Aircraft Evacuation Testing: Research and Technology Issues

Aircraft Evacuation Testing: Research andTechnology Issues

September 1993

OTA-BP-SET-121NTIS order #PB94-107620

Contents

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Findings and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Benefits and Limitations of Evacuation Demonstrations . . . . . . . . . . . . . . . . . . . . . . . . . . 1Models and Simulations for Evacuation Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2Data Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2Aircraft Evacuation Performance and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2

Chapter 1. Background and Regulatory Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Federal Aviation Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Recent History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Exits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Slides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Training and Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

National Transportation Safety Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Accident/Incident Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

United Kingdom Civil Aviation Authority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Chapter 2. Evacuation Demonstrations for Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Limitations of Full-Scale Demonstrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Test and Data Validity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Development of Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Other Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Chapter 3. Research and Technologies for Evacuation Systems . . . . . . . . . . . . . . . . . . . . 27Cabin Safety Research and Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Cabin Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Emergency Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Risk/Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Training and Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Evacuation Research and Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Evacuation Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Computer Modeling and Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Chapter 4. Findings and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Benefits and Limitations of Evacuation Demonstrations . . . . . . . . . . . . . . . . . . . . . . . . 43Models and Simulations for Evacuation Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Data Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Aircraft Evacuation Performance and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Appendix

Project Staff

John Andelin1

Assistant DirectorScience, Informationl and Natural

Resources Division

Peter Blair2

Assistant DirectorIndustry, Commerce, and International

Security Division

Nancy CarsonProgram ManagerScience, Education, and Transportation

PRINCIPAL STAFFKevin DepartProject Director

KELLEY SCOTTPrincipal Analyst

Greg WallaceResearch Assistant

OTHER CONTRIBUTING STAFFElizabeth Sheley/n-House Contractor

Jeanne OlivierDetailee, Port Authorityof New York and New Jersey

ADMINISTRATIVE STAFFMarsha FennTechnical Editor

Gay JacksonPC Specialist

Tamara KowalskiSecretary

1 Through August 1993.2 After August 1993.

Advisory Panel

Robert W. BakerExecutive Vice President, OperationsAmerican Airlines, Inc.Dallas, TX

Najeeb E. Halaby, Panel ChairChairman, Safair International

McLean, VA

William F. Ballhaus, Jr.PresidentMartin Marietta Aero & Naval SystemsBaltimore, MD

Robert A. DavisVice President, EngineeringBoeing Commercial Airplane GroupSeattle, WA

Donald D. EngenPresidentAOPA Air Safety FoundationFrederick, MD

Edmund S. GreensletPresidentESG Aviation ServicesPonte Vedra Beach, FL

David HaaseExecutive Central Air Safety ChairmanAir Line Pilots AssociationWhitehouse, OH

Jonathan HoweZuckert, Scoutt & RasenbergerWashington, DC

Noreene KoanChairperson, National Air Safety CommitteeAssociation of Flight AttendantsBelmont, CA

Richard LivingstonAir Transportation ConsultantArlington, VA

T. AlIan McArtorPresidentFEDEX Aeronautics Corp.Memphis, TN

Clinton V. Oster, Jr.ProfessorSchool of Public and Environmental AffairsIndiana UniversityBloomington, IN

Willard G. Plentl, Jr.DirectorDivision of Aviation for North CarolinaRaleigh, NC

Robert W. SimpsonDirectorFlight Transportation LabMassachusetts Institute of TechnologyCambridge, MA

Richard SwaugerAir Traffic ConsultantFairfax, VA

Patricia F. WailerDirectorTransportation Research InstituteUniversity of MichiganAnn Arbor, Ml

NOTE: OTA appreciates the valuable assistance and thoughtful critiques provided by theadvisory panel members. The panel does not, however, necessarily approve,disapprove, or endorse this background paper. OTA assumes full responsibility for thebackground paper and the accuracy of its contents.

Reviewers

Ricky DavidsonAccident Survial CommitteeAir Line Pilots Association

Steven EricksonAir Transport Associationand Aviation Rulemaking Advisory CommitteeEmergency Evacuation Issues

William HathawayOffice of Transport and Information

Resources ManagementVolpe National Transportation

Systems CenterU.S. Department of Transportation

Pierre HugginsEngineering & Air Safety DepartmentAir Line Pilots Association

Matthew McCormickSurvival Factors DivisionNational Transportation Safety Board

Jeffrey MarcusProtection and Survival LaboratoryCivil Aeromedical InstituteFederal Aviation Administration

Stephanie MarkosOffice of Systems EngineeringVolpe National Transportation

Systems CenterU.S. Department of Transportation

Nora MatthewsSurvival Factors DivisionNational Transportation Safety Board

Constantine SarkosFire Safety BranchTechnical CenterFederal Aviation Administration

William ShookMD-1 1/DC-10 Product DefinitionDouglas Aircraft Co.

George Veryioglou747/767 Payload SystemsBoeing Commercial Airplane Group

NOTE: OTA appreciates the valuable assistance and thoughtful critiques provided by thereviewers. They do not, however, necessarily approve, disapprove, or endorse thisbackground paper. OTA assumes full responsibility for the background paper and theaccuracy of its contents.

EXECUTIVE SUMMARY

Executive Summary

The U.S. air transportation industry has anoutstanding safety record. Yet passenger safetyaboard U.S. airlines remains a continuing issuefor the public, the Federal Aviation Admini-stration (FAA), and the U.S. Congress. Oneconcern is that aircraft be evacuated quicklyand safely in an emergency.

FAA certification criteria and test methodsare integral to evaluating the evacuation capa-bilities of new aircraft. In November 1991, theSubcommittee on Government Activities andTransportation of the House Committee onGovernment Operations requested that theOffice of Technology Assessment (OTA) “. . .study the prospects for improving existingmethods of evacuation testing in light of theneed to balance realism against the safety oftest participants. ” 1 For this study, OTA exami-ned regulatory, research, and technologyissues related to passenger safety and evacu-ation testing.

Pursuant to Federal aviation regulations, air-craft manufacturers conduct full-scale demon-strations to show an airplane’s basic evacuationcapability and to evaluate crew training. FAAfull-scale evacuation demonstration criteriainclude the following requirements:

• All passengers and crew must be evacu-ated from the aircraft to the groundwithin 90 seconds;

• The demonstration must be conductedduring the dark of night or with the darkof night simulated, so that the airplane’semergency lighting system provides theonly illumination of exit path and slides;

• specified mix of passengers “in normalhealth” must be used;

• Not more than 50 percent of the emer-gency exits may be used.

FINDINGS AND CONCLUSIONS

Benefits and Limitations ofEvacuation Demonstrations

• Full-scale demonstrations are costly andexpose participants to significant hazards.The cost of conducting a full-scale dem-onstration can exceed $1 million. Onaverage, approximately 6 percent of par-ticipants are injured during full-scaletests. While most injuries have beenminor, broken bones and paralysis haveoccurred. Fewer and less severe injuriesthan average occurred in the December1992 MD-1 1 certification test in whichslides were replaced with ramps.

• A full-scale demonstration simulatesevacuation for only a narrow range ofemergency conditions-an aborted take-off at night involving no structural dam-age, cabin fire, or smoke, for a distinctsubset of potential passengers (i. e., nochildren, persons with disabilities, ornon-English speaking passengers).

• Demonstrations provide only a benchmarkfor consistent evaluation of various seatingand exit configurations. The requirementto demonstrate complete evacuation with-in 90 seconds is not an adequate per-formance standard for measuring actualevacuation capabilities.

• Present evacuation certification rules donot encourage new technology develop-ment for extending the period of surviv-ability in post-crash fires. The evacuationdemonstration criteria are inflexible, re-gardless of the availability of technolo-gies that could extend the period ofsurvivability within the cabin.

1 Barbara Boxer$ chajr, Subcommittee on GovernmentActivities and Transportation, House Committee on Gov-ernment Operations, letter to John Gibbons, director,OffIce of Technology Assessment, Nov. 19, 1991.

2 ● Aircraft Evacuation Testing: Research and Technology Issues

Models and Simulations forEvacuation Certification

At present, neither certification by full-scale demonstration nor by purely ana-lytical certification methods is acceptableto all segments of the aviation commu-nity.

The certification process will likely con-tinue to rely on human test subjects in theforeseeable future. However, a combina-tion of analysis and partial demonstra-tions or component tests can bedeveloped to minimize the risk of injuryand provide more comprehensive data onaircraft performance than full-scale dem-onstrations.

Using aircraft manufacturers’ analyticalmodels, passenger egress rates throughexisting aircraft components are predict-able. The results of industry analysestypically correlate well with observedrates through doors, aisles, slides, andother components under consistent testconditions.

Human behavior in certification tests maybe empirically modeled using data fromprior demonstrations, but cannot yet bereliably “simulated.” Estimates for aver-age reaction times and egress rates areknown for evacuation during controlledconditions. Because few reliable dataexist on human behavior during acci-dents, the variations in human judgmentand decisionmaking that might be ex-pected for changing hazardous conditionscannot be predicted. These data cannotbe obtained from current demonstrationrequirements, which do not addressmotivational effects or other behavioralfactors that often exist in a real emer-gency.

Recent computer simulation efforts mayprovide the technology base for adynamic aircraft evacuation simulationcapability, but the additional psychologi-cal data required for validating behav-

ioral assumptions will be difficult toattain.

Data IssuesFAA and industry could collect additionalexperimental data to support and validateevacuation models/simulations. AlthoughFAA’s present test fuselage is adequatefor studying evacuation scenarios in sin-gle-aisle, narrow-body airliners, neitherFAA nor any other regulatory agency hasa facility that can be used to analyzeegress from double-aisle, wide-bodytransports.

Data on injuries related to aircraftevacuation testing are not readily avail-able, nor are they classified by severity.Data from actual emergency evacuationsare unevenly collected and analyzed.Neither FAA nor the National Transpor-tation Safety Board collect information onprecautionary evacuations.

Aircraft Evacuation Performanceand Safety

Ž Survivability in commercial air transportsis improving, largely through the intro-duction of highly fire-retardant materialsand more crashworthy seats, restraints,and overhead bins. Though still a signifi-cant threat, fire has become less of a riskin survivable accidents. In the early1980s, FAA attributed 40 percent offatalities in survivable accidents, ap-proximately 20 percent of total fatalities,to fire effects.2 Between 1985 and 1991,approximately 10 percent of fatalitiesaboard U.S. airlines were related to fire.3

Ž Crew training and passenger motivationare as crucial to successful evacuations asthe aircraft’s design and equipment.

2 Constantine P. Sarkos, mamger, Fire Safety Branch,FM Technical Center, persoml communication, June 3,1993.3 Offlce of Twhnology Assessment, based on FAA andNational Transportation Safety Board data.

Executive Summary ● 3

Flight attendant training, done in cabinmockups without passengers, may notprovide crew members with sufficientskills for assessing flow control problemsand motivating passengers to evacuatemore efficiently. Simulation technologiesmay enhance training in passenger man-agement and use of emergency equip-ment.

Passenger safety may be better improved byextending the period of survivability than byattempting to reduce the time required forevacuation. New technologies intended to delaydeadly heat and toxicity levels after a crashwould save more lives than feasible configura-tion changes intended to speed evacuation,according to a British analysis of aircraft acci-dents. Furthermore, demographic trends indi-cate that the average mobility of aircraftpassengers will decrease in the future.

Chapter 1

BACKGROUND ANDREGULATORY CONTEXT

CHAPTER 1

Background and Regulatory Context

Air travel, for business or pleasure, is anindispensable part of American life. The in-tegrity of the aircraft and air traffic manage-ment systems, and the vigilance and skill ofthose who operate them, are the cornerstonesof safe air travel. Should an emergency occur,passenger survival depends on the ability of theaircraft and its contents to withstand impact andthe post-crash environment, on the design andeffectiveness of escape routes and equipment,and on the crew’s ability to help passengersevacuate the aircraft as quickly as possible.Ensuring crashworthiness, prolonging surviv-able conditions within the cabin, and providingquick egress are major thrusts of Federal safetyprograms. One of the final measures of an air-craft’s readiness for operation is the full-scaleevacuation demonstration.

Pursuant to existing Federal aviation regula-tions, aircraft manufacturers must demonstratethat a new or substantially revised type of air-craft can be completely evacuated under speci-fied conditions in less than 90 seconds.Manufacturers have conducted more than 20full-scale evacuation demonstrations since1969, involving over 7,000 volunteers andairline crew personnel.1 On average, 6 percentof full-scale demonstration participants receiveinjuries, which typically range from scrapesand bruises to broken bones. In October 1991,a test participant became permanently para-lyzed after being injured during a McDonnellDouglas evacuation demonstration test for cer-tification of an MD-11 airplane.2 This renewed

concern on the part of Congress, manufactur-ers, and passenger, pilot, and flight attendantgroups about the safety of the certificationprocess. The air transportation community isstriving to find ways to reduce the likelihood ofinjuries in future tests.3 At the same time, thecommunity is considering the net benefits offull-scale demonstrations as a requirement fortype certification.

