEXT(189$); ; Model statements for module: Station 116 189$ STATION, Stair_lobby02; 937$ DELAY: 0.0,,VA:NEXT(190$); ; Model statements for module: Route 112 190$ ROUTE: 0.000000000000000,Floor 02; ; Model statements for module: Dispose 1 1$ ASSIGN: Dispose 1.NumberOut=Dispose 1.NumberOut + 1; 955$ DISPOSE: Yes; ; Model statements for module: Station 111 179$ STATION, Stair_lobby07; 985$ DELAY: 0.0,,VA:NEXT(180$); ; Model statements for module: Route 102 180$ ROUTE: 0.000000000000000,Floor 07; ; Model statements for module: Create 9 986$ CREATE, 1,SecondstoBaseTime(Response),Person:SecondstoBaseTime(POIS( 14.5 )),16:NEXT(987$); 987$ ASSIGN: Passenger Arrival Floor 08.NumberOut=Passenger Arrival Floor 08.NumberOut + 1:NEXT(782$); ; Model statements for module: Assign 67 ; Model statements for module: Decide 1 198$ BRANCH, 1: If,WantedFloor == 37,211$,Yes: If,Floor_entered == 37 && WantedFloor > 37,202$,Yes: If,Floor_entered == 37 && WantedFloor < 37,205$,Yes: If,WantedFloor > 37,208$,Yes: Else,196$,Yes; ; Model statements for module: Seize 79 205$ QUEUE, Floor 37_2.Queue; SEIZE, 2,Other: Floor37_36Stairs,1:NEXT(1269$); 1269$ DELAY: 0.0,,VA:NEXT(204$); ; Model statements for module: Route 117 204$ ROUTE: TRIA( 13.4, 19.6, 24.6 ),Floor 36; ; Model statements for module: Seize 80 208$ QUEUE, Floor 37.Queue; SEIZE, 2,Other: Floor38_37Stairs,1:NEXT(1271$);

281

1271$ DELAY: 0.0,,VA:NEXT(210$); ; Model statements for module: Assign 82 887$ ASSIGN: Floor_entered=29: WantedFloor=1:NEXT(90$); ; Model statements for module: Route 31 90$ ROUTE: EXPO( 20 ),Stair_lobby29; ; Model statements for module: Release 145 871$ RELEASE: Floor02_01Stairs,1:NEXT(868$); ; Model statements for module: Route 293 868$ ROUTE: TRIA( 13.4, 19.6, 24.6 ),Floor 03; ; Model statements for module: Create 43 1787$ CREATE, 1,SecondstoBaseTime(Response),Person:SecondstoBaseTime(POIS( 33 )),6:NEXT(1788$); 1788$ ASSIGN: Passenger Arrival Floor 01.NumberOut=Passenger Arrival Floor 01.NumberOut + 1:NEXT(920$); ; Model statements for module: Assign 110 920$ ASSIGN: Floor_entered=01: WantedFloor= DISC(0.027,38,0.054,37,0.081,36,0.108,35,0.135,34,0.162,33,0.189,32,0.216,31,0.243,30,0.27,29,0.297,28,0.324,27,0.351,26,0.378,25,0.405,24,0.432,23,0.459,22,0.486,21,0.517,20,0.54,19,0.567,18,0.594,17,0.621,16,0.648,15,0.675,14,0.702,13,0.729,12,0.756,11,0.783,10,0.81,9,0.837,8,0.864,7,0.891,6,0.918,5,0.945,4,0.972,3,1,2) :NEXT(918$); ; Model statements for module: Station 155 918$ STATION, Stair_lobby01; 1793$ DELAY: 0.0,,VA:NEXT(919$); ; Model statements for module: Route 296 919$ ROUTE: 0.000000000000000,Floor 01; ; Model statements for module: Create 44 1794$ CREATE, 1,HoursToBaseTime(0.0),TimerEntity:HoursToBaseTime(EXPO(1)):NEXT(1795$); 1795$ ASSIGN: Create 44.NumberOut=Create 44.NumberOut + 1:NEXT(925$); ; Model statements for module: Assign 111 925$ ASSIGN: StartTime=TNOW:NEXT(928$); ; Model statements for module: Hold 1 928$ SCAN: Dispose 1.Numberout == 598:NEXT(926$); ; Model statements for module: Record 1 926$ TALLY: Total System Time,INT(StartTime),1:NEXT(930$); ; Model statements for module: Assign 112 930$ ASSIGN: FinishReplication=1:NEXT(929$); ; Model statements for module: Dispose 2 929$ ASSIGN: Dispose 2.NumberOut=Dispose 2.NumberOut + 1; 1798$ DISPOSE: Yes;

282

0 2F lo o r

A r r iv a lP a s s e n g e r

0 3F lo o r

A r r iv a lP a s s e n g e r

0 4F lo o r

A r r iv a lP a s s e n g e r

2

S t a t io n1

D is p o s e

0 2

A p a r t m e n t

0 3A p a r t m e n t

0 4

A p a r t m e n t

D e la y 1

1R o u t e

D e la y 2

3R o u t e

D e la y 3 4

R o u t e

F in a l E xi t

0 5F lo o r

A r r iv a lP a s s e n g e r

0 6F lo o r

A r r iv a l

P a s s e n g e r

0 7F lo o r

A r r iv a lP a s s e n g e r

0 8F lo o r

A r r iv a lP a s s e n g e r

0 9F lo o r

A r r iv a lP a s s e n g e r

0 6

A p a r t m e n t

0 5

A p a r t m e n t

D e la y 5 6

R o u t e

D e la y 67

R o u t e

0 8

A p a r t m e n t

0 7

A p a r t m e n t

D e la y 78

R o u t e

D e la y 8 9

R o u t e

0 9

A p a r t m e n tD e la y 9 1 0

R o u t e

1 0F lo o r

A r r iv a lP a s s e n g e r

1 1F lo o r

A r r iv a lP a s s e n g e r

1 2F lo o r

A r r iv a lP a s s e n g e r

1 1

A p a r t m e n t

1 0A p a r t m e n t

1 1

D e la y

1 1

R o u t e

1 2D e la y

1 2R o u t e

1 2

A p a r t m e n t

1 3

D e la y

1 3

R o u t e

1 3F lo o r

A r r iv a lP a s s e n g e r

1 4F lo o r

A r r iv a lP a s s e n g e r

1 5F lo o r

A r r iv a lP a s s e n g e r

1 6F lo o r

A r r iv a lP a s s e n g e r

1 7F lo o r

A r r iv a lP a s s e n g e r

1 8F lo o r

A r r iv a lP a s s e n g e r

1 9F lo o r

A r r iv a lP a s s e n g e r

2 0F lo o r

A r r iv a lP a s s e n g e r

2 1F lo o r

A r r iv a lP a s s e n g e r

2 2

F lo o rA r r iv a l

P a s s e n g e r

2 3F lo o r

A r r iv a lP a s s e n g e r

2 4F lo o r

A r r iv a lP a s s e n g e r

2 5F lo o r

A r r iv a lP a s s e n g e r

1 3

A p a r t m e n t

1 4

A p a r t m e n t

1 5

A p a r t m e n t

1 4

D e la y

1 4R o u t e

1 5

D e la y

1 5

R o u t e

1 6

D e la y

1 6

R o u t e

1 7

A p a r t m e n t

1 6

A p a r t m e n t

1 7

D e la y

1 7

R o u t e

1 8

D e la y1 8

R o u t e

1 9

A p a r t m e n t

1 8

A p a r t m e n t

1 9

D e la y

1 9

R o u t e

2 0

D e la y

2 0

R o u t e

2 0

A p a r t m e n t

2 1

D e la y2 1

R o u t e

2 2A p a r t m e n t

2 1A p a r t m e n t

2 2D e la y

2 2

R o u t e

2 3D e la y

2 3

R o u t e

2 3A p a r t m e n t

2 4D e la y

2 4

R o u t e

2 4

A p a r t m e n t

2 5

D e la y

2 5

R o u t e

2 5

A p a r t m e n t

2 6

D e la y

2 6

R o u t e

2 6F lo o r

A r r iv a lP a s s e n g e r

2 7F lo o r

A r r iv a lP a s s e n g e r

2 8F lo o r

A r r iv a lP a s s e n g e r

2 9F lo o r

A r r iv a lP a s s e n g e r

3 0F lo o r

A r r iv a l

P a s s e n g e r

3 1F lo o r

A r r iv a lP a s s e n g e r

3 2F lo o r

A r r iv a lP a s s e n g e r

3 3F lo o r

A r r iv a lP a s s e n g e r

3 4

F lo o rA r r iv a l

P a s s e n g e r

3 5F lo o r

A r r iv a lP a s s e n g e r

3 6F lo o r

A r r iv a lP a s s e n g e r

3 7F lo o r

A r r iv a lP a s s e n g e r

6 5S t a t io n

3 8F lo o r

A r r iv a lP a s s e n g e r

6 6

S t a t io n

2 6

A p a r t m e n t

2 7A p a r t m e n t

2 8A p a r t m e n t

2 7

D e la y2 7

R o u t e

2 8D e la y

2 8R o u t e

2 9D e la y

2 9R o u t e

3 0A p a r t m e n t

2 9A p a r t m e n t

3 0D e la y

3 0R o u t e

3 1D e la y

3 1

R o u t e

3 2

A p a r t m e n t

3 1

A p a r t m e n t

3 2

D e la y3 2

R o u t e

3 3

D e la y

3 3

R o u t e

3 3

A p a r t m e n t

3 4

D e la y3 4

R o u t e

3 5

A p a r t m e n t

3 4A p a r t m e n t

3 5

D e la y3 5

R o u t e

3 6D e la y

3 6R o u t e

3 6

A p a r t m e n t

3 7

D e la y 3 7

R o u t e

3 7

A p a r t m e n t

3 8

D e la y

3 8

R o u t e

3 8

A p a r t m e n t3 9

D e la y

3 9

R o u t e

4 0

R o u t e

8 0S t a t io n

4 1

R o u t e

8 1

S t a t io n4 2

R o u t e

4 3R o u t e

8 2

S t a t io n

4 4

R o u t e

8 3

S t a t io n4 6

R o u t e

8 4

S t a t io n

4 8

R o u t e

8 5S t a t io n

5 0R o u t e

8 6

S t a t io n5 2

R o u t e

8 7

S t a t io n5 4

R o u t e

8 8S t a t io n 5 6

R o u t e

8 9

S t a t io n

5 8

R o u t e

9 0

S t a t io n6 0

R o u t e

9 1S t a t io n

6 2R o u t e

9 2

S t a t io n

6 4R o u t e

9 3S t a t io n

6 6

R o u t e

9 4S t a t io n

6 8R o u t e

9 5

S t a t io n

7 0

R o u t e

9 6

S t a t io n

7 2

R o u t e

9 7

S t a t io n

7 4

R o u t e

9 8S t a t io n

7 6R o u t e

9 9

S t a t io n7 8

R o u t e

1 0 0S t a t io n

8 0R o u t e

1 0 1

S t a t io n

8 2

R o u t e

1 0 2

S t a t io n8 4

R o u t e

1 0 3S t a t io n

8 6R o u t e

1 0 4S t a t io n

8 8R o u t e

1 0 5

S t a t io n9 0

R o u t e

1 0 6

S t a t io n

9 2

R o u t e

1 0 7

S t a t io n

9 4

R o u t e

1 0 8

S t a t io n9 6

R o u t e

1 0 9

S t a t io n9 8

R o u t e

1 1 0

S t a t io n

1 0 0

R o u t e

1 1 1

S t a t io n

1 0 2

R o u t e

1 1 2

S t a t io n

1 0 4

R o u t e

1 1 3

S t a t io n1 0 6

R o u t e

1 1 4

S t a t io n

1 0 8

R o u t e

1 1 5

S t a t io n1 1 0

R o u t e

1 1 6

S t a t io n1 1 2

R o u t e

1 1 7

S t a t io n

1 1 4R o u t e

1

S e iz e

2S e iz e

1R e le a s e

D e c id e 1

W a n t e d F lo o r = = 3 7F lo o r _ e n t e r e d = = 3 7 & & W a n t e d F lo o r > 3 7F lo o r _ e n t e r e d = = 3 7 & & W a n t e d F lo o r < 3 7W a n t e d F lo o r > 3 7E ls e

2A s s ig n

3 9

R e le a s e

1 1 5

R o u t e

7 8

S e iz e

1 1 7

R o u t e

7 9

S e iz e

1 1 8

R o u t e

8 0S e iz e 4 0

R e le a s e

4 1

R e le a s e

1 1 9

R o u t e

3 8A s s ig n

1 2 0

S t a t io n

1 2 0

R o u t e

8 1

S e iz e4 2

R e le a s e

D e c id e 3 8

W a n t e d F lo o r = = 3 6F lo o r _ e n t e r e d = = 3 6 & & W a n t e d F lo o r > 3 6F lo o r _ e n t e r e d = = 3 6 & & W a n t e d F lo o r < 3 6W a n t e d F lo o r > 3 6E ls e

1 2 1

R o u t e8 2

S e iz e

1 2 2

R o u t e

8 3

S e iz e

1 2 3

R o u t e

8 4

S e iz e4 3

R e le a s e

4 4

R e le a s e

1 2 4

R o u t e

1 2 1

S t a t io n

1 2 5

R o u t e

8 5

S e iz e4 5

R e le a s e

D e c id e 3 9

W a n t e d F lo o r = = 3 5F lo o r _ e n t e r e d = = 3 5 & & W a n t e d F lo o r > 3 5F lo o r _ e n t e r e d = = 3 5 & & W a n t e d F lo o r < 3 5W a n t e d F lo o r > 3 5E ls e

1 2 6

R o u t e

8 6

S e iz e

1 2 7

R o u t e

8 7

S e iz e

1 2 8

R o u t e

8 8

S e iz e4 6

R e le a s e

4 7

R e le a s e1 2 9

R o u t e

1 2 2S t a t io n

1 3 0

R o u t e

8 9

S e iz e4 8

R e le a s e

D e c id e 4 0

W a n t e d F lo o r = = 3 4F lo o r _ e n t e r e d = = 3 4 & & W a n t e d F lo o r > 3 4F lo o r _ e n t e r e d = = 3 4 & & W a n t e d F lo o r < 3 4W a n t e d F lo o r > 3 4E ls e

1 3 1R o u t e

9 0S e iz e

1 3 2R o u t e

9 1S e iz e

1 3 3R o u t e

9 2S e iz e

4 9R e le a s e

5 0R e le a s e

1 3 4R o u t e

1 2 3

S t a t io n

1 3 5

R o u t e

9 3

S e iz e5 1

R e le a s e

D e c id e 4 1

W a n t e d F lo o r = = 3 3

F lo o r _ e n t e r e d = = 3 3 & & W a n t e d F lo o r > 3 3

F lo o r _ e n t e r e d = = 3 3 & & W a n t e d F lo o r < 3 3

W a n t e d F lo o r > 3 3E ls e

1 3 6

R o u t e

9 4

S e iz e

1 3 7

R o u t e

9 5

S e iz e

1 3 8

R o u t e

9 6

S e iz e5 2

R e le a s e

5 3

R e le a s e1 3 9

R o u t e

1 2 4

S t a t io n

1 4 0

R o u t e

9 7

S e iz e5 4

R e le a s e

D e c id e 4 2

W a n t e d F lo o r = = 3 2

F lo o r _ e n t e r e d = = 3 2 & & W a n t e d F lo o r > 3 2

F lo o r _ e n t e r e d = = 3 2 & & W a n t e d F lo o r < 3 2

W a n t e d F lo o r > 3 2E ls e

1 4 1

R o u t e

9 8

S e iz e

1 4 2

R o u t e

9 9

S e iz e

1 4 3

R o u t e

1 0 0

S e iz e5 5

R e le a s e

5 6

R e le a s e1 4 4

R o u t e

3 9A s s ig n

4 0A s s ig n

4 1A s s ig n

4 2A s s ig n

4 3A s s ig n

1 2 5

S t a t io n

1 4 5

R o u t e

1 0 1

S e iz e5 7

R e le a s e

D e c id e 4 3

W a n t e d F lo o r = = 3 1

F lo o r _ e n t e r e d = = 3 1 & & W a n t e d F lo o r > 3 1F lo o r _ e n t e r e d = = 3 1 & & W a n t e d F lo o r < 3 1