The Federal Aviation Administration’s (FAA)Aviation Rulemaking Advisory Committee (ARAC)is studying options for the development ofperformance standards to replace evacuationsafety design criteria. Recent congressionalActivity includes 991 hearings by the subcommittee onGovernment Activities and Transportation of theHouse Committee on Government Operations,4

and a request for the General AccountingOffice (GAO) to assess both the implementationof recent cabin materials regulations and theadequacy of the 90-second evacuation testcriterion. GAO’s investigation of evacuationdemonstration issues is on hold, pending com-pletion of the ARAC Subcommittee’s efforts.

In November 1991, the Subcommittee onGovernment Activities and Transportation ofthe House Committee on Government Opera-tions requested that the Office of TechnologyAssessment (OTA) “. . . study the prospects forimproving existing methods of evacuationtesting in light of the need to balance realism

1 Webster C. Heath, manager, Technical Liaison,Industry Regulatory Affairs, Douglas Aircraft Co.,

5ersoml communication, Sept. 24, 1992.

At the time of the unsuccessful demonstration, theMD-1 1 was certificated to carry 390 passengers. TheDouglas Aircraft Company again attempted to certificatethe aircra!l for 410 passengers on December 11, 1992,employing ramps instead of slides to minimize thepotential for injury to test participants. The second,revised demonstration satisfied FAA’s requirements forcertification. See section on evacuation demonstrations inchapter 2.

3 In 1993, Boeing and Airbus wi]] attempt fd]+de

demonstrations with their respective B-767 and A-330aircrafi.4 See Us. COngress, House Committee on GovernmentOperations, “Issues in Aircraft Cabin Safety and CrashSurvivability: The USAir-Skywest Accident, ” HouseReport 102-501, Apr. 22, 1992.5 See us Genera] Accounting office, Auiazion SafefY:Slow Progress in Making Aircr@ Cabin InteriorsFireproof, GAO/RCED-93-37 (Washington, DC: U.S.Government Printing Office, January 1993).


against the safety of test participants.“6 Thesafety and utility of testing methods are twoprimary concerns. Investigating the evacuationperformance of an aircraft under actual emer-gency conditions would subject test participantsto significant risk of injury. Computer simula-tion has emerged as a potential tool for evaluat-ing numerous evacuations in changing fire andcabin configurations, trials too hazardous toconduct with human participants.

OTA examined a range of regulatory, re-search, and technology issues related to pas-senger safety and evacuation testing, includingthe scientific validity of the full-scale demon-stration as a measure of evacuation capability.While this document describes alternatives toand the relative merits of current full-scaledemonstration requirements, it does not provideresearch or regulatory policy options forCongress. Key issues this background paperdoes discuss include:

the current evacuation standards and therole of evacuation testing;data collection and analysis to evaluateperformance of evacuation systems inactual accidents or incidents;potential near-term improvements todemonstration tests that may reduce thelikelihood of injury to participants;the role of mathematical modeling and/orcomputer simulations in reducing theneed for human participation in theevaluation of evacuation procedures andequipment; and,economic concerns.

Federal authority for aircraft safety andevacuation standards lies with FAA. TheNational Transportation Safety Board (NTSB)investigates aviation accidents or incidents andmakes recommendations regarding safety im-provements, Outside the United States, theCivil Aviation Authority (CAA) of the UnitedKingdom is most active in improving the

6 Barbara Boxer, chair, Subcommittee on GovernmentActivities and Transportation, House Committee onGovernment Operations, letter to John Gibbons, director,Office of Technology Assessment, November 19, 1991.

evacuation capability of the aircraft and crewunder its authority. This section describes theroles of the U.S. agencies and CAA, alongwith requirements for emergency equipment,cabin safety operations, and crew training.

FEDERAL AVIATIONADMINISTRATION

FAA responsibility for cabin safety encom-passes the development and enforcement of theFederal Aviation Regulations (FAR);7 FAAconducts and sponsors research and develop-ment (R&D) programs related to cabin safety tosupport its rulemaking activities.

Cabin safety certification and complianceauthority rests primarily with FAA’s AircraftCertification Service, which manages airwor-thiness offices throughout the United States andsets airworthiness standards, and the FlightStandards Service, which regulates air carrieroperations and crew training and standards.The Certification Service establishes minimumstandards for the design and manufacture of allU.S. aircraft and certifies that all aircraft meetthese standards prior to introduction into serv-ice. g Certification authority for large com-mercial aircraft rests with FAA’s TransportAircraft Directorate in Seattle, Washington.

Under the Executive Director for SystemDevelopment, the FAA Technical Center inAtlantic City supports regulatory developmentthrough in-house and contracted R&D, particu-larly in the areas of crashworthiness and firesafety. FAA’s Civil Aeromedical Institute(CAMI) in Oklahoma City, under the purviewof the Office of Aviation Medicine, is anothercontributor to crashworthiness research andevacuation standards evaluation. CAMI con-ducts pilot training research and, along with the

7 Tlt]e 14, c~pter I Of the Code of ‘dera]Regulations.8 U.S. Congress, Office of Technology Assessment,we Skies for Tomorrow: Aviation Sa$ety in a CompetitiveEnvironment, OTA-SET-381 (Washington, DC: U.S.Government Printing Office, July 1988), p. 56, availablefrom OTA’s Science, Education, and TransportationProgram.

Chapter 1—Background and Regulatory Context ● 7

Technical Center, supports accident investiga-tion.

To obtain type9 certification, manufacturersof aircraft having more than 44 passenger seatsmust conduct emergency evacuation demon-strations that test the following:

• basic aircraft design;• the efficiency with which passengers can

safely be evacuated from the aircraft;• the emergency evacuation system; and• the manufacturer’s FAA-approved emer-

gency evacuation procedures. 10

Manufacturers typically elect to conduct thedemonstration to serve both the type and op-erating certification requirements. Figure 1-1shows the procedure for aircraft type certifica-tion, including airframe, seats, and evacuationdemonstration.

The number, duties, and location of flightattendants are specified in 14 CFR 121. Flightattendants perform numerous safety-relatedduties before, during, and after each flight. Theindividual air carriers provide flight attendantswith their initial emergency procedure training,and additional training each year thereafter.Current regulations require 1 flight attendantfor every 50 passenger seats.

The description and demonstration of emer-gency evacuation procedures are integral partsof the operating certificate application proce-dure.11 Once all application and demonstrationrequirements have been satisfied, FAA issuesan air carrier operating certificate, specifyingthe terms, conditions, and limitations of opera-tion. 12

Recent HistoryFlight and cabin safety comprises a signifi-

cant portion of FAA’s rulemaking and researchduties. In 1979, FAA formed the Special Avia-tion Fire and Explosion Reduction AdvisoryCommittee to assess related research and regu-latory needs. For several years, following thecommittee’s final report in 1980, FAA empha-sized the development of improved fire testmethods and cabin interior material criteria. 13

Several of the projects and rules related to im-proving fire safety are identified in table 1-1.

On August 22, 1985, as a Boeing B-737attempted to take off from Manchester Intern-ational Airport (England), its left engine disinte-grated, causing a fuel spill and a subsequentfuel-fed cabin fire. Of the aircraft’s 137 occu-pants, 55 died aboard the burning aircraft.Most were later found to have been incapaci-tated from smoke and toxic gas inhalation. Ac-cident analysis indicated that limited access tooverwing exits and competition among passen-gers delayed evacuation of the plane.

In September 1985, FAA convened a publictechnical conference related to emergencyevacuation from transport aircraft. 14 Discussioncentered on emergency exits and slides, full-scale evacuation demonstrations, and crewtraining. FAA formed an Emergency Evacu-ation Task Force to coordinate activities ofthree working groups established during theconference--Design and Certification, Trainingand Operations, and Maintenance and Reliabil-ity.15 In 1986, FAA agreed to develop andissue rulemaking and/or advisory material on

9 Type, as defined in 14 CFR 1.1, means those aircrafi

that are similar in design (e.g., DC-10 Series 30 andSeries 40, B-747-200 and B-747-400).10 Federa] Aviation Administration, “Evaluate FARPart 21 Emergency Evacuation/Ditching Procedures/Demonstration, ” Airworthiness Safety InspectorsHandbook, vol. 2 (Washington, DC: November 1988),ch. 77-1.11 U.S. General A c c o u n t i n g OffIce, AViarion s@?V:Procedures for Registering and Certifying Air Carriers,GAO/RCED-87-l 15FS (Washington, DC: May 5, 1987),

Y15.

2 Ibid., p. 18.

13 Consmntlne Sarkos, Federal Aviation Administration,“Full Scale Test Results and Status of FAA’s Cabin SafetyProgram, ” Proceedings of the Flight Safety Foundation/Federal Aviation Administration International Aircrafioccupant Safety Conference and Workshop,DOT/FAA/OV-89-2 (Washington, DC: U.S. Departmentof Transportation, August 1989), p. 179.14 50 Federa/ Register 32087 (Aug. 8, 1985).15 For fi~er description of the conference and itsworking groups, see U.S, Department of Transportation,Federal Aviation Administration, Tmk Force Report onEmergency Evacuation of Transport Airplanes, vol. 1,summary Report, DOT/FAA/VS-86/l,1 (Washington,DC: ]u]y 1986).


Figure l-l--Federal Aviation Administration Proceduresfor Issuing an Aircraft Type Certificate

Applicant submits application to theAircraft Certification Directoratea ofFAA accompanied by a three-waydrawing of the aircraft.

Certification Directorate makes aninitial determination of the adequacyof the proposal.

Certification Directorate inspects andtests the aircraft for airworthiness,including:

-flight tests-ground tests-compliance with structuralrequirements

FAA issues an aircraft Type Certificate after the applicant has met all requirements.I

aThe FAA office responsible for evaluating compliance with certification requirements for a given class of aircraft.

SOURCE: Office of Technology Assessment, 1993.

Chapter I-Background and Regulatory Context . 9

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

KEY:

Table 1-1—Federal Aviation Administration Fire Safety Program

Project/subject

Seat cushion fire blocking14 CFR Part 135 extension

Floor proximity lighting

Lavatory smoke detectors

Lavatory automatic fireextinguisher

Halon fire extinguishers

Class E cargo compartmentfire extinguishers

Class C & D cargo orbaggage compartments

Improved cargo liners

Crew member PBE forflight attendants

Heat release-interior materials

Smoke density-interior materials

Fuel system crash resistance

Small airplane crashresistant fuel systems

Passenger PBE

Cabin water spray system

Class C & D cargo compartments

Action

Final RuleFinal Rule

Final Rule

Final Rule

Final Rule

Final Rule

Final Rule

Final Rule

NPRM

Final Rule

Final Rule

Final Rule

NPRM

NPRM

Rulemaking dropped

R&D

R&D

Issued (compliance)

10/26/84 (11/26/87)11/25/87 (12/1/88)

10/26/84 (11/26/86)

3/26/85 (10/29/86)

3/26/85 (4/29/87)

3/26/85 (4/29/86)

3/26/85 (10/29/85)

5/16/86 (6/16/86)

10/28/87 (2 years)

5/26/87 (7/6/89)

7/21/86 (8/20/88and

8/20/90)

8/19/88 (8/20/90)

4/26/89

2/14/90

NPRM = Notice of Proposed Rulemaking; PBE=protective breathing equipment.

SOURCE: John J. Petrakis, “FAA Occupant Protection and Cabin Safety Overview,” Proceedings ofthe Flight Safety Foundation/Federal Aviation Administration International AircraftOccupant Safety Conference and Workshop, DOT/FAA/OV-89-2 (Washington, DC: U.S.Department of Transportation, August 1989), p. 56; and Office of TechnologyAssessment, 1993.

10 . Aircraft Evacuation Testing: Research and Technology Issues

29 specific proposals recommended by theworking groups. All but 2 of the 29 recom-mendations resulted in FAA action by the be-ginning of 1992. 16

Of the numerous elements of an aircraft’semergency evacuation system, exit and slidedesign, flight attendant training, and full-scaleevacuation demonstrations required for typecertification have engendered the most attentionand public debate. The key design and trainingrequirements and related areas of contentionare discussed below; full-scale demonstrationsare described in a subsequent section.