W a n t e d F lo o r > 3 1E ls e

1 4 6R o u t e

1 0 2S e iz e

1 4 7R o u t e

1 0 3S e iz e

1 4 8R o u t e

1 0 4S e iz e

5 8R e le a s e

5 9R e le a s e

1 4 9

R o u t e

1 2 6

S t a t io n

1 5 0

R o u t e

1 0 5

S e iz e6 0

R e le a s e

D e c id e 4 4

W a n t e d F lo o r = = 3 0

F lo o r _ e n t e r e d = = 3 0 & & W a n t e d F lo o r > 3 0

F lo o r _ e n t e r e d = = 3 0 & & W a n t e d F lo o r < 3 0

W a n t e d F lo o r > 3 0E ls e

1 5 1

R o u t e

1 0 6

S e iz e

1 5 2

R o u t e

1 0 7

S e iz e

1 5 3

R o u t e

1 0 8

S e iz e6 1

R e le a s e

6 2

R e le a s e1 5 4

R o u t e

1 2 7

S t a t io n

1 5 5

R o u t e

1 0 9

S e iz e6 3

R e le a s e

D e c id e 4 5

W a n t e d F lo o r = = 2 9

F lo o r _ e n t e r e d = = 2 9 & & W a n t e d F lo o r > 2 9

F lo o r _ e n t e r e d = = 2 9 & & W a n t e d F lo o r < 2 9

W a n t e d F lo o r > 2 9E ls e

1 5 6

R o u t e

1 1 0

S e iz e

1 5 7

R o u t e

1 1 1

S e iz e

1 5 8

R o u t e

1 1 2

S e iz e6 4

R e le a s e

6 5

R e le a s e1 5 9

R o u t e

1 2 8

S t a t io n

1 6 0

R o u t e

1 1 3

S e iz e6 6

R e le a s e

D e c id e 4 6

W a n t e d F lo o r = = 2 8

F lo o r _ e n t e r e d = = 2 8 & & W a n t e d F lo o r > 2 8

F lo o r _ e n t e r e d = = 2 8 & & W a n t e d F lo o r < 2 8W a n t e d F lo o r > 2 8E ls e

1 6 1

R o u t e

1 1 4

S e iz e

1 6 2

R o u t e

1 1 5

S e iz e

1 6 3R o u t e

1 1 6

S e iz e6 7

R e le a s e

6 8

R e le a s e1 6 4

R o u t e

1 2 9

S t a t io n

1 6 5

R o u t e

1 1 7

S e iz e6 9

R e le a s e

D e c id e 4 7

W a n t e d F lo o r = = 2 7

F lo o r _ e n t e r e d = = 2 7 & & W a n t e d F lo o r > 2 7

F lo o r _ e n t e r e d = = 2 7 & & W a n t e d F lo o r < 2 7W a n t e d F lo o r > 2 7E ls e

1 6 6

R o u t e

1 1 8

S e iz e

1 6 7

R o u t e

1 1 9

S e iz e

1 6 8

R o u t e

1 2 0

S e iz e7 0

R e le a s e

7 1

R e le a s e1 6 9

R o u t e

1 3 0

S t a t io n

1 7 0

R o u t e

1 2 1S e iz e

7 2

R e le a s e

D e c id e 4 8

W a n t e d F lo o r = = 2 6F lo o r _ e n t e r e d = = 2 6 & & W a n t e d F lo o r > 2 6

F lo o r _ e n t e r e d = = 2 6 & & W a n t e d F lo o r < 2 6

W a n t e d F lo o r > 2 6E ls e

1 7 1

R o u t e

1 2 2

S e iz e

1 7 2R o u t e

1 2 3S e iz e

1 7 3

R o u t e

1 2 4

S e iz e7 3

R e le a s e

7 4

R e le a s e1 7 4

R o u t e

4 4A s s ig n

4 5

A s s ig n

4 6A s s ig n

4 7A s s ig n

4 8A s s ig n

4 9A s s ig n

1 3 1S t a t io n

1 7 5R o u t e

1 2 5S e iz e

7 5

R e le a s e

D e c id e 4 9

W a n t e d F lo o r = = 2 5

F lo o r _ e n t e r e d = = 2 5 & & W a n t e d F lo o r > 2 5F lo o r _ e n t e r e d = = 2 5 & & W a n t e d F lo o r < 2 5W a n t e d F lo o r > 2 5E ls e

1 7 6R o u t e

1 2 6S e iz e

1 7 7

R o u t e

1 2 7

S e iz e

1 7 8

R o u t e

1 2 8

S e iz e7 6

R e le a s e

7 7

R e le a s e1 7 9

R o u t e

1 3 2

S t a t io n

1 8 0

R o u t e

1 2 9

S e iz e7 8

R e le a s e

D e c id e 5 0

W a n t e d F lo o r = = 2 4

F lo o r _ e n t e r e d = = 2 4 & & W a n t e d F lo o r > 2 4F lo o r _ e n t e r e d = = 2 4 & & W a n t e d F lo o r < 2 4W a n t e d F lo o r > 2 4E ls e

1 8 1

R o u t e

1 3 0

S e iz e

1 8 2

R o u t e

1 3 1

S e iz e

1 8 3

R o u t e

1 3 2

S e iz e7 9

R e le a s e

8 0

R e le a s e1 8 4

R o u t e

1 3 3

S t a t io n

1 8 5

R o u t e

1 3 3

S e iz e8 1

R e le a s e

D e c id e 5 1

W a n t e d F lo o r = = 2 3

F lo o r _ e n t e r e d = = 2 3 & & W a n t e d F lo o r > 2 3F lo o r _ e n t e r e d = = 2 3 & & W a n t e d F lo o r < 2 3W a n t e d F lo o r > 2 3E ls e

1 8 6

R o u t e

1 3 4

S e iz e

1 8 7

R o u t e

1 3 5

S e iz e

1 8 8

R o u t e

1 3 6

S e iz e8 2

R e le a s e

8 3

R e le a s e1 8 9

R o u t e

1 3 4S t a t io n

1 9 0R o u t e

1 3 7S e iz e

8 4R e le a s e

D e c id e 5 2

W a n t e d F lo o r = = 2 2F lo o r _ e n t e r e d = = 2 2 & & W a n t e d F lo o r > 2 2F lo o r _ e n t e r e d = = 2 2 & & W a n t e d F lo o r < 2 2

W a n t e d F lo o r > 2 2E ls e

1 9 1

R o u t e

1 3 8

S e iz e

1 9 2

R o u t e

1 3 9

S e iz e

1 9 3

R o u t e

1 4 0

S e iz e8 5

R e le a s e

8 6

R e le a s e1 9 4

R o u t e

1 3 5

S t a t io n

1 9 5

R o u t e

1 4 1

S e iz e8 7

R e le a s e

D e c id e 5 3

W a n t e d F lo o r = = 2 1F lo o r _ e n t e r e d = = 2 1 & & W a n t e d F lo o r > 2 1F lo o r _ e n t e r e d = = 2 1 & & W a n t e d F lo o r < 2 1W a n t e d F lo o r > 2 1E ls e

1 9 6

R o u t e

1 4 2

S e iz e

1 9 7

R o u t e

1 4 3

S e iz e

1 9 8

R o u t e

1 4 4

S e iz e8 8

R e le a s e

8 9

R e le a s e1 9 9

R o u t e

1 3 6

S t a t io n

2 0 0

R o u t e

1 4 5

S e iz e9 0

R e le a s e

D e c id e 5 4

W a n t e d F lo o r = = 2 0F lo o r _ e n t e r e d = = 2 0 & & W a n t e d F lo o r > 2 0F lo o r _ e n t e r e d = = 2 0 & & W a n t e d F lo o r < 2 0W a n t e d F lo o r > 2 0E ls e

2 0 1

R o u t e

1 4 6

S e iz e

2 0 2

R o u t e

1 4 7

S e iz e

2 0 3

R o u t e

1 4 8

S e iz e9 1

R e le a s e

9 2

R e le a s e2 0 4

R o u t e

1 3 7

S t a t io n

2 0 5

R o u t e

1 4 9

S e iz e9 3

R e le a s e

D e c id e 5 5

W a n t e d F lo o r = = 1 9

F lo o r _ e n t e r e d = = 1 9 & & W a n t e d F lo o r > 1 9

F lo o r _ e n t e r e d = = 1 9 & & W a n t e d F lo o r < 1 9

W a n t e d F lo o r > 1 9E ls e

2 0 6R o u t e

1 5 0S e iz e

2 0 7R o u t e

1 5 1S e iz e

2 0 8R o u t e

1 5 2S e iz e

9 4R e le a s e

9 5R e le a s e

2 0 9

R o u t e

5 0A s s ig n

5 1A s s ig n

5 2A s s ig n

5 3A s s ig n

5 4A s s ig n

5 5

A s s ig n

5 6A s s ig n

1 3 8

S t a t io n

2 1 0

R o u t e

1 5 3

S e iz e9 6

R e le a s e

D e c id e 5 6

W a n t e d F lo o r = = 1 8F lo o r _ e n t e r e d = = 1 8 & & W a n t e d F lo o r > 1 8

F lo o r _ e n t e r e d = = 1 8 & & W a n t e d F lo o r < 1 8

W a n t e d F lo o r > 1 8E ls e

2 1 1

R o u t e

1 5 4

S e iz e

2 1 2

R o u t e

1 5 5

S e iz e

2 1 3

R o u t e

1 5 6

S e iz e9 7

R e le a s e

9 8

R e le a s e2 1 4

R o u t e

1 3 9

S t a t io n

2 1 5

R o u t e

1 5 7

S e iz e9 9

R e le a s e

D e c id e 5 7

W a n t e d F lo o r = = 1 7

F lo o r _ e n t e r e d = = 1 7 & & W a n t e d F lo o r > 1 7

F lo o r _ e n t e r e d = = 1 7 & & W a n t e d F lo o r < 1 7

W a n t e d F lo o r > 1 7E ls e

2 1 6

R o u t e

1 5 8

S e iz e

2 1 7

R o u t e

1 5 9

S e iz e

2 1 8

R o u t e

1 6 0

S e iz e1 0 0

R e le a s e

1 0 1

R e le a s e2 1 9

R o u t e

1 4 0

S t a t io n

2 2 0

R o u t e

1 6 1

S e iz e1 0 2

R e le a s e

D e c id e 5 8

W a n t e d F lo o r = = 1 6

F lo o r _ e n t e r e d = = 1 6 & & W a n t e d F lo o r > 1 6

F lo o r _ e n t e r e d = = 1 6 & & W a n t e d F lo o r < 1 6

W a n t e d F lo o r > 1 6E ls e

2 2 1

R o u t e

1 6 2

S e iz e

2 2 2

R o u t e

1 6 3

S e iz e

2 2 3

R o u t e

1 6 4

S e iz e1 0 3

R e le a s e

1 0 4

R e le a s e2 2 4

R o u t e

1 4 1

S t a t io n

2 2 5

R o u t e

1 6 5

S e iz e1 0 5

R e le a s e

D e c id e 5 9

W a n t e d F lo o r = = 1 5

F lo o r _ e n t e r e d = = 1 5 & & W a n t e d F lo o r > 1 5

F lo o r _ e n t e r e d = = 1 5 & & W a n t e d F lo o r < 1 5W a n t e d F lo o r > 1 5E ls e

2 2 6

R o u t e

1 6 6

S e iz e

2 2 7

R o u t e

1 6 7

S e iz e

2 2 8R o u t e

1 6 8S e iz e

1 0 6R e le a s e

1 0 7R e le a s e

2 2 9R o u t e

1 4 2

S t a t io n

2 3 0

R o u t e

1 6 9

S e iz e1 0 8

R e le a s e

D e c id e 6 0

W a n t e d F lo o r = = 1 4

F lo o r _ e n t e r e d = = 1 4 & & W a n t e d F lo o r > 1 4

F lo o r _ e n t e r e d = = 1 4 & & W a n t e d F lo o r < 1 4W a n t e d F lo o r > 1 4E ls e

2 3 1R o u t e

1 7 0S e iz e

2 3 2R o u t e

1 7 1S e iz e

2 3 3R o u t e

1 7 2S e iz e

1 0 9R e le a s e

1 1 0R e le a s e

2 3 4R o u t e

1 4 3

S t a t io n

2 3 5R o u t e

1 7 3S e iz e

1 1 1

R e le a s e

D e c id e 6 1

W a n t e d F lo o r = = 1 3F lo o r _ e n t e r e d = = 1 3 & & W a n t e d F lo o r > 1 3

F lo o r _ e n t e r e d = = 1 3 & & W a n t e d F lo o r < 1 3W a n t e d F lo o r > 1 3E ls e

2 3 6

R o u t e

1 7 4

S e iz e

2 3 7R o u t e

1 7 5

S e iz e

2 3 8R o u t e

1 7 6S e iz e

1 1 2

R e le a s e

1 1 3

R e le a s e2 3 9

R o u t e

1 4 4

S t a t io n

2 4 0R o u t e

1 7 7S e iz e

1 1 4R e le a s e

D e c id e 6 2

W a n t e d F lo o r = = 1 2

F lo o r _ e n t e r e d = = 1 2 & & W a n t e d F lo o r > 1 2

F lo o r _ e n t e r e d = = 1 2 & & W a n t e d F lo o r < 1 2

W a n t e d F lo o r > 1 2E ls e

2 4 1

R o u t e

1 7 8

S e iz e

2 4 2

R o u t e

1 7 9

S e iz e

2 4 3R o u t e

1 8 0

S e iz e1 1 5

R e le a s e

1 1 6

R e le a s e 2 4 4

R o u t e

5 7A s s ig n

5 8A s s ig n

5 9A s s ig n

6 0A s s ig n

6 1A s s ig n

6 2

A s s ig n

6 3A s s ig n

1 4 5

S t a t io n

2 4 5

R o u t e

1 8 1

S e iz e1 1 7

R e le a s e

D e c id e 6 3

W a n t e d F lo o r = = 1 1F lo o r _ e n t e r e d = = 1 1 & & W a n t e d F lo o r > 1 1F lo o r _ e n t e r e d = = 1 1 & & W a n t e d F lo o r < 1 1W a n t e d F lo o r > 1 1E ls e

2 4 6

R o u t e

1 8 2

S e iz e

2 4 7

R o u t e

1 8 3

S e iz e

2 4 8R o u t e

1 8 4S e iz e

1 1 8R e le a s e

1 1 9

R e le a s e2 4 9

R o u t e

1 4 6

S t a t io n

2 5 0

R o u t e

1 8 5

S e iz e1 2 0

R e le a s e

D e c id e 6 4

W a n t e d F lo o r = = 1 0F lo o r _ e n t e r e d = = 1 0 & & W a n t e d F lo o r > 1 0F lo o r _ e n t e r e d = = 1 0 & & W a n t e d F lo o r < 1 0W a n t e d F lo o r > 1 0E ls e

2 5 1

R o u t e

1 8 6

S e iz e

2 5 2

R o u t e

1 8 7

S e iz e

2 5 3

R o u t e

1 8 8

S e iz e1 2 1

R e le a s e

1 2 2

R e le a s e2 5 4

R o u t e

1 4 7

S t a t io n

2 5 5

R o u t e

1 8 9

S e iz e1 2 3

R e le a s e

D e c id e 6 5

W a n t e d F lo o r = = 0 9F lo o r _ e n t e r e d = = 0 9 & & W a n t e d F lo o r > 0 9F lo o r _ e n t e r e d = = 0 9 & & W a n t e d F lo o r < 0 9W a n t e d F lo o r > 0 9E ls e

2 5 6

R o u t e

1 9 0

S e iz e

2 5 7

R o u t e

1 9 1

S e iz e

2 5 8

R o u t e

1 9 2

S e iz e1 2 4

R e le a s e

1 2 5

R e le a s e

2 5 9

R o u t e

1 4 8

S t a t io n

2 6 0

R o u t e

1 9 3

S e iz e1 2 6

R e le a s e

D e c id e 6 6

W a n t e d F lo o r = = 0 8F lo o r _ e n t e r e d = = 0 8 & & W a n t e d F lo o r > 0 8

F lo o r _ e n t e r e d = = 0 8 & & W a n t e d F lo o r < 0 8W a n t e d F lo o r > 0 8E ls e

2 6 1

R o u t e

1 9 4

S e iz e

2 6 2

R o u t e

1 9 5

S e iz e

2 6 3

R o u t e

1 9 6

S e iz e1 2 7

R e le a s e

1 2 8

R e le a s e2 6 4

R o u t e

1 4 9

S t a t io n

2 6 5

R o u t e1 9 7

S e iz e

1 2 9

R e le a s e

D e c id e 6 7

W a n t e d F lo o r = = 0 7

F lo o r _ e n t e r e d = = 0 7 & & W a n t e d F lo o r > 0 7F lo o r _ e n t e r e d = = 0 7 & & W a n t e d F lo o r < 0 7W a n t e d F lo o r > 0 7E ls e

2 6 6R o u t e

1 9 8S e iz e

2 6 7

R o u t e

1 9 9

S e iz e

2 6 8

R o u t e

2 0 0

S e iz e1 3 0

R e le a s e

1 3 1

R e le a s e2 6 9

R o u t e

6 4A s s ig n

6 5A s s ig n

6 6A s s ig n

6 7A s s ig n

6 8A s s ig n

1 5 0

S t a t io n

2 7 0

R o u t e2 0 1

S e iz e

1 3 2

R e le a s e

D e c id e 6 8

W a n t e d F lo o r = = 0 6

F lo o r _ e n t e r e d = = 0 6 & & W a n t e d F lo o r > 0 6

F lo o r _ e n t e r e d = = 0 6 & & W a n t e d F lo o r < 0 6

W a n t e d F lo o r > 0 6E ls e

2 7 1

R o u t e

2 0 2

S e iz e

2 7 2

R o u t e

2 0 3

S e iz e

2 7 3

R o u t e

2 0 4

S e iz e1 3 3

R e le a s e

1 3 4

R e le a s e2 7 4

R o u t e

1 5 1S t a t io n

2 7 5R o u t e

2 0 5S e iz e

1 3 5R e le a s e

D e c id e 6 9

W a n t e d F lo o r = = 0 5

F lo o r _ e n t e r e d = = 0 5 & & W a n t e d F lo o r > 0 5

F lo o r _ e n t e r e d = = 0 5 & & W a n t e d F lo o r < 0 5

W a n t e d F lo o r > 0 5E ls e

2 7 6R o u t e

2 0 6S e iz e

2 7 7R o u t e

2 0 7S e iz e

2 7 8R o u t e

2 0 8S e iz e

1 3 6R e le a s e

1 3 7R e le a s e 2 7 9

R o u t e

1 5 2

S t a t io n

2 8 0

R o u t e2 0 9

S e iz e

1 3 8

R e le a s e

D e c id e 7 0

W a n t e d F lo o r = = 0 4

F lo o r _ e n t e r e d = = 0 4 & & W a n t e d F lo o r > 0 4

F lo o r _ e n t e r e d = = 0 4 & & W a n t e d F lo o r < 0 4

W a n t e d F lo o r > 0 4E ls e

2 8 1

R o u t e

2 1 0

S e iz e

2 8 2

R o u t e

2 1 1

S e iz e

2 8 3

R o u t e

2 1 2

S e iz e1 3 9

R e le a s e

1 4 0

R e le a s e2 8 4

R o u t e

1 5 3

S t a t io n

2 8 5

R o u t e2 1 3

S e iz e

1 4 1

R e le a s e

D e c id e 7 1

W a n t e d F lo o r = = 0 3

F lo o r _ e n t e r e d = = 0 3 & & W a n t e d F lo o r > 0 3

F lo o r _ e n t e r e d = = 0 3 & & W a n t e d F lo o r < 0 3

W a n t e d F lo o r > 0 3E ls e

2 8 6

R o u t e

2 1 4

S e iz e

2 8 7

R o u t e

2 1 5

S e iz e

2 8 8

R o u t e2 1 6

S e iz e1 4 2

R e le a s e

1 4 3

R e le a s e2 8 9

R o u t e

1 5 4

S t a t io n

2 9 0R o u t e

2 1 7

S e iz e

1 4 4

R e le a s e

D e c id e 7 2

W a n t e d F lo o r = = 0 2

F lo o r _ e n t e r e d = = 0 2 & & W a n t e d F lo o r > 0 2

F lo o r _ e n t e r e d = = 0 2 & & W a n t e d F lo o r < 0 2W a n t e d F lo o r > 0 2E ls e

2 9 1

R o u t e

2 1 8

S e iz e

2 9 2

R o u t e

2 1 9

S e iz e

2 9 3

R o u t e

2 2 0

S e iz e1 4 5

R e le a s e

1 4 6

R e le a s e2 9 4

R o u t e

6 9A s s ig n

7 0A s s ig n

7 1A s s ig n

7 2A s s ig n

7 3A s s ig n

7 4A s s ig n

7 5A s s ig n

7 6A s s ig n

7 7A s s ig n

7 8A s s ig n

7 9A s s ig n

8 0A s s ig n

8 1A s s ig n

8 2A s s ig n

8 3A s s ig n

8 4A s s ig n

8 5A s s ig n

8 6A s s ig n

8 7A s s ig n

8 8A s s ig n

8 9A s s ig n

9 0A s s ig n

9 1A s s ig n

9 2A s s ig n

9 3A s s ig n

9 4A s s ig n

9 5A s s ig n

9 6A s s ig n

9 7A s s ig n

9 8A s s ig n

9 9

A s s ig n

1 0 0A s s ig n

1 0 1A s s ig n

1 0 2A s s ig n

1 0 3A s s ig n

1 0 4A s s ig n

1 0 5A s s ig n

1 0 6A s s ig n

1 0 7A s s ig n

1 0 8A s s ig n

1 0 9A s s ig n

D e c id e 7 3

W a n t e d F lo o r = = 3 8

E ls e

1 4 7

R e le a s e

2 9 5

R o u t e

0 1F lo o r

A r r iv a lP a s s e n g e r

1 5 5

S t a t io n2 9 6

R o u t e

1 1 0A s s ig n

D e c id e 7 4

F lo o r _ e n t e r e d = = 0 1 & & W a n t e d F lo o r > 0 1

E ls e

2 9 7R o u t e

2 2 1S e iz e

4 4C r e a t e

1 1 1A s s ig n

1R e c o r d

H o ld 12

D is p o s e

1 1 2A s s ig n

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0 0

Figure F2 – Overall Model Used for Stair Simulation

283

Appendix G

ARENA Results

Table G1: Lift Waiting Time, tLW (100 replications)

Lift Waiting Time-100% population Lift Waiting Time-25% population

2nd

floor fire 19th

floor fire 38th

floor fire 2nd

floor fire 19th

floor fire 38th

floor fire

717.06 707.48 680.26 313.37 298.74 343.14

725.01 722.97 718.92 312.24 246.39 244.45

658.28 644.88 675.10 335.08 264.66 272.30

770.99 804.44 766.92 302.16 296.53 287.19

760.50 653.93 725.87 259.76 304.24 275.68

601.69 679.86 678.05 211.23 276.04 266.97

731.41 743.00 705.11 303.64 316.57 311.80

731.52 751.88 658.73 296.09 243.80 298.96

705.94 683.51 689.57 273.11 245.58 253.95

667.20 727.76 726.46 301.28 325.45 269.07

695.39 720.55 813.62 283.76 299.83 283.49

702.97 656.49 804.02 322.37 311.53 348.73

672.46 677.19 715.59 265.54 279.50 277.75

754.20 640.13 718.01 299.77 309.55 307.19

669.93 729.94 685.37 306.17 299.44 246.41

716.58 708.03 630.57 308.97 267.09 300.64

700.94 783.57 800.97 283.03 307.39 265.46

800.59 733.33 764.56 345.33 351.36 309.30

696.85 659.46 688.20 259.62 295.17 315.40

721.70 625.47 658.39 321.09 289.32 273.38

703.36 737.21 788.47 264.79 323.27 304.78

681.01 747.12 660.68 344.61 327.18 315.63

734.25 760.89 715.62 345.32 320.99 303.44

733.91 746.61 706.69 368.90 317.54 335.41

723.56 716.90 700.41 326.93 268.60 273.90

722.59 792.07 743.59 284.95 267.11 380.21

789.46 697.49 800.16 326.82 301.01 312.27

710.00 763.75 796.73 292.52 304.03 258.58

757.26 732.68 702.82 378.35 305.61 332.54

741.60 732.17 739.09 343.29 301.91 378.46

676.32 732.37 694.68 247.31 329.95 301.63

721.48 758.92 777.81 324.62 325.85 337.77

723.91 766.84 745.47 352.79 321.28 285.83

699.70 717.51 766.56 305.60 318.27 328.38

743.45 753.32 747.34 365.18 304.96 323.90

737.23 810.20 728.20 302.93 260.89 266.89

704.74 682.95 647.58 424.95 303.55 292.86

723.70 696.30 704.78 251.19 281.30 319.17

716.62 724.86 687.46 295.39 283.77 309.67

671.66 710.74 801.20 307.98 285.45 299.79

701.37 711.95 679.68 310.75 262.04 254.31

733.87 741.76 771.82 321.03 248.50 301.17

801.11 798.97 727.75 320.67 282.36 315.73

284

775.36 758.59 692.43 305.14 336.14 325.10

795.21 805.87 727.97 301.11 387.43 320.72

778.93 774.57 728.59 276.53 232.71 299.75

672.04 735.83 756.77 310.32 326.09 351.58

748.42 758.38 695.23 271.16 252.91 314.67

773.02 684.56 757.03 289.44 285.53 327.43

701.26 670.80 719.57 250.85 266.40 328.78

767.12 739.28 750.75 361.09 309.77 260.17

744.89 680.45 804.48 264.26 284.56 320.59

775.04 746.46 839.67 257.48 361.35 316.69

716.03 724.70 712.29 259.36 294.52 297.74

770.43 746.73 690.23 280.91 332.62 286.41

726.98 708.12 691.51 251.51 267.15 253.86

827.20 702.55 749.26 293.79 357.46 307.16

717.48 699.50 714.20 293.18 256.43 257.46

639.08 737.43 744.89 248.45 294.64 403.91

708.36 710.21 731.75 286.54 252.37 323.18

721.75 713.94 693.47 245.33 249.17 242.10

691.20 669.66 735.78 265.54 324.78 281.92

677.66 743.48 745.23 279.27 308.13 265.23

709.14 702.62 720.63 260.34 337.19 270.83

678.89 706.64 676.16 313.17 280.54 289.55

663.70 736.38 744.95 288.58 240.41 299.71

682.50 714.74 709.72 351.29 324.11 249.33

724.83 736.44 705.94 324.07 374.12 327.41

693.76 631.04 725.52 273.63 244.73 284.36

800.03 610.03 706.41 316.83 317.17 290.15

701.96 726.58 760.93 292.17 363.09 267.22

729.07 721.77 717.04 268.67 324.04 288.30

725.58 729.48 712.90 300.11 265.53 279.31

691.39 728.75 730.41 313.85 271.47 305.21

719.16 709.72 773.34 289.39 317.90 332.44

770.10 731.21 701.21 377.65 242.89 241.85

694.05 672.24 730.05 374.72 373.00 324.86

737.17 695.17 714.54 306.79 344.81 321.72

734.95 724.36 741.33 314.24 313.30 289.00

688.88 717.12 740.35 353.24 270.71 318.10

694.42 679.82 645.94 272.21 287.42 279.04

736.61 716.94 715.91 308.34 316.05 346.02

674.29 681.96 652.29 315.76 361.87 298.28

715.95 735.89 657.46 317.10 306.95 325.93

723.45 723.34 809.09 267.81 258.25 312.98

682.29 708.60 674.14 284.29 237.85 257.28

785.61 740.96 702.46 312.22 318.80 345.54

718.56 772.53 711.36 289.91 233.23 284.86

693.59 736.10 655.71 244.52 329.05 244.72

725.73 666.55 669.65 295.80 208.79 244.16

796.37 747.45 761.53 373.63 298.12 301.48

704.39 761.96 767.03 278.06 304.32 362.18

723.76 708.64 657.46 256.01 297.85 266.25

730.90 704.17 727.93 283.14 229.12 279.18

770.62 753.80 768.18 279.22 311.67 280.99

285

675.95 761.74 680.12 252.89 276.68 299.11

791.87 692.73 754.97 319.38 298.01 281.13

636.12 687.96 605.19 281.61 265.73 308.02

707.97 698.94 710.22 289.56 270.65 268.41

709.86 692.60 736.57 249.04 226.63 279.65

Table G2: Lift Transportation Time, tLT (100 replications)

Lift Transportation Time -100% population Lift Transportation Time - 25% population

2nd

floor fire 19th

floor fire 38th

floor fire 2nd

floor fire 19th

floor fire 38th

floor fire

42.68 40.06 42.90 42.28 37.96 39.35

41.45 42.37 42.62 41.29 39.84 41.81

41.85 39.95 41.22 38.32 38.73 39.59

43.60 43.79 41.72 38.08 39.11 38.78

42.92 41.32 42.06 42.13 40.44 40.37

40.96 41.55 41.04 39.68 40.89 39.88

41.54 42.03 40.66 39.83 40.76 39.76

42.98 42.74 40.37 40.48 40.80 39.04

42.23 42.78 42.40 39.84 42.93 40.23

40.83 41.56 41.27 39.21 43.00 40.06

43.14 42.24 40.96 40.14 38.98 38.08

40.56 39.80 41.74 37.81 41.64 40.06

41.21 44.30 44.47 40.33 40.00 41.50

43.80 41.90 41.91 39.40 43.36 40.92

41.57 42.08 40.73 37.76 39.45 40.07

41.86 43.10 40.91 40.50 40.61 40.68

40.60 43.20 42.71 39.63 40.29 40.50

42.59 42.34 41.06 41.56 40.37 38.69

41.47 42.37 42.90 40.65 43.24 38.74

41.57 40.37 40.51 38.61 41.39 38.45

42.95 41.82 43.93 40.16 39.31 44.73

41.28 42.94 43.00 38.25 40.89 40.18

42.47 42.79 41.74 41.09 41.52 42.24

41.89 42.89 42.55 40.23 38.44 41.32

41.02 39.68 40.58 38.92 36.48 37.64

41.03 43.23 41.90 39.78 38.92 39.71

42.08 41.77 44.37 40.50 39.30 39.07

40.89 42.31 41.54 37.72 43.90 41.69

42.39 41.33 41.32 40.35 39.16 40.34

42.89 40.19 41.86 39.44 39.77 40.75

41.76 42.82 41.47 40.68 37.80 45.43

42.45 44.36 44.77 37.72 40.75 44.73

41.39 42.28 42.50 41.63 40.61 38.60

40.87 43.82 42.98 43.36 40.59 39.19

42.37 41.18 40.15 41.52 39.39 37.24

41.42 39.83 43.71 40.27 40.00 38.37

40.90 42.11 40.58 40.12 38.90 38.86

40.68 42.08 41.15 39.41 38.61 39.15

41.97 41.71 42.62 42.03 40.99 40.16

41.73 42.27 44.65 44.17 41.41 41.57

286

42.21 42.06 41.17 40.03 39.44 41.73

39.74 40.77 42.35 40.81 41.04 39.81

42.53 43.47 40.91 41.60 39.22 39.81

43.17 41.45 42.70 37.73 39.04 42.51

41.43 43.27 40.39 39.09 41.67 41.72

41.86 41.98 41.87 38.43 40.45 38.49

41.06 41.89 40.15 39.19 37.80 40.51

41.48 43.84 41.96 40.93 43.36 41.47

44.31 41.51 42.40 39.41 39.86 39.82

40.44 39.75 40.42 39.36 40.08 40.83

43.55 42.83 45.12 37.87 39.97 38.48

42.58 42.53 42.39 39.50 40.75 41.18

42.20 40.10 42.68 41.08 41.29 40.60

40.82 40.84 41.89 39.37 42.06 40.45

42.60 43.83 41.66 41.36 40.18 44.70

40.76 41.45 40.24 43.24 39.48 38.73

42.96 40.10 40.33 39.25 41.07 40.08

43.51 41.78 39.72 38.63 39.20 36.92

41.54 42.82 42.81 39.90 40.33 41.14

40.21 42.46 42.18 40.16 38.49 39.52

41.83 41.77 40.85 40.17 42.12 40.37

41.78 42.65 40.00 41.41 42.76 40.35

41.64 43.79 41.10 39.72 38.97 40.58

42.14 41.49 42.99 39.61 40.46 41.75

40.42 43.43 39.73 40.98 40.02 38.89

42.33 43.23 39.63 40.39 40.48 39.72

42.61 44.38 41.49 39.36 43.59 42.07

41.21 42.18 42.02 40.35 40.03 41.39

41.38 41.14 42.77 38.99 39.94 38.93

42.40 42.20 41.69 42.65 39.06 45.29

40.84 42.18 40.15 40.40 40.66 37.83

43.86 43.63 41.06 38.84 43.71 39.50

43.12 41.32 43.25 41.60 38.87 42.43

40.70 40.45 42.06 41.12 37.32 40.82

39.78 42.36 42.48 40.11 40.10 41.67

42.67 41.59 41.76 39.98 39.98 38.70

42.61 41.70 41.17 44.08 40.44 42.90

42.88 40.85 42.37 39.15 39.94 38.99

42.71 41.42 41.57 43.00 43.08 41.10

41.25 41.69 41.99 40.53 41.85 44.87

41.87 41.50 40.61 41.30 39.66 38.71

42.39 41.22 42.79 39.52 39.84 39.04

43.04 42.38 40.28 40.44 41.50 40.06

41.75 41.08 41.59 39.41 41.22 39.50

40.39 41.18 43.20 38.46 41.32 38.69

40.25 40.83 42.20 38.72 38.06 38.75

43.12 42.14 42.12 41.83 40.54 40.44

41.75 41.46 41.07 40.31 40.83 39.72

40.45 42.59 41.88 39.01 39.32 40.86

40.15 40.76 40.37 38.62 40.22 39.78

43.25 42.45 43.84 39.44 41.74 39.84

40.66 43.23 44.16 41.03 38.86 41.55

287

41.28 39.50 41.63 38.91 42.25 40.08

42.55 41.82 43.27 37.55 39.75 40.31

43.91 41.96 44.63 36.62 38.29 38.86

42.44 40.95 41.58 40.44 38.93 39.03

40.13 40.81 41.28 39.73 40.56 40.71

40.56 41.38 39.58 39.25 37.67 41.10

40.17 40.60 40.10 40.07 39.70 43.34

42.77 41.64 42.11 44.14 41.61 41.60

Table G3: Lift Evacuation Time, tET (100 replications)