ExitsSince 1967, FAA has regulated the location

of emergency exits on airplanes with the fol-1owing requirements:

Specific types and numbers of exits mustbe provided for given numbers of pas-sengers;Exits must be located to provide the mosteffective means of passenger evacuation;andExits must be distributed as uniformly aspractical with respect to passenger seat-ing.

Exit arrangement, deployment, and marking,and emergency lighting must meet specific cri-teria. 17 See box 1-A for a description of varioustypes of aircraft exits.

In 1986, after analysis of the Manchesteraccident indicated congestion at the overwingexit contributed to slow evacuation, CAA is-sued an airworthiness notice for alternateminimum requirements for seating next tooverwing exits. 18 FAA, in turn, authorized

16 tJ.s. Department of Transportation,Administration, Task Force ReportEvacuation of Transport Airplanes, vol.

Federal Aviationon Emergency

3, Final Repoti,FAA/AIR-92-01 (Washington, DC: Jan. 23, 1992), p. 11.17 14 CFR 25.807, Amendment 25-15, 32 FederalRegister 13263 (Sept. 20, 1967).18 Cjvil Aviatjon Authority, United Kingdom, “Access toand Opening of Type III and Type IV Emergency Exits, ”Airworthiness Notice No. 79, 1986.

CAMI to evaluate the proposed changes underconditions that would enable comparison withthe minimum requirements delineated in theFAR.19 CAMI conducted the evacuation testsin 1986 and 1991. In May 1992, FAA issued afinal rule requiring transport aircraft having 60or more passenger seats20 to make Type IIIoverwing exits more accessible (e.g., providewider passageways between seats or removethe seat adjacent to the exit) .21 With compliancerequired by December 1992, the rule also man-dated that all aircraft with Type III exits displayplacards that describe how to open and stow theexit, and state the exit door’s weight.

SlidesTo prevent injury to passengers and crew

escaping through floor-level exits located morethan 6 feet above the ground, assist devices(e.g., slides or slide-rafts) are required. Therapid deployment, inflation, and stability ofevacuation slides are critical elements of theevacuation system. Slide design and perform-ance requirements are contained in technicalstandard orders, while general slide require-ments are found in 14 CFR 25. In 1983, FAArevised the requirements to specify criteria forresistance to water penetration and adsorption,

19 Paul G. Wsmussen and Charles B. Chittum, meInfluence of Aaj”acent Seating Configurations on EgressZhrough a Type III Emergency Exit, Final Repoti,DOT/FAA/AM-89/14 (Washington, DC: December1989). Although the fiml report was not released until1989, the tests were authorized in 1986. Additional testswere conducted in 1991; see Garnet A. McLean et al.,Civil Aeromedical Institute, Eflects of Seating Con-figuration and Number of Type III Ekits on EmergencyAircr@ Evacuation, Final Report, DOTIFAA/AM-92/27(Washington, DC: U.S. Depanment of Transportation,August 1992).20 FAA considers that a minimum of 60 passenger seats,which typically requires at least 15 rows, enablesoperators to provide the additioml access through seat rowadjustment without a loss of revenue. 57 Federal Register19239 (May 4, 1992).21 57 Fe~era/ Register 19220 (Mav 4. 1992J.

Box l-A--Description of Passenger Emergency Exits

Type Aa Rectangular opening at least 42 inches wide by 72 inches high, with specifieddimensions for passageways to main and cross aisles. Floor-level Type A exitsmust be equipped with dual-lane emergency slide. Overwing Type A exits withstep-downsb outside the airplane typically have automatically deployed and erectedmeans of reaching the wing and ground.

Type Ia Floor level exit at least 24 inches wide by 48 inches high.

Type II Floor level exit at least 20 inches wide and 44 inches high. May also be locatedover the wing, with step-up inside the airplane of no more than 10 inches and step-down outside the airplane not exceeding 27 inches.

Type IIIa Rectangular opening at least 20 by 36 inches with step-up not to exceed 20 inches.Most often placed over the wing, having stepdown not exceeding 36 inches.

Type IVC Over-the-wing exit no less than 19 by 36 inches, with step-up of no more than 29inches and step-down no greater than 36 inches.

Taila Similar to the Type I exit in size, a ventral exit is a passage from the passengercompartment through the plane’s fuselage down a set of stairs to the ground. Tailcone exits lead directly out of the airplane’s tail onto an escape slide.

a Exit types most commonly used in large transport aircrafl.b ~feP-~uw is the ac~~ distance between the bottom of the required opening and a usable

foothold, extending out from the fuselage, that is large enough to be effective withoutsearching by sight or feel. Srep-up is the height from the floor of the cabin to the lower sillof the exit.

c Used in aircraft having fewer than 10 passenger seats.

Type A

Type I

7

Type II

Type Ill \

Type IV \

~

SOURCE: Office of Technology Assessment, 1993, based on 14 CFR 25.807; Daniel A. Johnson, Justin Case: A Passenger’s Guide to AirpZaneSafefy and Sumival, (New York, NY: Plenum Press, 1984), p. 148; and Mary Edwards and Elwyn Edwards, l%e Aircraj Cabin:Managing /he Human Factors (Hants, England: Gover Technical Publishing Co., 1990), p. 140.


puncture strength, radiant heat resistance, anddeployment as flotation platforms after ditch-ing .22

Training and OperationsFAA requires operating certificate holders

(airlines) to establish and maintain training pro-grams for each crew member. FAA also regu-lates cockpit crew hours but not flightattendants’ duty time. Activities required offlight attendants prior to takeoff include verify-ing that passengers’ seat belts are fastened,briefing passengers on emergency equipmentuse, and ensuring all galley items and carry-onluggage are securely stowed. Flight attendantsalso administer first aid and cope with other in-flight emergencies.

During flight attendant initial training, re-quired instruction topics include passengerhandling, cabin and galley equipment use, air-plane characteristics pertinent to in-flight emer-gency procedures, appropriate provisions of theFAR, and extensive emergency training. Re-current training includes a review of the crewmember’s state of knowledge of the airplaneand their duties, provides new instruction asnecessary in subjects required for initial groundtraining, and requires a competence check inassigned duties and responsibilities.23 Cabincrew members receive recurrent training every12 months .24

least one emergency evacuation drill.25 Duringinitial training and once each 24 months duringrecurrent training, crew members must performand observe additional emergency drills, Ingeneral, this is accomplished using cabinmockups, in which flight attendants and othercrew members operate exits and simulate thedeployment, inflation, and use of slides. Hands-on training with the slides is provided only ininitial training.26

NATIONAL TRANSPORTATIONSAFETY BOARD

Created in 1966 under the U.S. Departmentof Transportation, NTSB became an independ-ent executive branch agency in 1975. It investi-gates accidents27 for all transportation modes,including general aviation, selected public-useaircraft, and commercial transports; conductssafety studies; and issues recommendations forchanges in regulations and procedures. FAA isnot bound to accept NTSB regulatory changesuggestions .28

Aircraft operators must immediately notifyNTSB whenever an accident occurs or an air-craft evacuation involves use of an emergencyegress system.29 The information provided toNTSB must include the number of personsaboard the aircraft, and the number killed or

Along with instruction in procedures andequipment use, the emergency training mustprovide at least one firefighting drill and at

22 JOhn J. pe[rakis, Federal Aviation Administration,“FAA Occupant Protection and Cabin Safety Overview, ”Proceedings of the Flight Safety FoundationlFederalAviation Administration International Aircrafl OccupantSafety Conference and Workshop, DOTIFAAIOV-89-2(Washington, DC: U.S. Department of Transportation,August 1989), p. 47. See also Federal AviationAdministration, “Emergency Evacuation Slides, Ramps,and Slide/Raft Combinations, ” TSO C-69B, unpublishedre rt, Aug. 17, 1988.2~14 CFR 121.427; 35 Federal Register 90 (Jan. 3,1970).24 Flight deck crew training requirements are containedin 14 CFR 121, Subpart N.

25 The realism of the evacuation drills is of concern;e.g., United Airlines uses darkness in its flight attendanttraining. William Hathaway, U.S. Department of Trans-portation/Research and Special Programs Administration,Volpe National Transportation Systems Center, personalcommunication, Jan. 15, 1993.26 Noreene Koan, chairperson, National Air SafevCommittee, Association of Flight Attendants, persomlcommunication, Dec. 12$ 1992.27 An aircrafi accident is an occurrence associated withthe operation of an aircraft that takes place between thetime any person boards the aircraft with the intention offlight and all such persons have disembarked, and inwhich any person suffers death or serious injury, or inwhich the aircrafi receives substantial damage. incidentmeans an occurrence other than an accident, associatedwith the operation of an aircraft, that affects or couldaffect the safety of operations. 49 CFR 830.2.28 offjce of Technology Assessment, op. cit., footnote 8,

5953.49 CFR 830.5.

Chapter 1—Background and Regulatory Context ● 13

seriously injured .30 NTSB then assesses theaccident or incident and determines probablecause. However, NTSB is not required to keeptrack of the nature of passenger injuries (i.e.,whether the injuries occurred as a result of acollision or during evacuation from the air-crafit. 31 Because existing accident/incidentdatabases do not support assessment of theperformance of evacuation systems during ac-tual emergency conditions, this informationmust be painstakingly gleaned from investigatorreports.

Accident/Incident ReportsFAA’s Aviation Standards National Field

Office maintains a database of accidents andincidents officially reported to NTSB and re-ports filed by FAA field inspectors.32 NTSBadmits it does not collect all relevant data be-cause reporting requirements omit some typesof evacuations (i. e., those in which no seriousinjuries occurred). According to NTSB staff, asignificant number of occurrences are notmonitored because of a shortage of personnel,variability in reporting efforts, and an emphasison fatal accidents .33

The performance of evacuation systems hasnot been the focus of accident investigations.Reporting has improved over the years, accord-ing to Boeing staff, as investigators have begun

so Serfous ~~~u~ is defined as any injury: 1) requiringhospitalization for more than 48 hours, commencingwithin 7 days of receipt of the injury; 2) resulting infracture of any bone (except simple fractures of nose,fingers, or toes); 3) causing severe hemorrhages, nerve,muscle, or tendon damage; 4) involving any internalorgan; and 5) involving any second- or third-degree bums,or bums affecting more than 5 percent of the bodysurface.31 Matthew McCormick, chief, Survival FactorsDivision, and Stan Smith, chief, Data and AnalysisDivision, National Transportation Safety Board, personalcommunication, Dec. 17, 1991.32 U.S. Congress, Office of Technology Assessment,Transportation of Hazardous Materials, OTA-SET-304(Washington, DC: U.S. Government Printing Office, July1986), p. 71, available from OTA’s Science, Education,and Transportation Program.33 Nora Marshall, senior accident investigator, NationalTransportation Safety Board, personal communication,Jan. 8, 1992.

to pay more attention to crashworthiness aswell as airworthiness issues.34 A Boeing paperpresented at FAA’s 1985 technical workshopon evacuation safety cited a total of 583 knowninservice incidents in which aircraft wereevacuated. FAA neither maintains nor requiresmanufacturers to maintain records of evacu-ation-related injuries. According to safety inter-est groups, the manufacturers share this safetydata among themselves, but choose not to re-lease it to the public.35 Because the informationis proprietary, Boeing admits a reluctance toshare certification documents with pilot andflight attendant groups at the time FAA viewsthem.36

UNITED KINGDOM CIVIL AVIATIONAUTHORITY

Among other activities, CAA supports R&Drelated to cabin safety and evacuation. CAA’sprojects in the area of aircraft and safety regu-lation cover operational problems and airwor-thiness, including passenger survivability, andhuman factors in general .37

Currently, CAA efforts include:

determine the feasibility of developingcomputer models to assess seating con-figuration in relation to the number ofexits for both new aircraft and for aircraftoperating without the full complement ofexits available;develop models for predicting the behav-ior of fires in different aircraft cabinconfigurations; andassess the potential of cabin water spraysystems (CWSS) to extend evacuation

34 George VeVlog]ou, senior manager, 747/767 payloadSystems, Boeing Commercial Airplane Group, personalcommunication, Jan. 25, 1993.35 Mat~ew Finucane, director, Air Safety and Health,Association of Flight Attendants, personal communication,Dec. 18, 1991.36 VeV1oglou, op. cit., fmtnote 34.37 Civil Aviation Research and Development PfogrammeBoard, Prograrnme of Research and Development for CivilAviation Operations and National Air Traj?c Services,Issue 23 (Chehenham, England: Civil Aviation Authority,April 1992), p. iii.


time and save lives, and to study the feasibility of CWSS implementation ontransport aircraft.38

The United Kingdom’s Accident InvestigationBoard assumes many of the same responsibili-ties and investigation activities as NTSB.