Lift Evacuation Time -100% population Lift Evacuation Time - 25% population

2nd

floor fire 19th

floor fire 38th

floor fire 2nd

floor fire 19th

floor fire 38th

floor fire

2748.98 1829.56 2398.57 1275.75 1479.93 1333.93

2635.33 2274.70 2688.76 1669.52 1535.24 1279.55

1404.26 1428.23 1374.50 1702.19 1461.67 1479.06

2228.93 2126.23 2649.70 1443.70 1065.72 1290.06

2803.96 2396.17 2379.92 1086.68 1194.01 1755.74

1265.64 1499.16 1796.80 1036.52 905.33 881.96

1451.18 2051.85 1901.64 1563.82 987.10 939.05

2698.57 2826.74 2647.91 1730.38 1839.23 1824.76

2124.91 2091.09 2447.28 1476.18 1667.07 1680.70

2347.72 2223.33 1536.06 1467.12 1174.27 926.69

1349.78 1532.20 2008.87 1066.32 1003.41 885.08

2173.08 2663.32 2711.77 2208.90 1635.43 1452.51

2345.70 1687.92 1641.86 947.56 1296.58 1682.40

1957.81 1736.60 1900.42 1313.38 1157.68 1195.61

2192.14 2091.93 2185.29 1352.64 1257.25 1123.65

3052.42 2526.62 2418.75 2319.05 1848.81 1773.11

2667.24 3151.79 3403.47 2144.00 1806.90 1904.40

2918.06 2931.97 2815.44 1921.98 1715.97 2289.45

2611.97 2723.74 2256.57 1910.74 1940.82 2179.01

1642.97 1447.17 1331.00 1836.94 1972.33 1818.53

2098.70 2303.09 2195.80 1085.96 1145.53 1290.91

1598.25 1536.03 1450.76 1300.11 1002.93 1943.75

1450.30 2834.46 2865.11 1971.33 1274.44 1309.57

2004.97 1988.23 2291.11 1187.24 1073.05 1132.88

2432.96 2461.74 2473.95 1698.10 1877.70 1480.63

2369.74 2888.65 3004.48 1787.52 1972.95 2039.00

2260.39 2078.97 2162.86 1555.88 1418.15 1023.99

1158.68 1682.84 1688.64 1104.77 1793.97 1048.06

1998.88 2083.99 1784.71 1688.79 1626.60 1216.72

2724.10 3218.29 2840.22 1845.61 1637.61 1926.93

2860.65 3259.40 3000.33 2197.94 1706.69 2229.96

2688.81 1763.12 2199.51 1292.89 1722.17 1772.07

2058.49 2183.99 1180.71 1134.09 929.34 1101.41

2292.07 2073.96 2075.12 982.78 1319.35 876.05

2618.82 1782.50 2468.49 1517.73 933.53 979.03

2449.68 1294.04 2090.32 879.14 1326.09 1088.38

3123.82 2729.22 3176.55 2083.59 1956.05 1784.64

288

2100.82 2856.92 1696.81 1314.72 1585.28 1380.83

2968.50 3134.87 3119.61 2189.22 1677.14 1475.24

2791.37 2401.12 2960.13 1792.10 2059.58 1770.51

3038.19 2821.86 2889.85 2008.04 1620.32 1675.65

2810.35 2783.68 2706.93 1892.29 1550.87 1807.92

1414.38 1984.56 1398.76 1604.93 1754.31 1308.60

3169.29 3146.26 3142.99 1996.30 1817.96 1880.70

2327.73 1528.70 2028.06 1311.58 2136.79 1960.63

1898.25 2163.22 1290.36 1664.52 1749.13 1007.37

2742.92 2689.70 2576.40 1700.24 1834.16 1875.44

2999.73 2980.68 2725.57 1948.80 1654.67 2238.84

2688.22 3205.78 2872.52 1802.84 1772.29 1783.66

2548.78 1591.29 1860.39 1744.16 1165.15 973.96

2346.99 1684.07 1603.14 1779.27 2051.86 1143.36

2122.99 1922.12 1456.00 1017.99 1713.43 1379.04

2378.67 1375.47 1459.78 1551.53 2401.98 1667.39

2328.07 2230.80 1389.28 1450.69 1491.30 1422.67

2574.80 2640.07 3125.31 2242.60 1822.34 1626.70

2926.59 3021.87 2870.62 1720.54 1559.02 1930.16

2598.03 2788.79 2557.93 1515.18 1663.16 2200.65

2071.92 1983.43 1615.47 1629.82 1226.48 1360.20

2779.99 2766.19 3035.95 1968.01 1715.24 2154.41

3053.41 2859.22 2651.62 1999.52 1858.62 1912.35

2171.84 2504.12 2235.32 1218.39 1245.64 1338.76

1930.69 1634.92 1832.97 1979.95 1675.04 1274.52

2599.34 1815.35 1727.01 1635.20 1631.10 1263.74

1821.21 1330.27 1514.86 1572.93 1760.72 1813.69

2457.27 2825.51 2638.22 2085.83 1411.87 1449.43

2868.34 2761.14 2921.05 1667.21 1496.94 1755.88

2846.87 2823.32 2974.45 2003.51 1846.52 2051.28

2087.56 1749.80 1936.19 1514.57 1021.52 1196.39

1499.52 1358.98 1603.47 1782.39 1953.74 1677.68

1356.48 1176.37 1302.67 1829.33 2170.80 2206.20

3080.85 2548.14 2871.00 1308.35 1887.17 1908.06

1714.32 2511.72 2543.82 1613.34 1163.76 1201.35

2668.95 2747.69 2922.56 1858.40 2040.17 1899.46

1484.19 1539.31 1487.31 1721.73 1886.33 2306.55

1348.12 2004.62 1366.32 1833.20 2076.86 2044.47

3109.78 3291.86 2652.86 1983.47 1632.10 1599.81

1605.60 1740.47 2125.40 1010.74 1365.36 1097.46

1997.76 2056.48 1913.77 987.61 1354.51 1717.49

2388.80 2414.14 1919.91 1461.63 2007.37 1269.63

2682.85 2466.06 3053.14 2160.09 1988.92 1841.18

2622.52 2784.52 2436.25 1512.33 1951.86 1756.12

2611.89 2521.68 2644.98 2216.76 1962.19 2128.72

2156.90 1912.13 2399.05 1123.52 742.13 874.57

3082.70 2567.67 2902.78 2051.41 2059.86 2132.75

1795.42 1993.60 1419.23 1585.41 1361.48 1403.25

2043.86 2397.74 2325.84 1314.55 1301.58 1431.19

1948.83 1879.29 2765.37 1922.20 1364.03 1366.65

3008.02 2677.97 2430.14 1435.40 1768.96 1548.09

2123.80 2948.91 2426.60 918.71 1561.59 912.21

289

1625.44 1777.03 1613.06 1199.74 706.84 993.96

2703.63 2619.85 2690.75 1983.81 2096.21 1720.44

2043.86 1941.13 2213.71 1125.29 1194.34 1398.99

2628.61 2580.49 2889.92 1304.82 1804.65 1732.62

1636.04 1318.32 1496.65 1864.18 1852.78 2146.17

2598.83 2449.73 2598.38 1732.65 1773.12 1384.40

2892.66 3124.70 2805.89 2138.55 1568.79 2293.19

1592.05 2234.02 2470.24 785.97 1630.34 1216.28

2330.46 1533.93 1662.40 1608.34 1404.90 1638.59

3145.59 2566.52 3167.82 2052.15 2118.76 1914.47

2372.74 1836.73 1622.27 1584.53 1273.48 1312.46

Table G4: Stair Travelling Time, tST (100 replications)

Stair Travelling Time (100% Population) Stair Travelling Time (75% Population)

2nd

floor fire 19th

floor fire 38th

floor fire 2nd

floor fire 19th

floor fire 38th

floor fire

374.24 374.19 373.41 365.31 371.29 372.29

368.62 369.37 373.11 371.86 364.79 368.14

374.29 367.4 371.39 366.47 366.17 368.85

371.9 378.32 375.11 366.01 368.78 365.33

372.24 373.46 371.75 366.91 365.05 362.42

373.57 370.77 372.24 366.39 365.41 365.23

372.85 370.9 376.57 369.31 365.34 363.38

375.04 374.28 371.39 366.83 366.99 363.63

371.03 371.39 375.62 366.45 364.10 365.43

372.67 372.23 370.67 370.40 367.70 371.09

373.71 372.65 376.57 367.25 369.05 368.52

370.36 372.98 368.72 368.18 366.86 367.61

372.88 370.75 375.17 367.10 365.18 363.92

372.74 375.48 373.18 370.29 366.37 370.35

372.76 371.92 370.58 366.83 367.17 364.99

372.79 376.13 373.01 366.39 365.28 367.41

372.11 372.79 372.81 369.44 369.59 368.78

371.82 372.26 370.56 365.03 366.99 364.34

371.08 371.58 376.14 365.22 369.45 367.52

375.48 375.42 374.99 362.63 364.57 364.39

368.95 370.02 373.91 366.12 367.97 364.08

373.75 375.1 376.27 368.08 368.19 365.72

369.64 373.19 374.08 366.36 367.20 368.53

372.81 372.78 372.31 369.30 370.70 366.44

372.32 374.68 372.26 365.09 365.52 370.40

370.92 375.87 372.79 367.32 367.06 365.96

372.87 373.41 374.45 372.44 364.39 367.91

372.17 373.65 372.93 367.89 368.77 369.58

377.12 374.45 376.18 368.16 372.06 369.15

371.56 376.31 375.45 368.90 369.13 368.47

375.69 373.25 378.1 367.19 369.41 373.72

377.34 372.14 371.68 371.12 366.73 364.89

372.6 369.02 371.9 368.55 367.23 370.39

290

375.32 371.65 371.48 366.54 366.05 367.97

373.17 370.19 372.93 369.56 368.68 364.25

372.06 375.81 373.75 368.53 365.83 367.42

372.15 370.58 371.41 367.52 364.69 364.35

377.32 371.46 374.22 367.32 366.27 367.08

371.44 371.97 373.88 364.95 366.03 364.79

371.41 370.01 375.91 368.04 368.93 369.43

376.85 378.49 374.88 368.11 369.01 370.43

371.68 370.42 371.55 365.89 367.42 367.27

373.21 374.03 370.46 368.44 368.20 365.87

374.27 372.88 371.92 368.80 367.56 366.18

371.81 372.86 369.54 368.56 364.10 366.57

373.52 372.85 371.39 370.16 366.93 363.73

376.24 371.68 374.7 368.47 367.42 367.17

372.45 374.76 369.88 367.91 366.22 366.41

373.65 372.39 373.13 366.04 368.74 364.77

374.3 372.31 371.54 367.40 363.40 369.63

373.79 379.44 375.17 367.41 368.03 369.43

375.9 373.15 370.29 369.39 370.25 367.29

369.93 369.18 374.83 366.87 366.10 365.80

373.36 374.83 377.51 363.41 368.63 373.26

372.12 374.58 370.89 373.80 370.11 370.60

375.08 371.69 372.18 369.06 368.33 370.39

375.39 374.92 374.87 368.38 368.09 367.36

370.62 369.59 372.02 367.49 367.14 365.19

371.99 374.55 376.25 366.76 365.87 366.00

369.96 376.19 373.9 367.56 364.74 368.81

370.58 372.09 371.92 365.23 366.22 367.88

375.13 369.39 368.72 368.39 365.88 367.71

371.29 369.89 367.12 369.42 366.98 366.04

375.87 370.05 374.14 368.15 365.85 366.22

373.77 376.02 371.9 371.71 366.90 365.78

369.79 372.59 375 365.48 365.06 364.92

372.59 373.23 375.79 364.84 366.75 369.43

370.04 374.98 371.63 367.68 365.73 366.90

372.39 372.91 373.45 366.15 364.41 364.90

371 370.12 368.2 366.84 364.95 367.81

373.79 372.33 371.82 369.40 364.96 363.41

377.54 369.11 366.3 365.42 365.57 367.31

372.79 372.08 378.38 369.27 366.27 366.11

373.63 371.52 373.62 369.45 367.35 367.83

375.88 371.54 373.34 370.87 368.04 366.80

371.33 376.03 368.16 368.94 365.57 366.05

371.29 370.31 370.96 369.15 367.56 366.56

367.77 374.34 374.07 365.43 364.68 369.97

376.83 376.43 371.89 368.57 365.64 370.82

372.54 371.96 373.1 368.59 369.77 370.11

374.36 373.59 372.92 367.61 366.67 364.74

375.55 369.57 369.75 367.50 368.10 365.58

291

373.43 376.45 374.76 370.26 366.63 368.70

372.65 370.39 373.31 369.81 366.49 363.45

375.7 369.8 372.78 365.32 367.88 362.99

373.37 371.13 374.72 368.75 368.19 365.80

371.48 375.42 372.78 365.73 368.38 373.34

371.02 368.64 370.52 367.52 364.92 368.54

376.64 375.77 374.81 365.72 368.26 366.40

371.31 375.35 373.58 366.73 367.23 366.47

374 368.81 370.5 367.95 366.83 369.19

373.76 372.11 369.87 366.74 369.14 365.43

377.9 371.11 374.06 367.67 368.16 370.41

375.41 374.95 372.54 370.21 365.49 369.01

370.33 371.47 372.15 366.66 365.77 367.37

377.14 371.43 370.86 369.49 367.34 367.98

372.34 371.67 368.88 367.10 363.99 363.73

372.12 372.92 373.33 367.71 367.04 364.04

368.67 370.47 372.04 365.34 366.48 364.34

377.86 370.06 372.79 366.19 367.84 368.10

Table G5: Stair Evacuation Time, tSE (100 replications)

Stair Evacuation Time (100% Population) Stair Evacuation Time (75% Population)

2nd

floor fire 19th

floor fire 38th

floor fire 2nd

floor fire 19th

floor fire 38th

floor fire

1812.73 1731.27 1799.73 1476.53 1219.15 1540.93

1717.4 1936.55 1920.57 1204.92 1246.38 1623.37

1681.98 1321.1 1796.19 1187.36 1223.95 1520.77

1488.45 926.1 909.4 1204.72 1186.40 1446.89

1165.35 1102.33 822.04 1174.28 1225.69 1234.67

1711.33 1947.65 1676.86 1273.49 1210.89 1477.21

1827.44 1761.46 1826.44 1226.55 1265.24 1462.48

1829.56 1740.89 1605.64 1217.78 1232.12 1461.53

1105.69 1132.48 1314.98 820.43 705.61 779.43

1433.73 1287.19 1317.24 906.18 988.71 1146.36

1078.54 1587.65 1609.51 1036.58 1002.47 1100.91

1779.85 1734.67 1497.09 1219.89 1257.19 1359.00

1797.3 1730.81 2002.17 1244.41 1332.18 1431.55

1431.76 1743.24 1782.24 1242.17 1336.34 1618.81

1644.63 1766.27 1711.58 1199.76 1318.45 1439.13

1830.24 1857.11 1807.4 1212.00 1187.24 1470.11

1649.33 1618.56 1610.44 1168.81 1263.72 1428.99

1813.49 2021.96 1773.55 1238.98 1234.27 1622.78

1807.62 1609.3 1770.46 1353.38 1215.94 1438.25

1657.57 1695.44 1709.49 1158.86 1207.65 1572.65

1842.99 1595.22 1790.21 1237.21 1384.31 1278.49

1526.73 1641.66 1869.02 1261.29 1401.78 1275.23

1596.84 1931.41 1834.06 1196.19 1213.72 1582.10

1569.91 1632.72 1694.72 1296.11 1202.82 1588.84

1270.52 937.93 1196.22 854.19 833.11 1048.02

1337.3 1283.97 1550.94 787.94 731.12 858.91

292

1786.05 1607.7 1669.21 1217.97 1418.73 1408.21

1756.07 1871.31 1718.2 1232.38 1368.74 1310.75

1846.28 1823.98 1847.1 1402.89 1316.26 1267.93

946.26 1178.67 1227.99 805.02 847.82 755.34

1804.8 1784.13 1782.68 1249.71 1401.90 1586.49

1540.61 1314.18 1438.08 810.81 1276.07 936.90

1502.74 1549.45 1695.18 1080.03 1152.09 1322.43

1679.7 1708.54 1456.99 1177.35 1228.32 1286.19

1581.98 928.38 1099.34 1217.03 1215.32 1473.08

1220.8 1616.96 1585.92 1066.02 1065.56 982.34

800.35 1186.54 1417.15 901.95 952.96 921.10

1440.84 1446.33 1380.63 651.48 763.87 937.98

1532.72 1757.8 1784.9 1170.27 1177.89 1217.45

1528.53 1233.08 1407.8 878.89 916.11 1008.60

1779.33 1726.49 1847.52 1282.20 1258.69 1701.60

1579.68 1306.61 1453.33 951.37 951.34 1329.33

1687.17 1737.22 1799.89 1195.00 1333.04 1230.14

1772.34 1649.96 1678.97 1178.30 1411.98 1271.66

1471.23 1671.28 1576.19 1032.91 1292.43 1262.42

1132.14 1386.38 1643.8 1026.24 1056.20 1068.65

1265.19 1023.95 832.81 723.05 770.22 1010.43

1783.24 1753.82 1760.4 1223.47 1234.07 1551.95

1900.23 1813.73 1743.49 1238.33 1327.16 1579.26

1743.68 1789.07 1777.84 1217.58 1257.39 1671.25

1738.2 1920.56 1775.94 1222.56 1228.13 1494.19

921.73 1434.89 1369.77 838.07 983.13 897.27

1726.35 1590.21 1761.8 1146.66 1188.74 1285.84

1103.01 1101.83 1243.4 867.18 865.40 1018.33

1551.73 1795.99 1682.55 1181.77 1247.20 1533.23

1307.59 1364.29 1358.53 737.47 932.28 856.36

1061.4 1067.93 1204.06 1292.50 1252.30 1432.23

1158.44 1527.86 1374.3 1022.88 1044.51 1190.33

1299.32 1642.9 1544.1 992.79 992.75 988.84

884.51 1066.58 1030.77 1154.74 1323.33 1444.44

1716.23 1755.84 1876.95 1203.21 1224.78 1581.35

1854.11 1714.04 1657.84 1216.90 1245.85 1498.74

1947.13 1687.36 1757.79 1260.86 1424.52 1599.75

1950.26 1895.88 1716.87 1187.65 1329.41 1505.14

1643.46 1838.51 1514.96 1187.85 1462.94 1277.18

904.69 1209.38 870.81 884.45 898.58 906.43

1453.16 1170.31 1342.15 950.30 968.35 1275.96

1466.16 1763.61 1640.3 1115.24 1131.81 1342.11

1283.14 951.58 1422.12 945.66 956.28 1295.34

1392.56 1394.79 1527.45 1051.17 1054.40 1358.31

911.44 1562.76 1775.24 1210.55 1252.58 1264.59

1649.44 917.38 964.42 816.24 836.97 925.60

1828.68 2157.91 1673.66 1176.94 1289.24 1291.78

1728.43 933.48 1813.49 1222.37 1389.64 1472.43

1288.82 1099.92 953.29 729.12 749.48 878.47

293

1831.43 1565.72 1540.54 1162.66 1182.41 1454.51

852.84 1229.04 1425.11 815.73 854.38 1010.00

1282.06 995.86 846.67 1193.62 1211.64 1487.70

1846.93 1850.2 1872.71 1254.07 1265.70 1640.43

2048.76 1844.99 1813.83 1249.79 1291.67 1502.21

1549.31 1167.29 890.19 875.89 681.79 1010.70

1689.54 1471.84 1606.69 1217.04 1342.32 1400.00

1457.25 1741.25 1740.36 1195.50 1241.08 1494.85

1334.15 989.57 1144.95 619.42 731.00 672.45

1664.12 1720.43 1799.59 1192.71 1198.43 1487.35

1833.19 2002.72 1778.73 1157.91 1294.38 1414.68

1737.25 1846.32 1559.62 1207.21 1163.85 1375.09

1505.17 1823.33 1820.77 1225.37 1306.26 1638.20

2114.19 1605.02 1605.39 1244.24 1220.67 1389.25

1777.25 1601.88 1785.26 1178.83 1370.89 1391.34

1792.84 1695.71 1673.96 1185.76 1301.11 1535.04

1774.57 1372.89 1524.51 1249.50 1260.91 1360.70

1783.22 1754.42 1775.94 1228.15 1245.58 1639.45

1949.24 1832.03 1622 1214.91 1344.33 1387.66

1692.48 1709.71 1736.39 1127.12 1155.42 1371.82

1502.1 1107.54 1401.84 709.69 989.69 940.32

1812.81 1863.41 1656.92 1222.18 1231.59 1489.96

1730.7 1798.44 1954.55 1192.57 1212.97 1527.18

1058.44 1388 1109.07 1202.90 1201.40 874.57

1723.27 1728.74 1599.02 1182.78 1196.98 1484.52

Table G6 – Mean Lift Time Periods (average values of three levels)

Lift Waiting

Time (100%)

Lift

Transportation

Time (100%)

Lift

Evacuation

Time (100%)

Lift

Waiting

Time (25%)

Lift

Transportation

Time (25%)

Lift

Evacuation

Time (25%)

701.60 41.88 2325.70 318.42 39.86 1363.20

722.30 42.15 2532.93 267.69 40.98 1494.77

659.42 41.00 1402.33 290.68 38.88 1547.64

780.79 43.04 2334.95 295.30 38.66 1266.49

713.43 42.10 2526.68 279.89 40.98 1345.47

653.20 41.18 1520.53 251.41 40.15 941.27

726.51 41.41 1801.56 310.67 40.12 1163.32

714.04 42.03 2724.41 279.62 40.11 1798.12

693.01 42.47 2221.10 257.55 41.00 1607.98

707.14 41.22 2035.70 298.60 40.76 1189.36

743.19 42.11 1630.29 289.03 39.07 984.94

721.16 40.70 2516.06 327.54 39.84 1765.61

688.41 43.32 1891.82 274.26 40.61 1308.85

704.11 42.53 1864.95 305.50 41.23 1222.22

695.08 41.46 2156.45 284.01 39.09 1244.51

685.06 41.96 2665.93 292.23 40.60 1980.32

761.83 42.17 3074.17 285.29 40.14 1951.77

766.16 42.00 2888.49 335.33 40.21 1975.80

294

681.50 42.25 2530.76 290.06 40.88 2010.19

668.52 40.82 1473.71 294.60 39.48 1875.93

743.01 42.90 2199.19 297.61 41.40 1174.13

696.27 42.41 1528.35 329.14 39.77 1415.60

736.92 42.34 2383.29 323.25 41.62 1518.45

729.07 42.45 2094.77 340.61 39.99 1131.06

713.62 40.43 2456.22 289.81 37.68 1685.48

752.75 42.05 2754.29 310.76 39.47 1933.16

762.37 42.74 2167.41 313.37 39.62 1332.68

756.82 41.58 1510.05 285.04 41.10 1315.60

730.92 41.68 1955.86 338.83 39.95 1510.70

737.62 41.65 2927.54 341.22 39.98 1803.38

701.12 42.02 3040.13 292.96 41.30 2044.87

752.74 43.86 2217.15 329.41 41.07 1595.71

745.41 42.06 1807.73 319.97 40.28 1054.95

727.93 42.56 2147.05 317.42 41.05 1059.39

748.04 41.23 2289.94 331.35 39.38 1143.43

758.54 41.65 1944.68 276.90 39.54 1097.87

678.42 41.19 3009.86 340.45 39.29 1941.43

708.26 41.30 2218.18 283.89 39.06 1426.94

709.65 42.10 3074.33 296.28 41.06 1780.53

727.87 42.89 2717.54 297.74 42.38 1874.06

697.66 41.81 2916.63 275.70 40.40 1768.00

749.15 40.95 2766.99 290.24 40.56 1750.36

775.94 42.30 1599.23 306.25 40.21 1555.94

742.13 42.44 3152.85 322.13 39.76 1898.32

776.35 41.70 1961.49 336.42 40.83 1803.00

760.70 41.90 1783.94 269.67 39.12 1473.67

721.55 41.03 2669.68 329.33 39.17 1803.28

734.01 42.43 2901.99 279.58 41.92 1947.44

738.20 42.74 2922.17 300.80 39.69 1786.26

697.21 40.20 2000.15 282.01 40.09 1294.42

752.38 43.83 1878.07 310.35 38.77 1658.17

743.27 42.50 1833.70 289.80 40.48 1370.15

787.06 41.66 1737.97 311.84 40.99 1873.63

717.67 41.18 1982.72 283.87 40.62 1454.89

735.80 42.70 2780.06 299.98 42.08 1897.21

708.87 40.82 2939.69 257.51 40.49 1736.57

759.67 41.13 2648.25 319.47 40.13 1793.00

710.39 41.67 1890.27 269.02 38.25 1405.50

707.13 42.39 2860.71 315.67 40.46 1945.89

716.77 41.62 2854.75 287.36 39.39 1923.49

709.72 41.48 2303.76 245.53 40.89 1267.60

698.88 41.48 1799.53 290.74 41.51 1643.17

722.12 42.18 2047.23 284.21 39.76 1510.01

710.80 42.21 1555.45 289.45 40.61 1715.78

687.23 41.19 2640.34 294.42 39.96 1649.04

715.01 41.73 2850.18 276.23 40.20 1640.01

702.32 42.82 2881.55 308.24 41.67 1967.10

295

722.40 41.80 1924.52 341.87 40.59 1244.16

683.44 41.76 1487.32 267.57 39.29 1804.61

705.49 42.10 1278.50 308.05 42.33 2068.78

729.82 41.06 2833.33 307.49 39.63 1701.20

722.63 42.85 2256.62 293.67 40.68 1326.15

722.65 42.56 2779.73 281.65 40.97 1932.68

716.85 41.07 1503.60 296.84 39.75 1971.54

734.07 41.54 1573.02 313.24 40.63 1984.85

734.18 42.01 3018.17 287.46 39.55 1738.46

698.78 41.83 1823.82 357.53 42.47 1157.86

715.62 42.03 1989.34 324.44 39.36 1353.20

733.55 41.90 2240.95 305.51 42.39 1579.54

715.45 41.64 2734.02 314.02 42.42 1996.73

673.39 41.33 2614.43 279.56 39.89 1740.10

723.15 42.14 2592.85 323.47 39.47 2102.56

669.51 41.90 2156.02 325.30 40.67 913.41

703.10 41.47 2851.05 316.66 40.04 2081.34

751.96 41.59 1736.08 279.68 39.49 1450.05

688.34 41.09 2255.81 259.81 38.51 1349.11

743.01 42.46 2197.83 325.52 40.94 1550.96

734.15 41.43 2705.38 269.34 40.28 1584.15

695.13 41.64 2499.77 272.76 39.73 1130.84

687.31 40.42 1671.84 249.58 39.54 966.85

768.45 43.18 2671.41 324.41 40.34 1933.49

744.46 42.69 2066.23 314.85 40.48 1239.54

696.62 40.80 2699.68 273.37 40.41 1614.03

721.00 42.55 1483.67 263.81 39.21 1954.38

764.20 43.50 2548.98 290.63 37.92 1630.06

705.94 41.65 2941.08 276.23 39.47 2000.18

746.52 40.74 2098.77 299.51 40.34 1210.86

643.09 40.51 1842.26 285.12 39.34 1550.61

705.71 40.29 2959.98 276.21 41.04 2028.46

713.01 42.17 1943.92 251.78 42.45 1390.16

296

Table G7 – Mean Stair Time Periods (average values of three levels)

Stair Travelling

Time (100%

population)

Stair Evacuation

Time (100%

population)

Stair Travelling

Time (75%

population)

Stair Evacuation

Time (75%

population)

373.95 1781.24 369.63 1412.20

370.37 1858.17 368.26 1358.22

371.03 1599.76 367.16 1310.69

375.11 1107.98 366.70 1279.34

372.48 1029.91 364.79 1211.54

372.19 1778.61 365.68 1320.53

373.44 1805.12 366.01 1318.09

373.57 1725.36 365.82 1303.81

372.68 1184.38 365.33 768.49

371.86 1346.05 369.73 1013.75

374.31 1425.24 368.27 1046.65

370.69 1670.54 367.55 1278.69

372.93 1843.42 365.40 1336.04

373.8 1652.41 369.00 1399.11

371.75 1707.49 366.33 1319.11

373.98 1831.59 366.36 1289.79

372.57 1626.11 369.27 1287.17

371.55 1869.67 365.45 1365.34

372.93 1729.13 367.40 1335.86

375.3 1687.5 363.86 1313.05

370.96 1742.8 366.06 1300.00

375.04 1679.14 367.33 1312.76

372.3 1787.43 367.36 1330.67

372.63 1632.45 368.81 1362.59

373.08 1134.89 367.00 911.77

373.2 1390.74 366.78 792.66

373.58 1687.65 368.25 1348.30

372.91 1781.86 368.74 1303.96

375.92 1839.12 369.79 1329.03

374.44 1117.64 368.83 802.73

375.68 1790.54 370.11 1412.70

373.72 1430.96 367.58 1007.93

371.17 1582.46 368.72 1184.85

372.81 1615.07 366.85 1230.62

372.1 1203.23 367.50 1301.81

373.87 1474.56 367.26 1037.97

371.38 1134.68 365.52 925.34

374.34 1422.6 366.89 784.45

372.43 1691.81 365.25 1188.54

372.45 1389.8 368.80 934.53

376.74 1784.45 369.18 1414.16

371.22 1446.54 366.86 1077.35

372.57 1741.43 367.50 1252.73

297

373.02 1700.43 367.52 1287.31

371.4 1572.9 366.41 1195.92

372.59 1387.44 366.94 1050.36

374.21 1040.65 367.69 834.57

372.36 1765.82 366.85 1336.50

373.06 1819.15 366.52 1381.58

372.72 1770.2 366.81 1382.07

376.13 1811.57 368.29 1314.96

373.11 1242.13 368.98 906.16

371.31 1692.79 366.26 1207.08

375.23 1149.41 368.43 916.97

372.53 1676.75 371.50 1320.73

372.98 1343.47 369.26 842.04

375.06 1111.13 367.94 1325.67

370.74 1353.53 366.61 1085.91

374.27 1495.44 366.21 991.46

373.35 993.95 367.04 1307.50

371.53 1783.01 366.44 1336.45

371.08 1742 367.33 1320.50

369.43 1797.43 367.48 1428.38

373.35 1854.34 366.74 1340.73

373.9 1665.64 368.13 1309.32

372.46 994.96 365.15 896.49

373.87 1321.87 367.01 1064.87

372.22 1623.35 366.77 1196.39

372.92 1218.95 365.15 1065.76

369.77 1438.27 366.54 1154.63

372.65 1416.48 365.92 1242.57

370.99 1177.08 366.10 859.60

374.42 1886.75 367.22 1252.65

372.92 1491.8 368.21 1361.48

373.58 1114.01 368.57 785.69

371.84 1645.9 366.85 1266.53

370.86 1169 367.75 893.37

372.06 1041.53 366.69 1297.65

375.05 1856.62 368.34 1386.73

372.54 1902.53 369.49 1347.89

373.63 1202.26 366.34 856.13

371.62 1589.35 367.06 1319.79

374.88 1646.29 368.53 1310.48

372.12 1156.22 366.59 674.29

372.76 1728.05 365.40 1292.83

373.07 1871.55 367.58 1288.99

373.23 1714.4 369.15 1248.72

370.06 1716.43 366.99 1389.95

375.74 1774.87 366.79 1284.72

373.41 1721.46 366.81 1313.69

371.1 1720.84 367.99 1340.64

371.91 1557.32 367.10 1290.37

298

374.36 1771.19 368.75 1371.06

374.3 1801.09 368.24 1315.63

371.31 1712.86 366.60 1218.12

373.14 1337.16 368.27 879.90

370.96 1777.71 364.94 1314.58

372.79 1827.9 366.26 1310.91

370.39 1185.17 365.38 1092.96

373.57 1683.68 367.38 1288.10

Table G8: Floor-Wise Lift Waiting Time (fire on 2nd

floor)

Waiting Time Average Half

Width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 26.16 2.15 10.63 69.01 0.01 238.00

Floor 02.Queue 486.32 63.37 55.74 1655.52 0.08 1868.33

Floor 03.Queue 627.53 74.25 43.60 1738.81 0.21 1982.33

Floor 04.Queue 641.85 75.50 74.58 1645.42 0.05 2046.17

Floor 05.Queue 656.92 73.66 62.88 1607.76 0.43 1964.06

Floor 06.Queue 710.06 79.71 88.35 1767.21 0.31 1922.58

Floor 07.Queue 615.65 65.51 96.61 1524.02 1.32 1631.03

Floor 08.Queue 636.33 78.24 68.31 1728.73 2.37 1988.79

Floor 09.Queue 711.80 62.15 106.15 1309.53 0.02 1928.64

Floor 10.Queue 750.00 72.97 110.13 1494.49 1.34 1895.74

Floor 11.Queue 726.09 72.96 120.95 1695.20 3.13 1908.29

Floor 12.Queue 695.72 69.37 70.16 1575.41 2.56 1900.89

Floor 13.Queue 700.95 78.47 82.85 1781.33 0.36 1937.83

Floor 14.Queue 728.34 78.07 64.73 1725.82 1.48 1901.43

Floor 15.Queue 743.67 74.18 120.17 1755.58 0.15 1840.10

Floor 16.Queue 762.71 77.52 66.63 1610.28 0.14 1817.28

Floor 17.Queue 739.79 72.67 190.18 1805.31 0.89 2015.29

Floor 18.Queue 765.46 79.78 98.85 1771.46 0.68 1891.45

Floor 19.Queue 895.56 81.42 75.46 1916.37 1.36 2089.71

Floor 20.Queue 823.65 74.56 125.69 1800.62 0.43 1945.83

Floor 21.Queue 803.23 78.91 86.21 2225.58 0.44 2298.87

Floor 22.Queue 684.25 75.60 92.28 1709.34 1.59 2000.60

Floor 23.Queue 812.56 87.66 102.08 1930.13 1.12 1998.80

Floor 24.Queue 811.48 83.38 53.13 1751.92 0.92 1914.64

Floor 25.Queue 754.58 77.20 109.11 1615.62 0.07 1892.65

Floor 26.Queue 723.81 80.21 62.70 1561.69 0.58 1980.61

Floor 27.Queue 740.11 78.04 79.74 1967.46 0.06 2064.60

Floor 28.Queue 623.84 77.22 49.74 1762.76 0.09 1899.19

Floor 29.Queue 588.45 74.40 28.44 1700.80 0.01 2039.30

Floor 30.Queue 590.60 83.71 42.79 1628.52 0.09 1895.64

Floor 31.Queue 598.33 81.84 32.10 1898.37 0.01 2011.71

Floor 32.Queue 645.16 77.50 29.27 1644.15 0.00 1838.97

Floor 33.Queue 646.00 72.76 36.30 1637.27 0.01 2074.18

Floor 34.Queue 737.30 81.05 38.78 1572.82 0.03 1978.26

Floor 35.Queue 904.95 80.96 136.98 1751.41 0.02 2206.09

Floor 36.Queue 887.52 77.12 97.37 1689.04 0.11 2195.09

Floor 37.Queue 983.78 86.23 115.76 2016.86 0.00 2289.98

Floor 38.Queue 1004.49 97.85 135.84 2319.36 0.55 2371.36

299

Table G9: Floor-Wise Number of Evacuees in Queue (fire on 2nd floor)