38 Ibid., pp. 15-18.

Chapter 2

EVACUATIONDEMONSTRATIONS FOR

CERTIFICATION

CHAPTER 2

Evacuation Demonstrations for Certification

Beginning in 1965, the Federal AviationAdministration (FAA) required each air carrieroperating under Part 121 of the Federal Avia-tion Regulations to perform full-scale evacu-ation demonstrations under simulated emer-gency conditions prior to receiving operatingcertification for new aircraft or seating configu-rations. 1 The air carrier demonstration wasdesigned to evaluate crew training and the ade-quacy of evacuation procedures.2 FAA initiallyimposed a 120-second egress time limit forevacuating all passengers and crew. See box2-A on evacuation regulation chronology.

FAA attributed a 1967 change in maximumegress time to 90 seconds to advances in slidetechnology that had occurred since the initialstandard was released. g In 1982, after study ofactual and demonstrated emergency evacu-ations, FAA allowed certificate holders, underspecified conditions, to use the results of asuccessful demonstration conducted by eitherthe manufacturer or another airline rather thanconduct a new test.4

The stated goal of requiring the full-scaledemonstrations is to provide a benchmark bywhich FAA can consistently evaluate evacu-ation capability using various seating and exitconfigurations. 5 FAA claims that a consistentmeasure of success is achieved by requiring allmanufacturers to strive for the same 90-secondlimit. According to FAA, the demonstration. . . is not an acceptable evacuation perform-

1 29 Federa/ Register 18291 (WC. 24, 1964).2 An~ony J. Broderick, associate administrator for

regulation and certification, Federal Aviation Administra-tion, testimony at hearings before the House Committeeon Government Operations, Subcommittee on GovernmentActivities and Transportation, Apr. 11, 1991.3 31 F’edera/ Register 10276 (JuIY 29, 1966).4 Wa]ter S. Colemn, director, operations, Air TrMu+port Association, “Emergency Evacuations, CareerTraining, and Passenger Briefings, ” paper presented at theFAA Technical Conference on Emergency Evacuation ofTransport Category Airplanes, Sept. 3-6, 1985, Seattle,WA, p. 3.5 54 Federal Register 26692 (June 23, 1989).

ance standard.“G That is, manufacturers mustalso comply with specific equipment and mini-mum configuration requirements in addition tosuccessfully demonstrating complete evacuationwithin 90 seconds. Performance standards, onthe other hand, are expressed using objectiveperformance goals alone--no specific design oroperating criteria are established.

In addition to the 90-second time limit, FAAfull-scale evacuation demonstration criteria in-clude the following:

The demonstration must be conductedduring the6 dark of night or with the darkof night simulated--the airplane’s emer-gency lighting system can provide theonly illumination of exit paths and slides;A specified mix of passengers “in normalhealth” must be used--for example, atleast 30 percent must be females and atleast 5 percent must be over 60 years ofage;The passengers may not have participatedin a demonstration in the previous 6months; andNot more than 50 percent of the emer-gency exits may be used.

In a 1989 advisory circular (AC), FAA pro-vided guidance to manufacturers on how todetermine whether analysis and tests might beused in place of full-scale demonstration.7 TheAC also provided guidelines for set-up andconduct of the demonstration. Among otherthings, the AC identified two equivalent age-sex distributions, shown in table 2-1. Under the1989 FAA guidelines, manufacturers may re-place participants in the highest age category(i.e., the one most susceptible to injury) withgreater numbers of persons aged 51 to 60 yearsand need not use minors.

6 Ibid.

7 U*S, Department of Transportation, Federal AviationAdministration, Advisory Circular 25.803-1, Nov. 13,1989.

16

Chapter 2—Evacuation Demonstrations for Centification . 17

Box 2-A–Chronology of Changes to Evacuation Regulations

June 1965 Amendment 121-2 required all transport-category aircraft opera-

December 1978

January 1982

March 1990

tors to conduct demonstrations, to be completed in less than 120seconds, for all previously built and new aircraft.

October 1%7 Amendment 25-15 required manufacturers to conduct a 90-seconddemonstration, and required that aircraft be equipped with auto-matically deployed egress assist devices. 1

Amendment 121-30 revised the operators’ demonstration timelimit from 120 seconds to 90 seconds, and required retrofit ofautomatically deployed egress assist devices within 2 years for allpreviously built aircraft.

Amendments 25-46 and 121-149 revised requirements to permitmanufacturers and operators to concurrently demonstrate compli-ance with evacuation certification requirements, and allowedevacuation certification to be substantiated by a combination ofanalysis and tests at the discretion of the FAA Administrator.

Amendment 121-176 required, if an aircraft is certified to FAR25.803 per Amendment 25-46, the airline operator to demonstratecrew proficiency by showing that crew members can open half theexits and achieve usable slides within 15 seconds.

Amendment 121-124 established criteria for passengers seated inexit rows.

1 Egre~~ ~~l~t devi=~ ~1~ sIidcs, slide mft combinations, and overwing escape mutes. At ce~in efi~, slides mum ~ aum-matically erected as well.

KEY: FAR = Federal Aviation Regulations, Title 14, Chapter 1 of the CodC of Federal Regulations.

SOURCES: Federal Aviation Adminisuation Advisog Circuiar No. 25-803-1. Nov. 13, 1989; and Aviation Rulcmaking Advi-sory Cornmitte e, Emergency Evacuation Subcommittee, Performance Stamiards Working Group, “Emergency EvacuationRequirements and Methods ‘I%at Would Eliminate or Minimize the Potential for Injury to Full-Scate Evacuation DemonstrationParticipants, ” unpublished report, January 1993.

LIMITATIONS OF FULL-SCALE abilities to escape the aircraft. Participants inDEMONSTRATIONS

Full-scale demonstrations of evacuation sys-tems are both hazardous and costly. Intended toserve as a benchmark for functional ability ofemergency equipment and procedures, the testis not useful for system optimization.

The emergency evacuation scenario used infill-scale demonstrations does not representmost accident conditions, where impact forcesand fire effects frequently impair passengers’

demonstrations know they face no such dangerin their efforts to quickly exit the aircraft, sopanic is not present. However, the test stillexposes participants to a range of injuries, frombumps and bruises to serious, permanent in-jury. During seven full-scale demonstrationsconducted by manufacturers between 1972 and1980, 166 of 2,571 total participants receivedinjuries, or 6.5 percent. Of the 3,761 partici-pants in 12 demonstrations conducted between

18 . Aircraft Evacuation Testing: Research anti Technology Issues

Table 2-l-Equivalent Passenger Age-Sex Distributions for Evacuation Certification Participants,1989 Federal Aviation Administration Advisory Circular

Passenger distribution 1:

Age Percent of Percent oftotal females

21-50 80 30

51-59 15 40

>60 5 30

Passenger distribution 2:

Age Percent of Percent oftotal females

18-50 75 30

51-60 25 40

NOTE: Minors are precluded from participating in evacuation demonstrations under many state child laborlaws. Distribution 2 eliminates the need for participants older than age 60, who are most susceptibleto injury. In August 1993, relying on Civil Aeromedical Institute and industry data on the relativeevacuation rates of different age and sex mixtures, FM amended the age/sex distributionrequirement for evacuation demonstration participants as follows: (1) at least 40 percent of thepassenger load must be female; (2) at least 35 percent must be over 50 years of age; (3) at least 15percent must be female and over 50 years of age.

SOURCE: U.S. Department of Transportation, Federal Aviation Administration, Advisory Circular 25.803-1, Nov.13, 1989, and 58 Federal Register 45230 (Aug. 26, 1993).

Chapter 2-Evacuation Demonstrations for Certification . 19

1981 and 1991, 212 received injuries (5.6 per-cent),8

The cost of conducting full-scale evacuationdemonstrations, including test set-up, paymentsto volunteers, analysis, and so forth, reachesupward of $2 million for wide-body trans-ports. 9 The cost of evacuation demonstration isinsignificant compared to overall program andairplane construction; manufacturers assert it isthe hazard of serious injury, not test costs, thatgenerated their interest in modifying the exist-ing certification criteria and developing alter-native testing and assessment methods.

Since FAA first imposed the evacuation teston airlines and airframe manufacturers, therehave been only two major changes. First, im-provements in slide technology prompted FAAto reduce the maximum evacuation time in1967. Second, Federal and State occupationalsafety and healh laws proscribe the use of chil-dren under 18 years of age in the demonstra-tions. 10

As with other safety standards, the demon-stration for certification relates to a minimumlevel of safety; airline economics dictate thatmanufacturers strive for maximum seating ca-pability, not optimal safety for a given numberof passenger seats. Both cost and the potentialfor injury make manufacturers reluctant to con-duct any more than the minimum number oftests required of the industry.

The utility of FAA’s “benchmark” for evacu-ation capability hinges on the comparability oftest conditions and test results. The benchmarkenables FAA to determine only if an aircraft

8 Aviation Rulemaking Advisory Committee, Emer-gency Evacuation Subcommittee, Performance StandardsWorking Group, “Emergency Evacuation Requirementsand Compliance Methods That Would Eliminate orMinimize the Potential for Injury to Full-Scale EvacuationDemonstration Participants, ” unpublished report, January1993, p. 10.9 W e b s t e r C. Heath, managers Technical Liaison,Industry Regulatory Affairs, Douglas Aircraft Co., per-sonal communication, July 8, 1992.10 Wil]lam H. Shook, senior principal technical special-ist, Douglas Aircraft Co., personal communication, Dec.16, 1992.

achieves the same minimum level of perform-ance as other aircraft before it; the benchmarkdoes not permit quantitative assessment ofoverall safety or the relative performance ofelements within the aircraft’s evacuation sys-tem. The subjective nature of some of the testcriteria (e.g., the maximum level of illumina-tion possible to simulate the dark of night) in-troduces variability. Controlling variability is akey factor in the statistical validity of any test,as discussed below.

Test and Data Validity

In order to assess the validity of a test or itsdata, one must judge both the quality of the testprocedure and the measurement methodology.The identification of major variables and howthey affect the outcome of a test lends credibil-ity to the process, as does the repeatability ofresults. Achieving consistent test conditions isfundamental to limiting variability. FAA andindustry use the benchmark of 90 seconds togauge whether different cabin seating and exitconfigurations provide a minimum level ofaircraft evacuation safety. Human performance,a dominant variable in successful evacuationsunder real or imagined emergency situations, isnot easily controlled. The following factorsmay greatly affect the outcome of an actualemergency evacuation performance:

cabin and flight crew capabilities (e.g.,training, experience, and physical/mentalcondition);aircraft integrity and evacuation tech-nologies;passenger demographics, percent of seatsoccupied, and amount and mix of carry-on luggage;ambient lighting; and,actual accident conditions.

One potential problem with the test procedureis that the mix of test participants required byFAA is often not representative of the flyingpublic on a given flight. In general, passengerdemographics vary from region to region andseasonally. Tests conducted using passengerloads with higher percentages of women and


elderly persons, or with children and personswith disabilities, would likely generate longeraverage evacuation times. See box 2-B on FAAtests with persons with disabilities and figureA-1 in the appendix.

An unrealistic passenger mix, combined withthe absence of surprise, trauma, fright, andpanic, produces optimistic indications of anaircraft’s evacuation safety capability.11 How-ever, industry and many others are understand-ably loathe to subject demonstration partici-pants to the presence of fire, smoke, andadditional debris, for fear of increasing thelikelihood of injury. On the other hand, anychanges to the certification process designed toreduce the risk of injury require analysis of thecomparability of results.

In addition, without the benefit of repeatedtrials, one cannot be confident that a singlecertification test result truly represents an air-craft evacuation system’s capability. Neither amargin of error or confidence level can be de-termined (see figure A-2 in the appendix). Bycomparison, use of anthropomorphic dummiesallows auto manufacturers to conduct realisticcrash response tests repeatedly and with highvalidity, without threat to human safety, and todetermine performance relative to governmentstandards. 12 FAA and the aviation communitystruggle to achieve agreement on whether thevalue of but one full-scale evacuation demon-stration for certification warrants the risk. Aformal vehicle for this discussion was theEmergency Evacuation Subcommittee estab-lished by the FAA Aviation Rulemaking Advi-sory Committee (ARAC).13 The followingsection describes the subcommittee’s progress

11 SM. Vanstone, Vice chairman, Aircraft Designs andOperations Committee, International Federation of AirLine Pilots Association (IFALPA), “Emergency Evacu-ation and Cabin Safety, ” paper presented at the FAA Pub-lic Technical Conference on Emergency Evacuation ofTransport Airplanes, Sept. 3-6, 1985, Seattle, WA, p. 1.12 Jeffry H< Marcus, manager, ProtectIon and SumivalLaboratory, Civil Aeromedical Institute, personal com-munication, Jan. 13, 1992.13 Created February 5, 1991, AIU4C is comprised ofFAA officials and representatives from 58 aviationgroups.

and potential changes to the demonstration re-quirement.

DEVELOPMENT OF ALTERNATIVESIn February 1991, ARAC formed a sub-

committee to address a slate of regulatoryreforms in the evacuation area--reforms rec-ommended during the conferences and work-shops of the mid-1980s--and charged it withgiving advice and recommendations to the FAAFlight Standards and Aircraft Certification of-fices on regulatory standards for evacuation andpassenger safety. In turn, the subcommitteechartered a working group to address thepotential for using performance standards inplace of or in addition to design criteria forcertification. 14 The Performance StandardsWorking Group (PSWG--members are drawnfrom the various elements of the aviation com-munity) is charged with making a recommen-dation concerning whether new or revised stan-dards for emergency evacuation can and shouldbe stated in terms of safety performance ratherthan as specific design requirements. Theworking group must consider two questions:

Can standards stated in terms of safetyperformance replace, supplement, or bean alternative to any or all of the currentcombination of design and performancestandards that now address emergencyevacuation found in Federal AviationRegulations Parts 25 and 121?If a performance standard is recom-mended, how can FAA evaluate a minorchange to an approved configuration, ora new configuration that differs in eithera minor or a major way from an ap-proved configuration?

November 1991, PSWG expanded itsmission to include making a recommendation tothe Emergency Evacuation Subcommittee con-

14 AS pan of the 1993 renewal of ARAC’S charter, tiesubcommittees were redesigned as interest areas (e. g.,Emergency Evacuation Issues) and working groups nowreport directly to ARAC. Steve Erickson, assistant chair,Aviation Rulemaking Advisory Committee, EmergencyEvacuation Issues, personal communication, Aug. 16,1993.

Chapter 2—Evacuation Demonstrations for Certification . 21

Box 2-B–Evacuation of Persons With Disabilities

Because the need for assistance in emergency situations has limited the access of nonambulatorypersons to commercial air transportation, in the early 1970s, the Federal Aviation Administration (FAA)commissioned the Civil Aeromedical Institute (CAMI) to study aircraft evacuation using passengers withdisabilities. 1

CAMI testing showed that, when occupying window seats, passengers with disabilities spent 50percent of the total time required for egress in moving from their seats to the aisle. The data suggestedthat persons with disabilities should be seated along the aisle. However, this may compromise the safetyof the passengers in the outboard seats. CAMI’s evacuation trials also showed that total evacuation timeswere shorter when nonambulatory passengers were seated away from the exits. Other observations fromthe study included:

● Aide width and seat row pitch affect the ability ofother passengers to assist nonambulatory persons.● Passengers with disabilities may need to be reoriented before entering the slides.

The desire to ensure accessibility to all forms of transportation led to a 1982 Civil Aeronautics Boardruling that all passengers, regardless of impairment, should be given reasonable access to air travel andthe opportunity to use ordinary, unaltered airline services.2 While it may be technically feasible toderive optimum seating conjurations for different percentages of passengers with disabilities, politicaland ethical considerations likely preclude the implementation of any such plans.

Rule changes adopted in 1991, and revised in 1992, limit seating adjacent to exits to those passengerswho are proficient in the English language and do not have mobility, sensory (e.g., hearing and vision),or cognitive (e.g., schizophrenia) impairments.

1 JCG. BMMOW et al., Civil Aeromedical Institute, Emergency Escape of Handicapped Air Travelers, FAA-AM-77-11(Washington, DC: Federal Aviation Administration, July 19T7), pp. 1-2.

2 ~aq ~ward~ ~ ~~n Edwards , ?%e AUC@ ~Ln; Mmaging the Human Ftmrors (Hants, England: CloverTechnical Publishing Co., 1990), pp. 45-46.

cerning new or revised emergency evacuation with the working group’s report, describedrequirements and compliance methods thatwould eliminate or minimize the potential forinjury to fill-scale demonstration participants.PSWG released its report on methods of reduc-ing risk of injury to participants in emergencyevacuation demonstrations for certification inJanuary 1993. Table 2-2 lists the workinggroup’s conclusions and recommendations.

Despite months of effort to reach consensus,the report failed to satisfy the group as awhole. 15 Three letters of dissent, submitted

dissatisfaction with the process and report con-clusions. Key concerns were the perceived fail-ure of the group to “. . . undertake a systematicanalysis of the procedures used in conductingfull scale evacuation demonstrations, ” and theloss of valuable crew performance informationincurred by eliminating the requirement forfull-scale demonstrations.16 The Air Line PilotsAssociation expressed concern that the absence

15 ARAC was intended to speed the rulemaking processby including constituents at the front end of regulationdevelopment (i.e., before notice of proposed rulemakingand request for comments are released). However, thelength of time required by the Performance StandardsWorking Group to address its first mission caused some

concern on the part of the subcommittee’s chairman andmembers of Congress that the process is itself unwieldy.16 Association of Flight Attendants “Comments on per-formance Standards Working Group Report, ” unpublishedreport, Jan. 15, 1993, p. 1.


Table 2-2-Conclusions and Recommendations of the Performance StandardsWorking Group, 1992 Report

Conclusions:

●

●

●

●

●

The nature of the full-scale evacuation demonstration, as currently defined in FAR 25.803(c), issuch that injuries can occur.

The full-scale evacuation demonstration can be a useful tool for comparing the evacuationcapability of a new, unique airplane configuration with the current FAR 25.803 standard.

The full-scale evacuation demonstration test conditions can and should be revised to minimize thepotential for injuries to test participants.

Steps must be taken to ensure that testing with humans is strictly limited and controlled. Onlyafter all alternative means of obtaining necessary data have been deliberated should limitedexposure of test subjects to the evacuation demonstration test conditions of FAR 25.803(c) beconsidered.

Full-scale evacuation demonstrations should be conducted for only those airplane configurationswhere regulatory authority-approved test data are not available to support analysis.

Research recommendations:

CAMI, or another source FAA deems appropriate, carry out a study to determine which age andsex group(s) are least susceptible to injury and develop an appropriate age and sex mix for full-scale demonstration tests while maintaining the validity of the 90-second criterion.

Initiate a research program to develop a new, two-part emergency evacuation test protocol forescape slide testing and airplane flow rate tests without the use of escape slides.

Institute a high-priority research and development program to develop long-term revisions to theevacuation demonstration test protocol so as to further minimize injuries to test participants.

Develop a system or process for FAA to collect data on injuries sustained during emergencyevacuation demonstration testing.

Establish an FAA escape slide research and development program designed to further minimizeinjuries.

KEY: CAMI = FAA’s Civil Aeromedical Institute; FAA= Federal Aviation Administration; FAR = FederalAviation Regulations, Title 14, Chapter I of the Code of Federal Regulations.

SOURCE: Aviation Rulemaking Advisory Committee, Emergency Evacuation Subcommittee, PerformanceStandards Working Group, “Emergency Evacuation Requirements and Compliance Methods thatWould Eliminate or Minimize the Potential for Injury to Full Scale Evacuation DemonstrationParticipants,” unpublished report, January 1993.

Chapter 2-Evacuation Demonstrations for Certification . 23

of data on injury rates for alternatives to thetwo extremes of certification (analysis only orfill-scale demonstrations) made the workinggroup report biased toward FAA approvalbased solely on analysis. 17

AnalysisInjuries sustained over the years by demon-

stration participants became the basis for the1978 rule change18 providing that the demon-stration requirement may be waived if theAdministrator finds that a combination ofanalysis and component testing19 will providedata equivalent to that obtainable through full-scale demonstration.20 In 1982, Boeing Aircraftpresented to FAA an analysis approach thatrelied on a timeline summation of evacuationactivities from exit preparation to the arrival ofthe last evacuee on the ground. All segments ofthe timeline were derived using data fromFAA-witnessed tests and tests verifiable fromvideo or film records .21 Boeing’s modelapproach is outlined in figure A-3. The time ofexit flow is equal to the time elapsed betweenthe first evacuee and last evacuee reaching theground; this time is a function of the anticipatednumber of passengers and crew and the flowrate permitted by exits.22 Critical to the analysis(and evacuation performance) is the balancedloading of passengers with respect to exit sizeand location. To address the issue of passengermanagement (flow control), Boeing includes adiscussion of passenger distribution; exit per-formance capability, both preparation and

17 Rj~ky R. Davjdson, chairman, Air Line pllOtS Asso-

ciation, Accident Survival Committee, letter to JayAnema, chairman, Performance Standards WorkingGroup, Jan. 19, 1993.18 See FAR amendments 2546 and 121-149.19 Compnent tests and partial demonstrations examinethe performance of isolated elements within the evacuation

%st~?orge Veryloglou, engineer, Airframe SystemsTechnology, Boeing Commercial Airplane Group,“Emergency Evacuation System Certification via Analysisand Tests, ” paper presented at the FAA Technical Confer-ence on Emergency Evacuation of Transport Airplanes,Sept. 3-6, 1985, Seattle, WA, p. 5.21 Ibid., p. 10.22 Ibid., p. 11.

egress; and crew member performance ele-ments (e.g., time of travel to duty position).23

The Douglas Aircraft Company adopted asimilar approach for predicting evacuation per-formance, using data from prior demonstrationsand component tests, along with “industry-accepted averages, ” to estimate total evacuationtime. Douglas Aircraft Company staff believethe analytical model is more credible than afull-scale demonstration, which is affected bynumerous human factors.24 Industry in generalsupports testing of component performance,emergency procedures, and crew training toavoid exposing crew and demonstration volun-teers to the risk of injury, but there is somepolitical sensitivity to certification by analysis,as discussed in box 2-C.

Demonstrations With PlatformsOne suggestion for reducing the likelihood of

injury to demonstration participants entails re-placing the slides with level platforms or gentlysloped ramps. Slide performance data wouldthus be obtained with more controlled demon-strations that present fewer risks to participants.

On December 11 and 12, 1992, for its secondattempt to certificate the MD-11 for 410 pas-senger seats, McDonnell Douglas adopted sucha phased approach.25 McDonnell Douglas firstdeveloped the analytic methodology to equatethe existing 90-second test with slides to aramp-based test of an unknown time limit. Tofill in data gaps, McDonnell Douglas conductedcomponent tests to establish average openingtimes for doors with and without slides, and theflow rates (passengers per minute) throughdoors without slides.

McDonnell Douglas completed 10 tests with100 persons each to establish rates for Type A

23 ~id., pp. 11-12.24 Douglm Aircraft Co., “MD-1 1 Evacuation Demon-stration: Analysis and Changes Overview With AnalyticModel, ” paper submitted to the Federal Aviation Admini-stration, n.d., p. 6.25 According to McDonnell Doughs staff, the CaliforniaOccupatioml Safety and Health Agency would not allowthe manufacturer to repeat the full-scale demonstrationwith slides in total darkness.


Box 2-C–Political Sensitivity to Use of Analysis in Evacuation Certification

In 1984, Boeing proposed to deactivate one of five pairs of overwing exits on inservice passenger747s.1 Maximum passenger density would be reduced to 440 from 550, commensurate with the newnumber of Type A exit pairs. 2 However, the distance between doors would exceed 60 feet. (Existingregulations did not specify the maximum distance between exit doors.) The Federal Aviation

. Administration (FAA) Transport Aircraft Certification Office (Northwest Mountain Region) approvedBoeing’s request based on analysis.

Flight attendant unions protested the decisions and certification process, and called on Congress tointervene on grounds of diminished safety.3 A June 1985 hearing conducted by the House Committeeon Public Works, Subcommittee on investigations and Oversight brought public attention to both thepotential impact of allowing large distances between exits and the unscrutinized process in which thedeactivation was approved. At the hearing, FAA Administrator Donald Engen announced his disap-proval of sealing off the overwing exits. Subsequently, Admiral Engen appointed an EmergencyEvacuation Task Force toexamine the issue and reassess related emergency evacuation regulations.4

In October 1987, FAA published a notice of proposed rulemaking relating to new standard limitson transport category airplanes for the distance between any passenger seat and the nearest exit andthe distance between exits.5 Under the rule, type certification for the new 747-400 with only eightexits would not be approved, and operation within the United States of oreign-owned 747s havingeight exits would not be allowed.6 In 1989, FAA issued a final rule prohibiting airplane manufacturersand air carriers from increasing the distance between emergency exits beyond 60 feet.7 Boeingmaintains the rule was specifically applied to the 747 but not the Lockheed L-1011, which also haddistances greater than 60 feet between exits.8 Mathematical analysis of evacuation times for thedifferent configurations (i.e., 440 passenger seats with 8 exits or 550 passenger seats with 10 exits)would yield the same results because flow rates and door opening times were insensitive to variationsin internal configurations.