Number Waiting Average Half

Width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 0.21 0.02 0.08 0.56 0.00 11.00

Floor 02.Queue 5.61 0.76 0.52 20.39 0.00 32.00

Floor 03.Queue 7.21 0.88 0.46 21.17 0.00 32.00

Floor 04.Queue 7.36 0.88 0.95 18.04 0.00 32.00

Floor 05.Queue 7.62 0.90 0.66 20.50 0.00 32.00

Floor 06.Queue 8.12 0.93 1.08 21.69 0.00 32.00

Floor 07.Queue 7.08 0.78 1.09 18.51 0.00 33.00

Floor 08.Queue 7.28 0.91 0.88 21.21 0.00 33.00

Floor 09.Queue 8.14 0.72 1.27 15.89 0.00 33.00

Floor 10.Queue 8.65 0.88 1.24 19.53 0.00 32.00

Floor 11.Queue 8.39 0.87 1.20 21.00 0.00 33.00

Floor 12.Queue 8.04 0.83 0.74 19.16 0.00 33.00

Floor 13.Queue 8.09 0.92 0.95 21.12 0.00 32.00

Floor 14.Queue 8.39 0.94 0.78 22.85 0.00 33.00

Floor 15.Queue 8.63 0.91 1.37 20.59 0.00 33.00

Floor 16.Queue 8.80 0.91 0.70 20.08 0.00 33.00

Floor 17.Queue 8.53 0.86 1.98 20.48 0.00 33.00

Floor 18.Queue 8.83 0.94 1.07 21.13 0.00 33.00

Floor 19.Queue 10.32 0.95 0.82 21.49 0.00 33.00

Floor 20.Queue 9.51 0.89 1.53 22.34 0.00 34.00

Floor 21.Queue 9.27 0.92 1.06 24.73 0.00 32.00

Floor 22.Queue 7.84 0.85 1.09 18.14 0.00 32.00

Floor 23.Queue 9.39 1.04 1.08 21.56 0.00 33.00

Floor 24.Queue 9.34 0.96 0.64 19.89 0.00 33.00

Floor 25.Queue 8.69 0.89 1.19 20.03 0.00 32.00

Floor 26.Queue 8.35 0.93 0.69 19.22 0.00 33.00

Floor 27.Queue 8.48 0.89 0.91 21.65 0.00 34.00

Floor 28.Queue 7.18 0.90 0.59 21.27 0.00 34.00

Floor 29.Queue 6.78 0.86 0.32 20.52 0.00 32.00

Floor 30.Queue 6.75 0.94 0.54 18.57 0.00 33.00

Floor 31.Queue 6.84 0.90 0.39 19.17 0.00 32.00

Floor 32.Queue 7.43 0.89 0.35 18.39 0.00 33.00

Floor 33.Queue 7.41 0.82 0.39 17.70 0.00 33.00

Floor 34.Queue 8.60 0.96 0.41 19.11 0.00 33.00

Floor 35.Queue 10.50 0.95 1.68 23.56 0.00 34.00

Floor 36.Queue 10.27 0.88 1.14 20.71 0.00 34.00

Floor 37.Queue 11.30 0.97 1.41 22.97 0.00 32.00

Floor 38.Queue 11.66 1.17 1.52 23.90 0.00 33.00

300

Table G10: Floor-Wise Lift Waiting Time (fire on 19th floor)

Waiting Time Average Half

Width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 26.97 2.27 13.53 61.75 0.02 224.60

Floor 02.Queue 683.20 78.09 100.65 1759.93 1.61 2062.88

Floor 03.Queue 653.12 64.28 85.38 1404.80 0.21 1997.93

Floor 04.Queue 658.34 76.64 59.74 1625.67 0.36 1833.95

Floor 05.Queue 630.88 72.55 46.98 1808.77 0.50 2086.69

Floor 06.Queue 691.11 68.74 85.97 1666.60 0.34 1814.71

Floor 07.Queue 665.67 74.26 45.22 1683.63 0.45 1999.93

Floor 08.Queue 685.93 78.24 144.63 2096.00 2.29 2260.06

Floor 09.Queue 686.45 73.64 88.20 1689.76 1.04 2102.50

Floor 10.Queue 755.85 82.21 87.22 1783.25 0.64 2024.93

Floor 11.Queue 758.26 71.70 118.78 1723.59 0.87 2106.54

Floor 12.Queue 751.37 74.17 127.66 1547.48 0.65 2029.59

Floor 13.Queue 761.62 76.78 81.83 1901.58 0.92 2015.35

Floor 14.Queue 773.69 76.72 71.81 1831.71 0.89 2140.82

Floor 15.Queue 778.43 73.05 73.10 1750.03 0.15 2161.87

Floor 16.Queue 751.06 82.09 145.23 1950.21 0.54 2076.69

Floor 17.Queue 728.02 68.85 163.15 1699.30 2.58 1849.20

Floor 18.Queue 658.77 75.52 60.62 1705.65 1.65 1928.71

Floor 19.Queue 661.78 69.76 126.25 1508.59 0.17 1737.91

Floor 20.Queue 783.63 76.03 129.80 1733.56 0.54 2029.08

Floor 21.Queue 798.39 72.88 94.77 1583.46 1.43 1944.72

Floor 22.Queue 804.79 74.40 65.24 1837.08 0.34 1897.02

Floor 23.Queue 819.30 73.68 198.28 1788.42 2.58 1867.20

Floor 24.Queue 826.84 78.66 113.16 1859.88 1.36 2054.16

Floor 25.Queue 740.62 76.84 103.82 1633.84 0.24 1991.94

Floor 26.Queue 711.78 76.09 63.38 1588.32 0.18 1863.37

Floor 27.Queue 689.25 80.11 62.22 1692.42 0.48 1924.10

Floor 28.Queue 638.02 69.10 70.06 1481.95 0.05 1838.55

Floor 29.Queue 520.74 77.89 31.35 1494.43 0.03 1960.52

Floor 30.Queue 543.67 81.64 27.54 1832.21 0.02 1981.69

Floor 31.Queue 527.83 73.38 45.56 1662.53 0.03 1914.93

Floor 32.Queue 595.23 77.08 31.96 1783.42 0.06 2250.99

Floor 33.Queue 704.95 76.90 44.73 1650.98 0.16 2202.82

Floor 34.Queue 771.98 75.19 50.04 1597.23 0.03 2077.66

Floor 35.Queue 806.33 75.79 103.78 1601.22 0.04 2084.80

Floor 36.Queue 897.59 82.28 156.75 1946.84 0.11 2314.22

Floor 37.Queue 906.66 90.95 120.38 2063.47 0.15 2323.26

Floor 38.Queue 1077.01 89.29 151.29 2178.60 16.53 2402.26

301

Table G11: Floor-Wise Number of Evacuees in Queue (fire on 19th floor)

Number Waiting Average Half

Width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 0.21 0.02 0.09 0.52 0.00 10.00

Floor 02.Queue 7.80 0.90 1.18 19.76 0.00 32.00

Floor 03.Queue 7.51 0.76 1.06 16.79 0.00 32.00

Floor 04.Queue 7.60 0.92 0.67 19.91 0.00 32.00

Floor 05.Queue 7.16 0.82 0.57 20.20 0.00 32.00

Floor 06.Queue 7.88 0.78 1.01 19.35 0.00 33.00

Floor 07.Queue 7.70 0.90 0.51 20.84 0.00 32.00

Floor 08.Queue 7.90 0.92 1.54 23.05 0.00 32.00

Floor 09.Queue 7.88 0.86 0.93 20.02 0.00 32.00

Floor 10.Queue 8.77 1.01 0.98 22.19 0.00 32.00

Floor 11.Queue 8.71 0.84 1.31 21.48 0.00 32.00

Floor 12.Queue 8.58 0.84 1.47 18.76 0.00 32.00

Floor 13.Queue 8.69 0.87 1.00 19.74 0.00 32.00

Floor 14.Queue 8.87 0.90 0.90 21.29 0.00 33.00

Floor 15.Queue 8.94 0.86 0.90 19.88 0.00 32.00

Floor 16.Queue 8.66 0.97 1.59 23.20 0.00 33.00

Floor 17.Queue 8.36 0.81 2.14 21.06 0.00 32.00

Floor 18.Queue 7.59 0.90 0.75 20.13 0.00 32.00

Floor 19.Queue 7.70 0.86 1.26 19.44 0.00 32.00

Floor 20.Queue 9.03 0.89 1.51 20.64 0.00 33.00

Floor 21.Queue 9.20 0.87 1.02 19.05 0.00 32.00

Floor 22.Queue 9.25 0.88 0.76 22.13 0.00 33.00

Floor 23.Queue 9.47 0.88 2.21 20.98 0.00 34.00

Floor 24.Queue 9.62 0.95 1.24 19.90 0.00 34.00

Floor 25.Queue 8.58 0.92 1.40 19.36 0.00 33.00

Floor 26.Queue 8.25 0.89 0.63 19.16 0.00 32.00

Floor 27.Queue 7.93 0.91 0.66 19.42 0.00 33.00

Floor 28.Queue 7.29 0.77 0.79 15.00 0.00 33.00

Floor 29.Queue 5.93 0.88 0.33 18.44 0.00 33.00

Floor 30.Queue 6.20 0.89 0.27 17.92 0.00 33.00

Floor 31.Queue 6.08 0.83 0.47 16.91 0.00 33.00

Floor 32.Queue 6.84 0.86 0.33 19.08 0.00 32.00

Floor 33.Queue 8.09 0.87 0.57 19.06 0.00 33.00

Floor 34.Queue 8.84 0.84 0.63 18.58 0.00 34.00

Floor 35.Queue 9.29 0.88 1.30 18.39 0.00 33.00

Floor 36.Queue 10.21 0.88 1.87 20.44 0.00 33.00

Floor 37.Queue 10.46 1.04 1.52 23.01 0.00 33.00

Floor 38.Queue 12.48 1.06 1.74 25.38 0.00 34.00

302

Table G12: Floor-Wise Lift Waiting Time (fire on 38th floor)

Waiting Time Average Half

Width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 26.71 2.09 12.71 68.84 0.00 248.24

Floor 02.Queue 681.07 73.55 74.60 1751.90 0.80 1928.53

Floor 03.Queue 644.37 80.38 43.41 1927.50 0.40 2107.77

Floor 04.Queue 728.19 79.73 60.16 1727.69 0.83 2161.11

Floor 05.Queue 661.27 78.49 103.28 1558.48 0.07 1916.80

Floor 06.Queue 670.39 76.82 68.28 1948.20 0.79 2106.31

Floor 07.Queue 665.40 77.39 51.19 1717.77 0.54 1868.08

Floor 08.Queue 649.69 66.61 38.90 1808.16 0.56 1951.74

Floor 09.Queue 740.58 84.09 62.00 1789.25 0.51 1873.45

Floor 10.Queue 711.12 83.35 64.61 2006.97 0.12 2081.17

Floor 11.Queue 702.82 68.95 76.10 1584.51 1.82 1710.08

Floor 12.Queue 663.99 70.31 59.52 1819.13 3.40 1877.61

Floor 13.Queue 773.38 71.08 150.82 1703.08 0.78 1910.49

Floor 14.Queue 755.84 68.91 188.86 1704.80 3.11 1830.96

Floor 15.Queue 752.94 82.52 89.94 1966.89 0.08 2100.41

Floor 16.Queue 771.74 81.66 79.17 1790.04 0.23 1872.82

Floor 17.Queue 753.66 77.84 54.90 1730.19 1.79 1836.85

Floor 18.Queue 811.62 73.70 188.93 1850.73 3.43 1940.76

Floor 19.Queue 819.76 72.19 139.51 1971.81 3.02 2032.46

Floor 20.Queue 802.01 83.18 87.89 1922.62 1.55 1979.82

Floor 21.Queue 848.50 81.06 122.31 1988.96 0.02 2118.84

Floor 22.Queue 781.98 76.36 53.34 1850.90 0.53 2001.95

Floor 23.Queue 827.28 79.92 89.37 1872.78 1.62 2092.31

Floor 24.Queue 761.27 76.29 102.35 1647.48 2.21 2033.31

Floor 25.Queue 880.49 88.38 72.03 2051.23 0.19 2133.40

Floor 26.Queue 716.66 73.84 41.26 1528.29 0.03 1749.96

Floor 27.Queue 675.75 74.15 25.08 1579.95 0.12 1972.29

Floor 28.Queue 665.30 78.92 93.09 2121.02 0.08 2190.04

Floor 29.Queue 506.20 63.68 55.68 1688.51 0.14 2031.17

Floor 30.Queue 508.72 71.99 35.71 1375.16 0.08 1752.51

Floor 31.Queue 593.22 75.25 44.83 1486.70 0.02 1768.00

Floor 32.Queue 593.40 78.92 40.18 1657.46 0.15 2037.64

Floor 33.Queue 660.78 77.75 60.75 1706.38 0.01 1885.96

Floor 34.Queue 732.33 79.32 82.06 1757.74 0.01 1980.46

Floor 35.Queue 827.24 75.98 90.75 1794.73 0.04 2051.41

Floor 36.Queue 891.06 73.32 115.46 1688.14 0.11 2059.78

Floor 37.Queue 943.78 90.36 79.92 2001.48 0.00 2406.48

Floor 38.Queue 820.23 73.74 164.80 1613.44 0.79 2152.75

303

Table G13: Floor-Wise Number of Evacuees in Queue (fire on 38th floor)

Number Waiting Average Half

width

Minimum

Average

Maximum

Average

Minimum

Value

Maximum

Value

Floor 01.Queue 0.21 0.02 0.09 0.56 0.00 11.00

Floor 02.Queue 7.85 0.87 0.90 19.62 0.00 32.00

Floor 03.Queue 7.38 0.92 0.53 21.12 0.00 32.00

Floor 04.Queue 8.39 0.94 0.62 21.04 0.00 32.00

Floor 05.Queue 7.66 0.94 1.11 19.59 0.00 32.00

Floor 06.Queue 7.70 0.87 0.72 22.05 0.00 32.00

Floor 07.Queue 7.67 0.92 0.61 20.06 0.00 32.00

Floor 08.Queue 7.44 0.77 0.43 21.94 0.00 33.00

Floor 09.Queue 8.54 0.99 0.77 21.50 0.00 32.00

Floor 10.Queue 8.15 0.95 0.77 21.61 0.00 33.00

Floor 11.Queue 8.04 0.79 0.72 17.57 0.00 33.00

Floor 12.Queue 7.69 0.84 0.71 20.43 0.00 32.00

Floor 13.Queue 8.96 0.86 1.57 19.61 0.00 33.00

Floor 14.Queue 8.66 0.79 2.49 20.10 0.00 32.00

Floor 15.Queue 8.67 0.96 1.15 23.09 0.00 33.00

Floor 16.Queue 8.91 0.97 0.89 22.66 0.00 33.00

Floor 17.Queue 8.68 0.91 0.61 19.73 0.00 32.00

Floor 18.Queue 9.34 0.85 2.08 21.36 0.00 32.00

Floor 19.Queue 9.45 0.86 1.89 23.15 0.00 32.00

Floor 20.Queue 9.26 0.98 1.04 23.67 0.00 33.00

Floor 21.Queue 9.79 0.96 1.33 24.25 0.00 33.00

Floor 22.Queue 9.04 0.92 0.59 20.61 0.00 33.00

Floor 23.Queue 9.57 0.94 1.17 20.85 0.00 33.00

Floor 24.Queue 8.75 0.87 1.30 19.67 0.00 33.00

Floor 25.Queue 10.19 1.05 0.87 23.13 0.00 32.00

Floor 26.Queue 8.31 0.87 0.42 18.08 0.00 33.00

Floor 27.Queue 7.78 0.86 0.30 18.37 0.00 33.00

Floor 28.Queue 7.63 0.87 1.04 20.40 0.00 32.00

Floor 29.Queue 5.85 0.72 0.67 18.49 0.00 32.00

Floor 30.Queue 5.87 0.82 0.47 17.74 0.00 33.00

Floor 31.Queue 6.85 0.88 0.50 17.36 0.00 32.00

Floor 32.Queue 6.84 0.91 0.43 20.20 0.00 33.00

Floor 33.Queue 7.63 0.89 0.67 18.17 0.00 33.00

Floor 34.Queue 8.50 0.93 0.91 20.26 0.00 33.00

Floor 35.Queue 9.56 0.85 0.99 19.55 0.00 33.00

Floor 36.Queue 10.27 0.83 1.46 20.35 0.00 33.00

Floor 37.Queue 10.83 1.00 1.08 22.92 0.00 34.00

Floor 38.Queue 9.51 0.85 1.98 19.49 0.00 33.00

304

Appendix H

Verification of ARENA Model

Table H1: Lift Simulation Model – Lift Evacuation Time (1618.88 seconds)

Table H2: Listing of ELVAC Analysis of 38 Story Building

*************************************************************************** ELVAC VERSION 1.00 WRITTEN BY DANIEL M. ALVORD AND JOHN H. KLOTE CONTRIBUTION OF THE NATIONAL INSTITUTES OF STANDARDS AND TECHNOLOGY (U.S.) NOT SUBJECT TO COPYRIGHT FOR COMPILED VERSION ONLY - PORTIONS (C) COPYRIGHT MICROSOFT CORPORATION, 1988 ALL RIGHTS RESERVED. DOCUMENTATION: KLOTE,J.H., ALVORD,D.M., AND DEAL,S., ANALYSIS OF PEOPLE MOVEMENT DURING ELEVATOR EVACUATION, NATIONAL INSTITUTE OF STANDARDS AND TECHNOLOGY, (U.S.), NISTIR 4730, 1992. **************************************************************************** Do you want to read about the model (Y or N)? N Enter the title of this run.