1 At ~ time, Boeing offe~ the 747 in various configurations, including a passenger model with 10 Type A ~in deckexits; convcrnbte and combi 747s wirh 10, 8, or 6 main deck Type A exits; and the special performance 747, with 8 such exits.George Veryiglou, senior manager, 747/767 Payload Systems, Boeing Commercial Airplane C3roup, personal cornmunicaion,Jan. 25, 1993.

2 14 ~ ~.307 ~tes CA Type A exit IM& at 110 @*ngers.3 U.S. Cmgmss, OffEe of Technology Assessment, S@e Skies for Tomorrow: A~’~”~ &@Y in a ~efitivc

Environrnenr, OTA-SET-381 (Wash@tom DC: U.S. Oovenunent Printing Office, July 1988), p. 57, available from OTA’sScience, Education, and Transportation Program.

4 ~oti Imus, dlr~r of ~@f for Co~~s~ James L, Oberstar. “Keynote Address.” Procectigs of ‘he HigtiS@ety Foundan” onlFederal Aviation Administration Internadonal Aircraft Occupan! Safety Conference and Workshop,DOT/FAA/OV-89-2 (Washington, DC: U.S. Department of Transportation, August 1989), p. 12.

5 NpW 87-10,52 Federal Reg&er 39190 (OCL ~, 1987).6 Off&of T~~lo~ AwMme~, Op. Cit., foomote 3} P. 57”7 ,4 cn 121.310, ~~cnt 121-205, M FederaZRegLrfer 26696 (June 23! 19$9)*8 Vqloglmo oP. c ‘it. foomote 1. The L-1OI 1 is still operated under Part 121 with these distances.

and Type I doors. Based on the test results, staff concluded that the cabin could be effec-McDonnell Douglas proposed a maximum time tively managed with 9 (the minimum numberlimit of 62 seconds for the modified certifica- required by FAA) instead of 10 flight atten-tion demonstration.26 Additionally, after three dants.evacuations using different procedures and The evacuation test was completed in 56 sec-flight attendant stations, McDonnell Douglas ends; a time margin analysis like that espoused

by the Working Group for future certifications26 Shmk, Op. cit., footnote 10. by analysis yielded 51 seconds, well above the

Chapter2-Evacuation Demonstrations for Certification . 25

10 percent factor.27 The entire testing programresulted in only four minima1 injuries, althoughpast experience suggested one or two fractureswould occur.28 FAA held the test to be suffi-cient for certification. Boeing and Airbus willlikely adopt use of component tests and analysiswhen possible. However, this approach tellslittle about the system effects of new slide con-figurations, a major factor in evacuation per-formance and one that has often changed.29

Limitations to AnalysisThe analytical models provide only estimates

of flow rates under ideal conditions; the modelsdo not take into account the effects of passen-ger motivation or the presence of fire, smoke,and injuries. The results of the October 1991evacuation test for the MD-1 1 evacuation certi-fication, in which test conditions were appre-ciably harsher,30 illustrate this limitation. Inaddition, flow control is difficult to analyzemathematically because the calculations are in-sensitive to architectural changes within thecabin or differences in passengers’ decision-making abilities.31 Another concern over rely-ing on analysis and component tests for certifi-cation is that, without full-scale demonstrations,it will be difficult to acquire information onpassenger management strategies.

Industry asserts that its mathematical analysismethods are valid and that demonstrationsusing volunteers are no longer necessary.Boeing and McDonnell Douglas provided OTA

27 me time margin ~lysis sums over al] exits the dif-

ference between maximum allowable egress time (e.g., 90seconds) and that achieved during the demonstration. ThePSWG-recommended margin of 10 percent of the maxi-mum equaled 6.2 seconds for the December 12, 1992,McDonnell Douglas evacuation test.28 Shmk, op. cit., footnote 10.29 George Veviogiou, senior manager, 747/767 payloadSystems, Boeing Commercial Airplane Group, personalcommunication, Dec. 14, 1992.30 FAA intevre~tion of the simulated dark of night re-quirement resulted in a pitch-black environment outsidethe aircraft; even the light from video monitors used fordata collection was shielded from passengers’ view.Additionally, a combimtion of cabin crews from differentcountries was used, contributing to poor coordimtion offlight attendant actions.31 Shmk, op. cit., footnote 10.

incomplete data to perform a statistical analysisof the parameters used in their models.

Other AlternativesIn addition to the combination of analysis and

component testing, the use of “professional”demonstration participants has been suggested(i.e., reduce the chance of injury by replacingthe “naive” volunteers required for full-scaledemonstrations with trained professionals).

The Civil Aeromedical Institute employs twodifferent test protocols for its evacuation stud-ies. In the first, participants repeat evacuationdrills several times to gain experience beforeexperimental variables are changed. Experi-mentation begins after no significant differencein evacuation times is reached. The secondprotocol entails exposing participants to thesame combination of experimental variables indifferent orders to average the experience fac-tor between subjects32 (Latin square or coun-terbalanced experiment design).

Begging the question of whether or not thecertification test represents reality, a “. . . sys-tems test with naive subjects allows the evalu-ation of design factors such as cabin lighting,tactile clues for exit locations, etc., whose in-fluence would be lost or minimized with expe-rienced test subjects. “33 Tests with young par-ticipants with similar athletic abilities couldminimize the risk of injury but provide onlyoptimistic estimates of evacuation performance.The comparability of test results with those ofearlier demonstrations would be suspect.

Continuing research and technology devel-opment have been integral to improving theoverall evacuation capability of an aircraft aswell as developing new methods of assessingevacuation performance for certification pur-poses. The next chapter describes the majorresearch and technology issues and programsrelated to evacuation performance.

32 Marc~, op. cit., fOOtnOte 12”

33 Ibid.

Chapter 3

RESEARCH ANDTECHNOLOGIES FOR

EVACUATION SYSTEMS

CHAPTER 3

Research and Technologies for Evacuation Systems

The aircraft evacuation system has three keyelements: exits and slides, efficient means ofreaching the exits, and the crew and passengerswho use them. To be able to leave one’s seat,move toward an exit door or hatch, and escapefrom the aircraft depends on the passenger’sphysical and mental condition, and tolerance tocrash and fire hazards. These hazards, in turn,depend on the strength of seat attachments andrestraints, airframe energy absorption, and thefire resistance of the cabin lining and seatingmaterials.

Evacuation performance thus requires en-hanced cabin safety to preclude incapacitationfrom impact, smoke, heat, and toxic gases be-fore egress can be achieved. Evacuation per-formance also depends on the design andoperation of emergency equipment and flightattendant training. Cabin safety research andevacuation testing are essential elements of anyeffort to assess and improve evacuation safety.

CABIN SAFETY RESEARCH ANDTECHNOLOGIES

The Federal Aviation Administration (FAA)researches and regulates several facets of cabinsafety for transport airplanes, rotorcraft, andgeneral aviation aircraft. The majority of theresearch and testing is accomplished at theTechnical Center and the Civil AeromedicalInstitute (CAMI). FAA also relies on theNational Aeronautics and Space Administration

.“ and the National Institute of Standards andTechnology (NIST) for contract or cooperativework in crashworthiness and fire safety, re-spectively. In passenger transport, after theUnited States, the United Kingdom is the sec-ond largest contributor to cabin safety researchand technology (R&T). Other foreign investiga-tors in fire safety research include Canada,Germany, the Nordic countries, Japan, andAustralia. 1

1 Richard B~kowski, senior researcher, Building andFire Research Laboratory, Natioml Institute of Standardsand Technology, persoml communication, Apr. 8, 1992.

In 1980, FAA’s Special Aviation Fire andExplosion Reduction (SAFER) AdvisoryCommittee published several recommendationsto improve fire safety and survivability.2 FAAused the committee’s recommendations to di-rect its research and development (R&D)efforts, and produced new and modified regu-lations in a number of areas.3 The success ofFAA’s programs rests primarily on the devel-opment of representative fire scenarios and testmethods. Currently, research is concentrated intwo categories: in-flight fires, where safety ismeasured by the ability to prevent, detect, andcontain a fire in the immediate vicinity of igni-tion as well as discriminate from false alarms;and postcrash fires, which in turn involve eithermaking the environment inhabitable for alonger time or evacuating passengers morequickly. 4 The key programs relating to cabinmaterials, emergency equipment, and trainingare discussed below.

Cabin MaterialsAccording to FAA, the most important recent

improvement in cabin safety was the addition offire-blocking layers to seat cushions.5 FAA,with NIST participation, established in the mid-1980s the methodology for determining the rateat which hot gases are emitted from burningseat cushions. The fire blocking has beenshown to extend evacuation and survival timeby at least 40 seconds in one representative

2 Federal Aviation Administration, Find Repoti of theSpecial Aviation Fire and Explosion Reduction (SAFER)Advisory Committee, vol. 1, Report FAA-ASF-80-4

!?’ashington, DC: June 26, 1980).R.G. Hill et al., “Aircraft Interior Panel Test Criteria

Derived From Full-Scale Fire Tests, ” DOT/FAA/CT-85/23 (Atlantic City, NJ: Federal Aviation Administra-tion Technical Center, Septemkw 1985), p. 1.4 Constintlne p. Sarkos, manager, Fire Safety Branch!

FAA Technical Center, personal communication, Apr. 22,1992.5 U.S. Depafiment of Transportation, Federal Aviation

Administration, Aircraft Safety Research Plan (AtlanticCity, NJ: Federal Aviation Administration TechnicalCenter, November 1991), p. 123.

27


postcrash fire scenario by delaying the onset ofmaterial ignition and reducing the spread offlames and toxic products of combustion.6

The FAA Technical Center developed thestandard test protocol for assessing cabin mate-rial flammability through comparison of labora-tory studies and fill-scale fire testing (using areconfigured C-133 fuselage). In simulatedpostcrash fires, evaluation of combustion gasand temperature profiles indicated that the oc-currence of cabin flashover7 dictated surviv-ability, and that flashover can be bestcharacterized by heat release levels.8 Thisprompted the development of the current heatrelease standard instead of limits on specificcombustion products.9 Today FAA continues toinvestigate fire behavior, smoke toxicity, thebehavior of composite materials, and the effec-tiveness of potential safety improvements usingthe FAA Technical Center’s DC-10 and B-707test craft. 10

CAMI has extensively studied the effects offire on aircraft interiors, supporting rulemakingfor crew member protective breathing equip-ment (PBE). Continuing fire safety researchtopics include smoke release and relative toxic-ity of materials used in cabin finishings, andmethods to improve evacuation under toxicsmoke conditions.

Over the years, FAA’s Technical Centercontracted out portions of its materials safety,

6 Job J , Petrakis, “FAA Occupant Protection andCabin Safety Overview, ” in Proceedings of the FlightSafety FoundationlFederal Aviation AdministrationInternational Aircrt@ Occupant Sa$ety Conference andWorkshop, DOTIFAAIOV-89-2 (Washington, DC: U.S.Department of Transportation, August 1989), p. 43.7 F]ashov~r is the sudden, rapid, and uncontrolledgrowth of fire throughout the cabin, generating high tem-peratures and toxic gases and robbing the cabin atmos-

!here of oxygen.

Constantine P. Sarkos, manager, Fire Safety Branch,FAA Technical Center, personal communication, Jan. 15,1993.9 Also known as the 65/65 rule, which refers to themaximum allowable rate of heat release, in kW/m2, andthe total heat reiease, in kW-min/m2, under specified testcriteria.10 Alan S. Brown, “Fire Rule Changes Aircraft Mate-rials Mix, ” Aerospace America, March 1991, pp. 20-24.

fire performance, and toxicology research toNIST. NIST conducts in-house research at theBuilding and Fire Research Laboratory andfunds additional research through its UniversityGrants Program.11 According to NIST staff,recent gains in scientific knowledge and theadvent of measurement technology will shiftfire safety regulation toward performance stan-dards rather than design criteria.

The measurement technology required forquantitatively assessing evacuation system per-formance, including human factors, has notbeen developed to the same degree. The Avia-tion Rulemaking Advisory Committee efforts toreplace evacuation design criteria with per-formance standards suffer from the lack ofsophisticated analytic tools and human per-formance data.

Emergency EquipmentAnalysis of the 1985 Manchester aborted

takeoff and subsequent fuel-fed fire promptedseveral recommended design changes, includ-ing improved access to overwing exits andcabin interior hardening, most of which havebeen implemented.12 The accident also renewedinterest in cabin water sprays and passengerprotective breathing equipment. The relativemerits and disadvantages of these proposals arediscussed below, along with the topic ofrisk/risk assessment.