Example 38 Story Building

Enter 1 for SI units or 2 for English units: 1 Enter floors that the elevators serve (for example- B G 1-6) 1-38 Typical floor to floor height (m)? 3.0 Height exceptions: Floor range, height(m) (enter END to stop) END Discharge floor? 1 Time to outside after leaving elevator (s)? 0 Trip inefficiency factor (for example .1)? .1 Number of elevator cars in the group? 4 Normal operating velocity of car (m/s)? 3.15 Car acceleration (m/s**2)? 1.0 Elevator Full Load (people)? 16

305

------ MENU OF DOOR TYPES ------ ------------------------------------------------------------------------------------------------------------- Door type Door width Time to Open Door Transfer and Close Inefficiency

-------------------------------------------------------------------------------------------------------------- A Single-Slide 900mm (36in) 6.6 0.10 B Two-Speed 900mm (36in) 5.9 0.10 C Center-Opening 900mm (36in) 4.1 0.08 D Single-Slide 1100mm (42in) 7.0 0.07 E Two-Speed 1100mm (42in) 6.6 0.07 F Center-Opening 1100mm (42in) 4.6 0.05 G Two-Speed 1200mm (48in) 7.7 0.02 H Center-Opening 1200mm (48in) 5.3 0 I Two-Speed 1400mm (54in) 8.8 0.02 J Center-Opening 1400mm (54in) 6.0 0 K Two-Speed 1600mm (60in) 9.9 0.02 L Center-Opening 1600mm (60in) 6.5 0 M Two-Speed, 1600mm (60in) 6.0 0 Center-Opening N OTHER ***Pick one of the door choices A - M. If you wish to specify another type, enter N. H Other transfer inefficiency? 0.0 The start up time for automatically operated elevators is 41.25 seconds. Do you want to enter another value (Y or N)? N Typical Number of People per Floor? 32 People per floor exception: Floor Range, people (enter END to stop) END Percent of people on typical floor using elevator? 100 Percent usage exceptions: Floor Range, Percent (enter END to stop) END

306

Results:

Example 38 Story Building People per floor is 100. Distance between floors is 3.0 m or 10.0 ft. Elevator usage percent is 100.000% Normal car velocity is 3.00 m/s or 590.55 fpm. Car acceleration is 1.20 m/s2 or 3.94 ft/s2. Car full load is 16 people. Full load standing time is 40.26 s. Other transfer inefficiency is 0.0000 Trip inefficiency is 0.100 Door type: H Center-Opening 1200mm (48in) wide Doortime s 5.300 Door inefficiency 0.000

Figure H1 – ELVAC Lift Evacuation Time (1590.5 seconds)

307

Verification of Stair Model from Equation 2.8 (Chapter 2):

Listing of equation and variables used for verification

s

s

tWF

nNt +=1

where

n = 37 floors

N = 16 persons

Fs = 0.5 persons/meter/second

W =1.2 m

ts = 30 seconds

Then

t1 = 1116 seconds

308

Appendix J

Building Characteristics, HRR and Temperature

The enclosure characterisations for fire scenarios are given in Table J1. It is assumed

that the door of apartment is open while the occupants left the unit. The window

opening (0.9 m wide × 0.6 m high) from fire-affected units is considered in the

simulation. Wind pressure is exerted on window openings for a reasonably worst

fire scenario.

Table J1: Enclosure Characterisation of a Hypothetical Building (57m x 20m)

Enclosure Enclosure use Dimensions

L x b x h (m)

Openings

(m)

1 Kitchen/ Living/

Dining room/

Bedroom (fire-

affected SOU)

12m x 8.4m x 2.7m To # 2 – 2.0m x 0.9m – Door

(toward corridor)

2 Corridor 42m x 2.0m x 2.7m To # 1 – 2.0m x 0.9m – Door

To # 3 – 6m x 2.7m – Opening

via smoke lobby

To # 4 – 2.0m x 0.9m – Door

3 Lift lobby 40m x 2.0m x 2.7m To # 2 – 6m x 2.7m – Opening

via smoke lobby

4 Stair smoke lobby 3.0m x 2.0m x 2.7m To # 2 – 2.0m x 0.9m – Door

To # 5 – 2.0m x 0.9m – Door

5 Stair 2.0m x 2.0m x 2.7m To # 4 – 2.0m x 0.9m – Door

6 Opposite SOU 12m x 8.4m x 2.7m To # 2 – 2.0m x 0.9m – Door

(toward corridor)

The area of smoke lobby in double protected lift lobby is considered 1.5 times that of

stair smoke lobby as lifts are concentrated for full evacuation and stairs for half.

Heat Release Rate

The apartment unit of fire origin is approximately 96 m2 in the building. The

ceilings, floors and walls of the apartment unit are assumed to be composed of

concrete. The combustible fuel is based on the wood typically found in drawing

rooms of apartment buildings. The fire source was approximated as a rectangular

object (2.5 m × 2 m × 0.6 m). The fire growth is assumed based on a medium t-

squared curve fire Q= αt2 {where Q is the heat release rate (kW) and α is the fire

growth coefficient (0.01172 kW/s2)} to a constant peak value. The heat release rate

(HRR) is 1000 kW/m2. The combustion yields are given below:

309

SOOT_YIELD = 0.032

NU_O2 = 4.53

NU_CO2 = 4.12

NU_H2O = 3.21

MW_FUEL = 98

EPUMO2 = 11020

The graphical representations of HRR and temperature during Fire Scenario 1 are

given in Figures J1 and J2.

0

200

400

600

800

1000

1200

0 50 100 150 200 250 300 350

Time (second)

HRR (kW)

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

0 500 1000 1500 2000 2500 3000 3500 4000

Time (second)

HRR (kW)

Figure J1 – HRR during Fire Scenario 1 (HRR for 300 seconds and one hour)

310

0

100

200

300

400

500

600

700

800

900

0 500 1000 1500 2000 2500 3000 3500 4000

Time (second)

Temperature (C)

Figure J2 – Temperature during Fire Scenario 1

311

Appendix K

Occupants’ Movement Causing Door Opening and Closing

Table K1: Door Opening and Closing for Stair and Lift Lobbies and Lift Shaft by the

Entire Population

Protected lift lobby

door opening and

closing by 100 %

floor population

Stair door opening

and closing by 50%

floor population

(1st door)

Stair door opening

and closing by 50%

floor population

(2nd

door)

Lift landing door

opening and

closing

105 – 108

126 – 129

138 – 141

155 – 158

178 – 181

196 – 199

208 – 211

225 – 228

244 – 247

260 – 263

277 – 280

293 – 296

310 – 313

322 – 325

338 – 341

354 – 357

368 – 371

387 – 390

405 – 408

421 – 424

435 – 438

454 – 457

465 – 468

483 – 486

493 – 496

505 – 508

518 – 521

536 – 539

551 – 554

572 – 575

583 – 586

598 – 601

105 – 108

144 –147

183 – 186

215 – 218

235 – 238

266 – 269

296 – 299

336 – 339

375 – 378

418 – 421

452 – 455

486 – 489

499 – 502

520 – 523

545 – 548

574 – 577

620 – 623

108 – 111

147 – 150

186 – 189

218 – 219

238 – 241

269 – 272

299 – 302

339 – 342

378 – 381

421 – 424

455 – 458

489 – 492

502 – 503

523 – 526

548 – 551

577 – 580

623 – 626

For 2nd floor fire

303 – 313

1157 – 1167

1695 – 1705

For 19th floor fire

932 – 942

1634 – 1644

1920 – 1930

For 38th floor fire

390 – 400

1982 – 1992

2105 – 2115

Above table indicates that 100% of the population for lifts and 100% of the

population for two stairs (or 50% for one stair) are using lift lobby or stair lobby

312

doors. Lifts are also serving the floor and lift landing doors are opening and closing

in the lift lobby (see last column). The timings were determined from the stochastic

evacuation model and used in the FDS model for smoke leakages.

Table K2: Door Opening and Closing for Lift and Stair for Partial Population

Protected lift lobby

door opening and

closing by 25 %

floor population

Stair door opening and

closing by 37.5% floor

population

(1st door)

Stair door opening

and closing by 37.5%

floor population

(2nd

door)

105 – 108

178 – 181

244 – 247

310 – 313

368 – 371

435 – 438

493 – 496

551 – 554

105 – 108

144 –147

183 – 186

235 – 238

266 – 269

296 – 299

375 – 378

418 – 421

452 – 455

499 – 502

520 – 523

545 – 548

108 – 111

147 – 150

186 – 189

238 – 241

269 – 272

299 – 302

378 – 381

421 – 424

455 – 458

502 – 503

523 – 526

548 – 551

Above table indicates that lifts are used by 25% of the population while two stairs are

used by 75% of the population (or 37.5% for one stair).

313

Appendix L

Visibility Determination at a Focal Point

Generally, visibility is determined with the help of line of sight method. This

research determined the visibility at a focal point in the lift lobby, lift shaft and stair

shaft. For comparing visibility at a line of sight and at a focal point, an analysis is

given for lift lobby. Figure L1 illustrates the method for determining the visibility via

line of sight. Extinction coefficients were obtained from the averages of three

measurements along the lines of sight (LS).

Figure L1 – Visibility Determination at Three Lines of Sight and Lift Lobby

From the averages of extinction coefficients, it was determined that the time to

exceed tenability limit for visibility arrives at 198 seconds for line of sight 1 (LS1 –

stair1 exit), 230 seconds for LS2 (stair2 exit) and 200 seconds for LS3 (lift exit) (see

Figure L2). Time to exceed tenability limit for visibility arrives at 241 seconds in lift

lobby in Fire Scenario 1. The extinction coefficient is 0.5 m-1.

314

Visibility in corridor and lift lobby

0

5

10

15

20

25

30

35

40

45

0 500 1000 1500 2000 2500Time (second)

Extinction coefficient

(1/m

)

LS_1(TC_1)

LS_1(TC_2)

Occupant_TC

LS_3(TC_2)

LS_3(TC_1)

LS_2(TC_2)

LS_2(TC_1)

Figure L2- Recorded Visibility at Three Lines of Sight

The evacuees have the option of using either stair2 or lifts, after the stair1 becomes

invisible. Considering the walking speed, the occupants are required to travel an

additional 10 m for lifts and 30 m for stair2. Under the deteriorating situation, the

evacuees would be near the lift lobby at the earlier stage then to approach 20 m more

for stair2. Once the evacuees are inside the lift lobby, line of sight method can not be

applied due to small compartment.

315

Appendix M

Species Concentration and Fractional Effective Doses of Smoke, Gases and Heat

Concept Design B (Protected Lift Lobby): Figure M1 gives the results of smoke

extinction coefficient, concentration of gases, temperature and radiant heat flux in

Fire Scenarios 7 to 12. However, slight traces of smoke, gases or minor changes in

temperature and radiant heat flux in the protected lift shaft are not visible in a few

fire scenarios.

Fire Scenario 7 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 7 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

316

Fire Scenario 7 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 7 (Smoke and Gases in Stair)

0

2

4

6

8

10

12

14

16

18

20

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

317

Fire Scenairo 8 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 8 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

318

Fire Scenario 8 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 8 (Smoke and Gases in Stair)

0

2

4

6

8

10

12

14

16

18

20

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

319

Fire Scenario 9 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 9 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

320

Fire Scenario 9 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 9 (Smoke and Gases in Stair)

0

2

4

6

8

10

12

14

16

18

20

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

321

Fire Scenario 10 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 10 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

322

Fire Scenario 10 (Smoke and Heat in Lift Shaft)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 10 (Smoke and Gases in Stair)

0

4

8

12

16

20

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

323

Fire Scenario 11 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 11 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

324

Fire Scenario 11 (Smoke and Gases in Lift Shaft)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 11 (Smoke and Gases in Stair)

0

2

4

6

8

10

12

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

325

Fire Scenario 12 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440

Time (second)

Extinction coefficient

(1/m

) and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 12 (Temperature and Heat in Lift Lobby)

0

10

20

30

40

50

60

0 360 720 1080 1440

Time (second)

Temperature (C)

1

1.5

2

2.5

3

3.5

4

Radiant heat flux (kW/m

2)

Temperature Radiant heat flux

326

Fire Scenario 12 (Smoke and Gases in Lift Shaft)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenairo 12 (Smoke and Gases in Stair)

0

2

4

6

8

10

12

14

16

18

20

0 360 720 1080 1440

Time (second)

Extinction coefficient (1/m

)

and CO (ten ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Figure M1 – Smoke, Gases and Heat in the Lift Lobby, the Lift Shaft and the Stair

Shaft (Fire Scenarios 7 to 12)

327

Concept Design B (Protected Lift Lobby with 25% Population): Figure M2 gives

the results of smoke, gases, temperature and radiant heat flux in Fire Scenarios 13 to

18. Concept Design B is considered for lift and stair partial evacuation. Slight traces

of smoke, gases or minor changes in temperature and radiant heat flux in protected

lift shaft are not visible in a few fire scenarios.

Fire Scenario 13 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 13 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

328

Fire Scenario 13 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 13 (Smoke and Gases in Stair)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

329

Fire Scenario 14 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 14 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

330

Fire Scenario 14 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

0

5

10

15

20

25

Radiant heat flux (kW/m

2)

Extinction coefficient CO CO2 O2

Fire Scenario 14 (Smoke and Gases in Stair)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

331

Fire Scenario 15 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 15 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m2)

Temperature Radiant heat flux

332

Fire Scenario 15 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 15 (Smoke and Gases in Stair)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

333

Fire Scenario 16 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 16 (Temperature and Heat in Lift Lobby)

0

5

10

15

20

25

30

35

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

334

Fire Scenario 16 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 16 (Smoke and Gases in Stair)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

335

Fire Scenario 17 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 17 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

336

Fire Scenario 17 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2(%

)

Extinction coefficient CO CO2 O2

Fire Scenario 17 (Smoke and Gases in Stair)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

337

Fire Scenario 18 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

0 360 720 1080 1440 1800 2160

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 18 (Temperature and Radiant Heat Flux in Lift

Lobby)

0

5

10

15

20

25

30

35

40

0 360 720 1080 1440 1800 2160

Time (second)

Temperature (C)

1.5

1.6

1.7

1.8

1.9

2

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

338

Fire Scenario 18 (Smoke and Gases in Lift Shaft)

0

2

4

6

8

10

12

14

16

0 360 720 1080 1440 1800 2160

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 18 (Smoke and Gases in Stair)

0

10

20

30

40

50

60

70

80

90

100

0 360 720 1080 1440 1800 2160

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Figure M2 – Smoke, Gases and Heat in the Lift Lobby, the Lift Shaft and the Stair

Shaft (Fire Scenarios 13 to 18)

339

Concept Design C (Double Protected Lift Lobby): Figure M3 gives the results of

smoke, gases, temperature and radiant heat flux in Fire Scenarios 19 to 24. Slight

traces of smoke, gases or minor changes in temperature and radiant heat flux in

protected lift shaft are not visible in a few fire scenarios.