Protective Breathing EquipmentTime and the thermo-toxic environment are

two critical aspects of survival in aircraft acci-dents involving fire.13 Based on R&D done atCAMI, criteria for PBE for air transport crew

11 Bukowski, op. cit., fOOtnOte 1.12 Arthur Reed, “Technology Safety . . . For Cabin-FireSurvival, ” Air Transport Worki, October 1991, pp. 101-106.13 Garnet A. Ivfch et al., Civil Aeromedical Author-ity, l%e Eflects of Wearing Passenger Protective Breath-ing Equipment on Evacuation Times Through Type III andType IV Emergency Aircraft fiits in Clear Air and Smoke,Final Repoti, DOT/FAA/AM-89/12 (Washington, DC:U.S. Department of Transportation, November 1989), p.1

Chapter 3—Research and Technologies for Evacuation Systems . 29

members were issued in June 1983.14 Consist-ing of a full-face oxygen mask or combinationsmoke goggles and oxygen mask, crew memberPBE is required equipment for all aircraft op-erating under 14 CFR 121.

Although the investigation of the Manchesteraccident resulted in a recommendation for pro-vision of passenger PBE, or smokehoods, rulesmandating their installation on transport aircrafthave not been issued.15 Two general types ofsmokehoods, filter and oxygen-generating,have been proposed. The lightweight filter typeis susceptible to carbon monoxide contamina-tion and becomes ineffective when cabin oxy-gen is depleted. Either type can delay evacu-ation because passengers stop moving towardthe exits to don the masks. Smokehoods canalso impede egress through smaller doors, pre-vent passengers from hearing crew instructions,and reduce vision. 16

In addition, while the Civil Aviation Author-ity (CAA) of the United Kingdom issued a draftspecification for passenger smokehoods in1986, it rejected requiring smokehood equip-ment after a joint review of regulatory policyby U. K., U. S., French, and Canadian authori-ties showed that the implementation of othersafety measures (e.g., seat fire blocking andcabin material improvements) has improvedsurvivability to the extent that smokehoodshave become less useful .17 Because the timeavailable to evacuate an aircraft is the mostcritical element of survival, the additional time

14 Federal Aviation Administration Technical StandardOrder, TSO-C99, June 27, 1983.15 The aviation industry first concentrated On the role ofsmoke and toxic gases in hindering evacuations after theNovember 1965 aircraft accident at Salt Lake City, Utah.FAA published a summary of CAMI studies related topassenger smokehoods performed during the period ofNovember 1965 to February 1987. See E.A. Higgins,Summary Report of the History and Events Pertinent to theCivil Aeromedical Imrtitute’s Evaluation of ProvidingSmoke/Fume Protective Breathing Equipmenl for AirlinePassenger Use, DOT/FAA/AM-8715 (Washington, DC:U.S. Department of Transportation, Federal AviationAdministration, June 1987).16 Helen Gavaghan, “Aircraft Fires: Living Through theSmoke, ” New Scientist, Aug. 6, 1987, pp. 54-57.17 Reed, op. cit., fOO~Ote 12.

spent donning smokehoods during the periodwhen conditions permit the fastest egress re-duces their potential to save lives and may evenresult in more deaths. 18

Water SprayFAA commissioned an early cost/benefit

study of fire management systems and safetyimprovements, completed in 1983. CAA re-viewed worldwide accidents involving fire-re-lated deaths over the 1966 to 1985 period, andconcluded that the benefit attributable to havingan onboard cabin fire suppression capability(e.g., a water spray system) is likely to be sub-stantial and exceeds the benefit attributable tosystems that do nothing to delay the onset orprogress of fire.19 In June 1989, FAA beganworking with CAA and Transport Canada todevelop and evaluate a cabin water spray sys-tem (CWSS).20

The present heat release standard has driventechnology to the point where it is unlikely thatfurther cabin materials research and improve-ments over the near term will lead to apprecia-ble delays of flashover.21 Water spray worksindependently of fire origin and has morepotential to delay flashover under a variety offire scenarios; its benefits include cooler cabintemperatures, suppressed ignition of cabin ma-terials and delay of flashover, absorption ofcombustion gases, and the washout of smokeparticles.22 Full-scale tests of one cabin sprin-

18 ~ulse Speltel and Richard G. Hill, St@ of Be~fits

of Passenger Protective Breathing Equipment FromAnaiysis of Past Accidents, Final Report, DOTIFAAICT-88/03 (Atlantic City, NJ: FAA Technical Center, March1988), p. 4.19 Lionel Virr, “The Feasibility of Improved Fire Pro-tection Systems for Aircraft Occupants, ” in Proceedings ofthe Flight Sa$ety Foundation/Federal Aviation Admini-stration International Aircraft Occupant Safety Conferenceand Workshop, DOTIFAAIOV-89-2 (Washington, DC:U.S. Department of Transportation, August 1989), p. 200.20 Water spray is Urgeted because the use of foams andHalon-based suppression systems present health and/orenvironmental obstacles.21 Sarkos, op. cit., footnote 8; and Kent pofler~ con-

tributing engineer, New Large Airplane Payload Systems,Boeing Commercial Airplane Group, personal communi-cation, Aug. 11, 1993.22 Sarkos, op. cit., fOOtnOte 4“


kler concept (the SAVE system) have suggestedthat survival times can be extended by 2 to 3minutes.23 By comparison, fire blocking andimproved cabin interior materials provided ex-tensions on the order of 40 seconds and 17seconds, respectively .24

The possibility of inadvertent system dis-charge during flight, the weight/cost of systemimplementation, and reduced visibility duringevacuation are key drawbacks. At FAA’s re-quest, manufacturers participated in a disbenefitstudy (i.e., estimating the consequences of bothcommanded and accidental use). Estimatedweight penalties for narrow- and wide-bodyaircraft were on the order of 600 and 2,100pounds, respectively.25 Boeing estimated thecosts of installing SAVE CWSS to be approxi-mately $800,000 for a 757 airplane, and nearly$1.7 million for the newest model 747.26 Es-timated costs of retrofitting the world’s fleet ofcurrent production aircraft exceeded $6 bil-lion.27

Recognizing that these penalties and risksmust be reduced before system implementationis feasible, FAA has explored zoned use of thesprinklers, or spraying water only in the im-mediate vicinity of the fire, to decrease theamount of water required. Full-scale effective-ness tests with the zoned CWSS showed that,along with improved visibility, temperature and

23 Richard G. Hill et al., “Evaluation and Optimizationof an On-Board Water Spray Fire Suppression System inAircraft, ” paper presented at the Twelfth Meeting ofUnited States-Japan Panel on Fire Research and Safety,Tsukuba and Tokyo, Japan, Oct. 27-Nov. 2, 1992, p. 1.24 Petrakis, op. cit., footnote 6, p. 43; and U.S. GeneralAccounting OffIce, Aviation Safety: Slow Progress inMaking Aircrafl Cabin Interiors Fireproof, GAOIRCED-93-37 (Washington, DC: U.S. Government Printing Of-fice, January 1993), p. 26.25 Tho~s L. Reynolds, “Study of the Disbenefits Cre-ated by the Installation of Water Spray Systems for FireProtection of Aircraft Cabins, ” Final Report and IndustryConcerns Preface, December 1992, p. iv. Figures citedfor systems installed aboard Boeing 757-200 and 747-400airplanes, respectively. For the 747, the correspondingrevenue loss, based on $1 million and 200 lbs. per passen-

5er seat, would be roughly $10 million per year.6 Ibid .

27 Ibid., p. vii.

gas concentration levels were lower, and thesurvival times greater than those in a fullysprayed cabin.28 The optimal amount of waterand its distribution requirements have yet to bedetermined. The drawbacks associated withusing a system with a small fraction of thewater required by the original concept shouldbe reassessed.

FAA is also evaluating the effectiveness ofanother CWSS concept, one which employssheets of water to act as curtains between sec-tions of the aircraft and contain the fire within asmall region of the cabin. Using nozzles fash-ioned by British Petroleum and sensor/activa-tion systems developed by GEC Avionics, theBP/GEC system would function similarly to thefirst design (see figure 3-l). Relative systemeffectiveness for equivalent water supplies hasnot yet been determined. Other options forminimizing the weight penalty of CWSS in-clude the use of potable water and, in the longterm, water reclamation systems.

A CAA study of turbine-engine aircraft acci-dents involving fire deaths compared the po-tential benefits of five improvements to cabinsafety .29 Assuming each improvement wasapplied uniquely, CAA found that the expectedsaving of life was much higher for water sprayand smokehoods than the other options. Indus-try has argued that the study was biased towardwater spray because the majority of the aircraftincluded in the assessment were first- and sec-ond-generation models that lacked many oftoday’s fire safety improvements and hadhigher accident rates.30 Changing demograph-ics indicate that the average age of airplanepassengers will be increasing, suggesting thatthe ability of passengers to move about and

28 Hill et al., op. cit., footnote 23, P. 1.29 The five proPsed improvements were: smokehoods,CWSS, improved access to overwing exits, more fire-re-tardant cabin lining materials, and minimum spacingrequirements for seats. Ron Ashford, “Air Safety Regula-tion and Its Commercial Impact, ” l%e AeronauticalJournal, vol. 95, No. 943, March 1991, p. 85.30 Thomas L. Reynolds, “Study of the DisbenefitsCreated by the Installation of Water Spray Systems forFire Protection of Aircraft Cabins, ” draft report, July1992, p. 63.

Chapter 3 —Research and Technologies for Evacuation Systems . 31

I

=—--. .——

/,// “’z~, -:


rapidly exit the aircraft if necessary will bediminished. In general, then, technologies thatfurther mitigate the thermo-toxic effects andextend cabin survivability periods would havegreater benefits than attempts to further speedthe evacuation rate.

Risk/Risk AssessmentWhen the interactive effects of introducing a

new technology are considered, the overall re-sult may be less rather than more safety. Waterspray systems may reduce the risk of fire-re-lated fatalities but could contribute to an overallincrease in risk to passenger safety-for exam-ple, inadvertent discharge during takeoff orlanding phases of flight may distract pilots orcause critical avionics to fail. Similarly, whilesmokehoods could extend survivable conditionsfor a fraction of passengers, other passengerswho might also have survived may, by delayingtheir escape in order to don smokehoods, beovercome by fire and smoke despite thebreathing assistance.

In addition to technical feasibility andcost/benefit analyses, risk/risk assessmentsmust be an essential part of the decisionmakingprocess when the likely safety improvementafforded by new technology is marginal. This isespecially true of commercial aviation, wherethe overall fatality risk to passengers is alreadyless than 1 in 10 million per flight.

Training and OperationsThe ability of flight attendants to quickly

.“ assess and respond to an in-flight or groundemergency affects passenger safety as much asthe design of the aircraft and the performanceof emergency equipment. The National Trans-portation Safety Board (NTSB) believes that asthe crashworthiness of aircraft and survivabilitycontinues to improve, flight attendants “. . . areassuming a more critical role for ensuring pas-senger safety. “31

31 NatjOm] Trmspwlalim Safety Board, ~lighf~fletiTraining and Performance During Emergency Situations,Special Investigation Report NTSB/SIR-92/02(Washington, DC: June 9, 1992), p. 37.

Flight attendants’ spokespersons cite fatiguefrom lengthy duty times as providing potentialfor diminished capability during emergencies.However, the quality of their initial and recur-rent training is perhaps more crucial. Flightattendants rely heavily on this training in emer-gency situations because real emergencies arerarely encountered in commercial aviation,providing little opportunity to practice the nec-essary skills. 32 Technologies assuming a largerrole in training flight attendants include motion-based cabin simulators, fill-scale cabin/cockpitevacuation trainers, cabin evacuation simula-tors, and actual aircraft.33 Some operators alsouse computer-assisted instruction.34 However,the training provided in mockups does not testthe flight attendants’ ability to manage passen-ger flow, which has become increasingly im-portant as seat density has increased.35

NTSB recommends that FAA requireevacuation drills and group exercises duringrecurrent training, and that flight attendantsdemonstrate proficiency in managing passengerflow with verbal commands when competitivebehavior is displayed.36

No matter how well-designed an aircraft orwell-trained the flight attendants, passengerscan undermine the safety capability by bringingon board excessive or inappropriate carry-onbaggage, damaging safety equipment, ordrinking to the point of becoming unable torespond to emergency instructions. In the 1992evacuation from an L-1011 (see box 3-A), onepassenger insisted on keeping a set of largeanimal horns while he exited the plane. 37 More

32 Ibid., p. 1.33 Ibid., p. 18.34 Ibid., p. 19.35 Nora Marshall, senior accident investigator, NationalTransportation Safety Board, persoml communication,

NOV. 16, 1992.36 Natio~] Transportation Safety Board, Flight Afleti~Training and Pe@ormance During Emergency Situations,Special Investigation Report, NTSB/SIR-92/02

$Washington, DC: June 9, 1992), p. 31.7 Independent Federation of Flight Attendants,

“Recommendations to the National Transportation SafetyBoard Concerning Trans World Airlines Flight 843, ” July30, 1992, p. 3.