Fire Scenario 19 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient (1/m

)

and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 19 (Temperature and Radiant Heat Flux in Lift

Lobby)

19.7

19.8

19.9

20

20.1

20.2

20.3

20.4

20.5

20.6

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Temperature (C)

1.67

1.672

1.674

1.676

1.678

1.68

Radiant heat flux

(kW/m

2)

Temperature Radiant heat flux

(Temperature and radiant heat flux have not increased significantly in double protected lift shaft)

340

Fire Scenario 19 (Smoke and Gases in Lift Shaft)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

(Smoke and gases have not increased significantly in double protected lift shaft)

Fire Scenario 20 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

341

Fire Scenario 21 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 22 (Smoke and Gases in Lift Lobby)

0

10

20

30

40

50

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

342

Fire Scenario 23 (Smoke and Gases in Lift Lobby)

0

5

10

15

20

25

30

0 360 720 1080 1440 1800 2160 2520 2880 3240 3600

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Fire Scenario 24 (Smoke and Gases in Lift Lobby)

0

20

40

60

80

100

120

0 360 720 1080

Time (second)

Extinction coefficient

(1/m

) and CO (ppm)

0

5

10

15

20

25

CO2 (%) and O2 (%)

Extinction coefficient CO CO2 O2

Figure M3 – Smoke, Gases and Heat in the Lift Lobby and the Lift Shaft (Fire

Scenarios 19 to 24)

343

FED – Concept Design B (Protected Lift Lobby): Figure M4 gives the

representations for FED smoke, asphyxiant gases and heat in Fire Scenarios 7 to 12.

Traces of asphyxiant gases and temperature less than 60°C are not shown in a few

fire scenarios.

Fire Scenario 7 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 7 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

(Traces of CO are invisible)

344

Fire Scenario 7 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 8 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

345

Fire Scenario 8 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

(Traces of CO are invisible)

Fire Scenario 8 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

346

Fire Scenario 9 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 9 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

(Traces of CO are invisible)

347

Fire Scenario 9 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 10 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (second)

FED smoke and FED asphyxiant

Smoke Asphyxiant

348

Fire Scenario 10 (FED in Lift shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 10 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

349

Fire Scenario 11 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

Fire Scenario 11 (FED in Lift Shaft)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

350

Fire Scenario 11 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

(Fire on 38th level – Smoke accumulated in the stair top and has not diluted)

Fire Scenario 12 (FED in Lift Lobby)

0

1

2

3

4

5

0 5 10 15 20 25 30

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

351

Fire Scenario 12 (FED in Lift Shaft)

0

1

2

3

4

5

0 5 10 15 20 25 30

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 12 (FED in Stair)

0

1

2

3

4

5

0 5 10 15 20 25 30

Time (minute)

FED smoke

Smoke

Figure M4 – FED in Fire Scenarios 6 to 12

352

FED – Concept Design B (Protected Lift Lobby with Partial Evacuation): Figure

M5 gives the FED for smoke, gases and heat in Fire Scenarios 13 to 18. Concept

Design B is considered for lift and stair partial evacuation.

Fire Scenario 13 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 13 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

353

Fire Scenario 13 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 14 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

354

Fire Scenario 14 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 14 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

355

Fire Scenario 15 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 15 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

356

Fire Scenario 15 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 16 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

357

Fire Scenario 16 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Fire Scenario 16 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

(Traces of CO are invisible)

358

Fire Scenario 17 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 17 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

359

Fire Scenario 17 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 18 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

360

Fire Scenario 18 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 18 (FED in Stair)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

Figure M5 – FED in Fire Scenarios 13 to 18

361

FED – Concept Design C (Double Protected Lift Lobby): Figure M6 gives the

FED for smoke, gases and heat in Fire Scenarios 19 to 24.

Fire Scenario 19 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

Fire Scenario 19 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

362

Fire Scenario 20 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 20 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

363

Fire Scenario 21 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 21 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

364

Fire Scenario 22 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED asphyxiant

Smoke Asphyxiant

Fire Scenario 22 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

365

Fire Scenario 23 (FED in Lift Lobby)

0

1

2

3

4

5

0 10 20 30 40

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

Fire Scenario 23 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 10 20 30 40

Time (minute)

FED smoke

Smoke

366

Fire Scenario 24 (FED in Lift Lobby)

0

1

2

3

4

5

0 5 10 15 20 25

Time (minute)

FED smoke and FED

asphyxiant

Smoke Asphyxiant

Fire Scenario 24 (FED in Lift Shaft)

0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20 25

Time (minute)

FED smoke

Smoke

Figure M6 – FED in Fire Scenarios 19 to 24

367

Appendix N

Calculation of Fractional Effective Doses of Smoke, Gases and Heat

The fractional effective doses (FEDs) are calculated for Fire Scenario 1. The FEDs

are related to smoke, asphyxiant toxic gases and radiant and convective heats. The

safe and incapacitation doses are shown in bold. The safe criterion is taken as one-

tenth of incapacitation.

TIME (minutes) 0 1 2 3 4 5 6 7 8 9 10

Extinction coefficient (1/m) 0 0 0 0 0.23 0.97 2.20 3.90 6.37 9.06 11.99

CO (ppm) 0 0 0 0 10.42 44.78 104.88 189.88 316.39 447.33 588.22 CO2 (%) 0 0 0 0 0.09 0.39 0.91 1.65 2.75 3.89 5.11 O2 (%) 20.72 20.72 20.72 20.72 20.60 20.22 19.55 18.61 17.20 15.74 14.18

Temperature (oC) 20.00 20.00 20.00 20.00 24.28 33.61 42.39 48.97 55.23 53.54 51.43

Heat Flux (kW/m2)

Lift lo

bby

1.67 1.67 1.67 1.68 1.68 1.75 1.89 2.08 2.29 2.34 2.35

Extinction coefficient (1/m) 0 0 0 0 0.08 0.28 0.52 1.15 2.00 2.17 2.34

CO (ppm) 0 0 0 0 3.61 12.52 23.36 52.43 91.61 98.85 105.71 CO2 (%) 0 0 0 0 0.03 0.11 0.20 0.46 0.80 0.86 0.92 O2 (%) 20.72 20.72 20.72 20.72 20.68 20.58 20.46 20.13 19.69 19.61 19.54

Temperature (oC) 20.00 20.00 20.00 20.00 21.43 23.91 25.11 28.25 30.31 27.73 25.93

Heat Flux (kW/m2)

Lift sh

aft

1.67 1.67 1.67 1.67 1.68 1.68 1.68 1.69 1.71 1.71 1.72

FED smoke 0 0 0 0 0.46 1.93 4.40 7.80 12.75 18.11 23.97

FID CO 0 0 0 0 0 0 0 0.01 0.01 0.02 0.02

FID CO2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.73 2.18 2.78

FID O2 0 0 0 0 0 0 0 0 0 0 0

FED asphyxiant 0 0 0 0 0 0 0 0.01 0.02 0.03 0.06

ΣΣΣΣFED asphyxiant 0 0 0 0 0 0 0.01 0.01 0.03 0.06 0.12

FED heat 0 0 0 0 0 0 0 0 0 0 0

ΣΣΣΣFED heat

Lift lo

bby

0 0 0 0 0 0 0 0 0 0 0

FED smoke 0 0 0 0 0.16 0.56 1.04 2.30 3.99 4.35 4.67

FID CO 0 0 0 0 0 0 0 0 0 0 0

FID CO2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

FID O2 0 0 0 0 0 0 0 0 0 0 0

FED asphyxiant 0 0 0 0 0 0 0 0 0 0 0

ΣΣΣΣFED asphyxiant 0 0 0 0 0 0 0 0 0.01 0.01 0.01

FED heat 0 0 0 0 0 0 0 0 0 0 0

ΣΣΣΣFED heat

Lift sh

aft

0 0 0 0 0 0 0 0 0 0 0

369

TIME (minutes) 11 12 13 14 15 16 17 18 19 20

Extinction coefficient (1/m) 13.72 15.21 15.17 16.40 20.28 23.82 27.43 33.28 36.49 39.08

CO (ppm) 686.43 770.39 758.96 939.33 1151.45 1292.32 1387.30 1590.02 1707.85 1814.67 CO2 (%) 5.96 6.69 6.59 8.16 10.00 11.23 12.05 13.81 14.84 15.76 O2 (%) 13.08 12.15 12.28 10.27 7.91 6.34 5.29 3.03 1.72 0.53

Temperature (oC) 57.88 62.00 57.87 105.83 102.45 85.81 61.41 42.86 36.47 33.98

Heat Flux (kW/m2)

Lift lo

bby

2.54 2.67 2.59 4.30 4.55 3.77 2.80 2.32 2.16 2.12

Extinction coefficient (1/m) 1.81 3.37 4.18 4.76 6.45 5.81 4.87 5.26 9.51 8.22

CO (ppm) 81.19 153.47 190.60 222.04 314.39 276.91 224.50 251.25 545.83 494.88 CO2 (%) 0.71 1.33 1.66 1.93 2.73 2.41 1.95 2.18 4.74 4.30 O2 (%) 19.81 19.00 18.59 18.23 17.20 17.62 18.21 17.91 14.61 15.18

Temperature (oC) 24.12 27.74 28.32 35.25 49.30 41.13 31.35 42.18 107.90 125.83

Heat Flux (kW/m2)

Lift sh

aft

1.72 1.73 1.76 1.82 2.00 1.87 1.92 2.30 3.55 7.94

FED smoke 27.43 30.41 30.33 32.79 40.56 47.63 54.86 66.56 72.97 78.16

FID CO 0.02 0.03 0.03 0.03 0.04 0.05 0.05 0.06 0.06 0.07

FID CO2 3.30 3.81 3.74 5.11 7.39 9.44 11.14 15.84 19.44 23.40

FID O2 0.02 0.03 0.03 0.09 0.33 0.76 1.35 4.57 9.28 17.62

FED asphyxiant 0.10 0.14 0.13 0.26 0.63 1.20 1.91 5.48 10.48 19.16

ΣΣΣΣFED asphyxiant 0.22 0.36 0.49 0.75 1.38 2.58 4.49 9.96 20.44 39.60

FED heat 0.04 0.07 0.06 0.24 0.23 0.15 0.07 0.05 0.04 0.04

ΣΣΣΣFED heat

Lift lo

bby

0.04 0.11 0.18 0.42 0.65 0.80 0.87 0.92 0.95 0.99

FED smoke 3.61 6.75 8.37 9.51 12.90 11.62 9.74 10.52 19.03 16.44

FID CO 0 0 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01

FID CO2 1.00 1.00 1.00 1.00 1.73 1.62 1.48 1.55 2.58 2.36

FID O2 0 0 0 0 0 0 0 0 0 0

FED asphyxiant 0 0 0.01 0.01 0.02 0.01 0.01 0.01 0.04 0.03

ΣΣΣΣFED asphyxiant 0.01 0.02 0.02 0.03 0.04 0.06 0.07 0.08 0.11 0.15

FED heat 0 0 0 0 0 0 0 0 0.23 0.47

ΣΣΣΣFED heat

Lift sh

aft

0 0 0 0 0 0 0 0 0.23 0.70

370

TIME (minutes) 21 22 23 24 25 26 27 28 29 30

Extinction coefficient (1/m) 39.92 39.79 39.64 39.41 39.17 38.84 38.81 39.03 39.05 38.34

CO (ppm) 1858.65 1853.66 1847.71 1840.73 1833.29 1826.36 1822.89 1817.62 1810.51 1798.25 CO2 (%) 16.15 16.10 16.05 15.99 15.93 15.87 15.84 15.79 15.73 15.62 O2 (%) 0 0 0 0 0 0 0 0 0 0

Temperature (oC) 34.80 34.96 35.17 35.80 36.43 37.90 37.51 34.90 33.52 37.09

Heat Flux (kW/m2)

Lift lo

bby

2.17 2.17 2.21 2.20 2.19 2.14 2.13 2.02 1.99 2.08

Extinction coefficient (1/m) 10.29 11.35 10.88 13.11 15.79 18.46 21.03 28.06 35.13 39.60

CO (ppm) 698.44 857.93 813.19 861.59 1052.40 1076.88 1099.92 1515.43 1685.19 1801.34 CO2 (%) 6.07 7.45 7.06 7.48 9.14 9.36 9.56 13.17 14.64 15.65 O2 (%) 12.90 11.11 11.61 11.07 8.93 8.66 8.40 3.75 1.85 0.00

Temperature (oC) 177.10 228.44 227.52 162.43 170.44 114.50 74.29 86.21 46.11 34.17

Heat Flux (kW/m2)

Lift sh

aft

9.68 18.87 17.93 12.82 10.26 5.39 4.05 3.75 2.51 2.01

FED smoke 79.85 79.59 79.28 78.82 78.34 77.68 77.63 78.06 78.10 76.68

FID CO 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07

FID CO2 25.26 25.04 24.79 24.49 24.17 23.88 23.74 23.52 23.24 22.75

FID O2 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48

FED asphyxiant 25.18 25.16 25.14 25.11 25.08 25.06 25.04 25.03 25.00 24.96

ΣΣΣΣFED asphyxiant 64.78 89.94 115.08 140.19 165.27 190.33 215.37 240.39 265.39 290.35

FED heat 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.04

ΣΣΣΣFED heat

Lift lo

bby

1.03 1.07 1.11 1.15 1.19 1.23 1.26 1.30 1.33 1.37

FED smoke 20.57 22.70 21.77 26.21 31.58 36.92 42.06 56.12 70.26 79.20

FID CO 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.04 0.05 0.05

FID CO2 3.37 4.44 4.11 4.47 6.22 6.49 6.76 13.92 18.69 22.87

FID O2 0.02 0.06 0.04 0.06 0.19 0.22 0.25 3.10 8.65 23.48

FED asphyxiant 0.09 0.16 0.14 0.17 0.37 0.41 0.46 3.69 9.52 24.62

ΣΣΣΣFED asphyxiant 0.23 0.40 0.53 0.70 1.07 1.48 1.94 5.62 15.14 39.76

FED heat 1.14 2.72 2.65 1.03 1.05 0.32 0.13 0.15 0.05 0.03

ΣΣΣΣFED heat

Lift sh

aft

1.84 4.56 7.20 8.23 9.28 9.60 9.72 9.87 9.92 9.96

371

TIME (minutes) 31 32 33 34 35 36 37 38 39 40

Extinction coefficient (1/m) 38.31 38.18 38.05 37.86 37.71 37.75 37.69 38.66 39.66 40.25

CO (ppm) 1797.17 1794.63 1791.25 1788.54 1785.84 1780.32 1774.39 1789.59 1801.42 1806.53 CO2 (%) 15.61 15.59 15.56 15.54 15.51 15.47 15.41 15.55 15.65 15.69 O2 (%) 0 0 0 0 0 0 0 0 0 0

Temperature (oC) 37.19 37.74 38.21 39.33 40.07 38.83 38.29 33.11 27.36 23.75

Heat Flux (kW/m2)

Lift lo

bby

2.08 2.09 2.11 2.14 2.16 2.12 2.10 2.00 1.84 1.75

Extinction coefficient (1/m) 38.67 38.55 37.23 38.16 37.64 37.59 38.64 39.79 40.32 40.69

CO (ppm) 1775.32 1764.11 1757.81 1762.48 1750.65 1748.82 1741.10 1757.44 1771.22 1773.84 CO2 (%) 15.42 15.33 15.27 15.31 15.21 15.19 15.13 15.27 15.39 15.41 O2 (%) 0 0 0 0 0 0 0 0 0 0

Temperature (oC) 39.43 39.47 49.99 42.51 45.88 46.28 36.80 29.20 26.28 23.76

Heat Flux (kW/m2)

Lift sh

aft

2.11 2.17 2.37 2.29 2.41 2.48 2.12 1.88 1.82 1.76

FED smoke 76.61 76.37 76.11 75.72 75.43 75.49 75.37 77.32 79.32 80.51

FID CO 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07

FID CO2 22.70 22.60 22.47 22.37 22.26 22.05 21.82 22.41 22.87 23.08

FID O2 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48

FED asphyxiant 24.95 24.94 24.93 24.92 24.92 24.90 24.88 24.93 24.97 24.99

ΣΣΣΣFED asphyxiant 315.31 340.25 365.18 390.11 415.02 439.92 464.80 489.73 514.69 539.68

FED heat 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03

ΣΣΣΣFED heat

Lift lo

bby

1.41 1.45 1.49 1.53 1.57 1.60 1.64 1.68 1.71 1.73

FED smoke 77.34 77.10 74.47 76.32 75.28 75.17 77.28 79.57 80.64 81.38

FID CO 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05

FID CO2 21.86 21.44 21.20 21.37 20.94 20.87 20.60 21.19 21.70 21.80

FID O2 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48 23.48

FED asphyxiant 24.55 24.52 24.51 24.52 24.49 24.49 24.47 24.51 24.54 24.55

ΣΣΣΣFED asphyxiant 64.31 88.83 113.34 137.86 162.35 186.83 211.30 235.80 260.34 284.89

FED heat 0.04 0.04 0.05 0.04 0.05 0.05 0.04 0.03 0.03 0.03

ΣΣΣΣFED heat

Lift sh

aft

10.00 10.04 10.09 10.13 10.18 10.23 10.27 10.30 10.33 10.36