Chapter 3-Research and Technologies for Evacuation Systems . 33

Box 3-A–TWA Flight 843 Evacuation

On July 30, 1992, shortly before 6 pm, TWA Flight 843 from New York to San Franciscoaborted a takeoff from JFK airport. The plane quickly came to rest to the left of the runway andcaught fire. Despite having but three of eight operable exit doors, there were no fatalities, in part dueto the presence of off-duty flight attendants.

According to preliminary National Transportation Safety Board (NTSB) investigations, theLockheed L-101 1 took off as normal and rose 50 to 100 feet before returning to the ground. Somepassengers and flight attendants commented that something felt amiss with the plane prior to andduring liftoff, but they could not be any more specific. Crew members and witnesses indicated that theaircraft landed very hard, causing the wings to flex excessively. A crew member in a plane awaitingtakeoff reported that he saw and smelled jet fuel emanating from the plane immediately after it camedown.

A fire quickly ensued and engulfed the aft portion of the plane, preventing the evacuation ofpassengers from all but three forward exits. By all accounts, the flight attendants responded swiftly,and evacuation was complete in approximately 2 minutes. Of the 273 passengers, 10 were injured,only 1 seriously. Flight attendants aboard the L-1011 stated that some passengers panicked and lefttheir seats before they were told to do so and before the plane completely stopped. Investigators notedthat a significant number of passengers climbed over the seat backs in order to exit the plane.

Nine flight attendants were assigned to flight 843, three more than the six required by FederalAviation Regulations, and five off-duty flight attendants were on board as passengers. According to anIndependent Federation of Flight Attendants report, the eight additional flight attendants played asignificant role in the safe evacuation of the passengers. For example, the on-duty flight attendantassigned to the L-2 exit could not see if there were flames outside through the door’s prismaticwindow. When she moved to a passenger seat window to get a better view, an off-duty flight attendanttook over her post and prevented passengers from crowding the exit. The off-duty attendant thenopened the hatch when the on-duty flight attendant verified that it was safe to do so. Subsequently,passengers became jammed at L-2, and the on-duty attendant instructed them to proceed to the L-1exit.

SOURCES: National Transportation Safety Board, Office of Aviation Safety, “Factual Report of Investigation, ” unpublishedreport, 1992; and Independent Federation of Flight Attendants, “Recommendations to the National Transportation Safety BoardConcerning Trans World Airlines Flight 843, ” July 30, 1992.

thorough enforcement of carry-on luggage rules evacuations may in fact indicate that additionalhas also been sought by flight attendant unions.

In 1985, the Training and Operations Work-ing Group, established for the FAA’s technicalconference on emergency evacuation, recom-mended that FAA conduct research in commu-nication techniques, behavioral sciences, andoptimum learning situations to further improvecomprehension and retention of safety instruc-tions by passengers. FAA responded that thenumber of passenger-initiated unwarranted

passenger training could have a negative effecton overall passenger safety .38 Rather thanwithhold information that may assist passengersin surviving a real emergency, crew coordina-

38 Federa] Aviation Administration, Emergency Evacu-ation Task Force, “Report of the Training and OperationWorking Group Meeting, December 3-4, 1985,Washington, DC, ” Td.rk Force Report on EmergencyEvacuation of Transpoti Airplanes, DOTIFAAIVS-8611,11(Washington, DC: U.S. Department of Transportation,July 1986), p. 18.


tion and communication could be improved toreduce the potential for unwarranted evacu-ations. Other technologies and training aids,

. including computer simulation, that may fosterbetter communication between the flight crewand attendants should be explored. In addition,some operators use videos (on newer modelaircraft) to heighten passengers’ attention to theairline’s safety briefing.39

Passenger education is only briefly mentionedin the National Plan for Aviation Human Fac-tors; flight attendant training is not. Accordingto FAA, a forthcoming revision to the NationalPlan is expected to address the cabin environ-ment.

EVACUATION RESEARCH ANDTECHNOLOGIES

Little information is available on the behaviorof evacuation systems and passengers in realaccidents except for data recounted by wit-nesses and survivors after the fact. It is impos-sible to realistically simulate an emergency en-vironment without exposing participants toconsiderable danger. To study the effects ofvarious human behaviors on overall evacuationperformance, researchers have used controlledand carefully staged emergency scenarios. Inaddition, researchers have developed and im-plemented complex evacuation models whoseresults depend on pre-set values or randomvariables representing human behavior. Anumber of persons interviewed for this paperfelt that more “realistic” evacuation tests thatattempt to introduce panic by exposing testparticipants to significant hazards would beunethical.40

In the United Kingdom, CAA has attemptedto introduce competition among passengers

39 The sou~em California Safety Institute, a spinoff ofthe University of Southern California’s Safety ScienceDepartment, produces aviation safety videos and conductstraining programs, audits airlines’ safety programs, and soforth.40 AI1 Federal biomedi~l and behavioral research utiliZ-ing human test subjects is governed by a common rule forthe protection of human subjects. See 14 CFR Part 11,June 18, 1991.

during evacuation testing by offering financialincentives to limited numbers of test partici-pants. Additional work using smoke and cabinwater spray has been recently completed. Ac-cording to Lionel Virr of Europe’s Joint Avia-tion Authority: “. . . the issue of competitivebehavior must be resolved to allow harmoniza-tion of future evacuation standards. “41 CAMIinvestigators are considering initiating coopera-tive research with the United Kingdom’sCranfield Institute of Technology (CIT) tocompare motivational techniques.42

Building evacuation research has included thedesign and development of several computermodels to predict egress under various fire sce-narios. These have some application to the de-velopment of models for aircraft. FAA hassupported evacuation model development inprevious years. In 1991, the Air TransportAssociation (ATA) began funding research bythe Southwest Research Institute (SwRI) intosimulation of passenger behavior in aircraftaccidents.

This section discusses evacuation R&T pro-grams, including testing to support rule changes(e.g., CAMI test of seat row separation stan-dard for overwing Type III exit), from whichan improved understanding of evacuation issuesmay be derived. It also discusses developmentsin computer modeling and simulation of pas-senger response during an emergency evacu-ation.

Evacuation TestingIn 1986, CAMI studied flow rates through

the overwing exits and exit preparation timesunder the following four different seating con-figurations:

41 Advanc~ Ru]emaking Advisory committee, Emer-gency Evacuation Subcommittee, minutes of the Jan. 24,1992, meeting, Washington, DC.42 Jeffrey H. Marcus, mawgcr, protection and su~ival

Laboratory, Civil Aeromedical Institute, persoml com-

munication, Jan. 13, 1992.


Figure 3-2-- Schematic Representation of the

AFAA standard

Four Evaluated Seating Configurations

c CAA proposal . Exit hatch

B CAA standard

37” r \

a ~~ard Seat removed.

KEY:FAA= Federal Aviation AdministrationCAA=United Kingdom, Civil Aviation Authority

SOURCE: Federal Aviation Administration, 1992.

the existingments;the minimumairworthiness

CAA minimum require-

requirements of the CAAnotice (see section on

“Exits” in chapter 1);the existing Federal Aviation Regulations(FAR) minimum requirements; and,an alternative proposed by FAA, i nwhich the seat adjacent to the exit is re-moved.

Observed egress rates were faster for theproposed CAA configuration and FAA’s alter-native arrangement than for the configurationspecified in the existing FAR (see diagrams in

figure 3-2).43 FAA observed no statisticallysignificant difference in exit preparation timesfor the various configurations.

After releasing a notice of proposed rulemak-ing for improved access to overwing exits inApril 1991, the Regulations Branch of theTransport Airplane Directorate requested thatCAMI conduct a second study of egress effi-ciency for different seating arrangements.44

4 3 pau] G. Msmussen and Char les B. Chittum, me

Influence of A~”acent Seating Configurations on Egressllrough a Type III Emergency Exit, Final Report,DOT/FAA/AM-89/14 (Washington, DC: December1989).44 Garnet A. Mc~n et a]., Civil Aeromedical Institute,Eflects of Seating Configuration and Number of Type IIIExits on Emergency Aircrajl Evacuation, Final Report,


Test results indicated that, of the total time re-quired to evacuate through a single Type IIIexit, the amount of time a passenger needs tomove from the center aisle through the seatsand out the exit depends greatly on the ergo-nomic restr ic t ions encountered at the exi topening (i. e., increasing the pathway width ordecreasing the r e s t r i c t ed d i s t ance to betraversed results in shorter egress times) .45

Based on the results from evacuation trialswith a dual Type 111 exit configuration, FAAhypothesized that arranging the seat rows suchthat only one pathway leads to each exit wouldmaximize the flow rates to and through thehatches. Aircraft with exit centerlines 29 inchesapart (e.g., the Fokker 100) would have diffi-culty achieving this configuration. 46 In May1992, FAA issued a final rule revising seatspacing standards for rows that lead to over-wing exits; the implementation deadline wasDecember 1992. 47

In 1987, CIT commenced a CAA-sponsoredprogram of research into passenger behaviorduring emergency evacuations. Analyses ofaircraft accidents indicated significant conges-tion occurred during some emergency evacu-ations at galley entrances and overwing (TypeIII) exits.48 CIT research sought to determinewhether an optimum aisle width through thecabin divider (bulkhead) near the Type I exit oran optimum seating configuration adjacent toType III exits existed. Two independent seriesof evacuation trials using different bulkheadapertures and seating configurations were per-formed, with one series employing financialincentives to foster competitive behavior amongtest participants.

CIT efforts to introduce as much realism aspossible during the test included:

DOT/FAA/AM-92/27 (Washington, DC: U.S. Depart-ment of Transportation, August 1992), p. 1.45 Ibid., p. 5.46 Ibid., p. 6.47 57 Federa/ Register 19220 (May 4, 1992).48 Helen C. Muir and Trevor J. Gilpin, “Egress UP-

date, ” paper presented at the Eighth AMuid InternationalAircraft Cabin Safety Symposium, Costa Mesa, CA, Feb.4-7, 1991, p. 2.

using an actual aircraft, a Trident Three;training and dressing researchers as cabinstaff; andproviding pre-flight briefings and playingback a sound recording of an aircraftstarting up and taxiing to a runway, ex-periencing an aborted takeoff, and beingshut down.49

comparing evacuation rates between theseries, CIT researchers concluded that increas-ing the width of the bulkhead aperture leads toan increase in passenger flow rates through theadjacent Type I exit. CIT researchers alsoconcluded that changes to the distances betweenseat rows on either side of an overwing exitinfluence flow rates; however, complete re-moval of the seat row adjacent to the Type 111exit allowed passengers to pool together andresulted in slower evacuation rates than thosemeasured for vertical projections between seatrows in the range of 13 inches to 25 inches (seefigure 3-2).5°

A preliminary investigation into effects fromthe presence of nontoxic smoke was initiated in1989, during which CIT again conducted aseries of evacuations using varying bulkheadapertures and distances between seat rows nextto overwing exits. CIT found that the presenceof smoke significantly reduced the rate at whichtest volunteers were able to orderly evacuatethe aircraft. At CAA’s request, CIT also inves-tigated the effects of nontoxic smoke and cabinconfiguration using competitive behavior. CITfound significant differences in egress rates forfour alternative seat spacings adjacent to TypeIII overwing exits, but observed no statisticallysignificant differences for evacuations throughvarious bulkhead apertures.51

After comparing the results of these tests withdata from the earlier noncompetitive evacuationtrials involving nontoxic smoke, CIT research-

49 Ibid., p. 8.50 Ibid., p. 12.51 Hoc. Muir et al., Cranfield Institute of Technology,Applied Psychology Unit, Aircrafl Evacuations: Competit-ive Evacuations in Conditions of Non-Toxic Smoke, CAApaper 92005 (London, England: Civil Aviation Authority,March 1992), pp. 9, 11.

Chapter 3-Research and Technologies for Evacuation Systems . 37

ers determined that the presence of a competi-tive element had a significant impact on egressrates for evacuations through the bulkheads,but did not affect the rate of evacuation throughthe Type III exit.52 In the latter case, the dif-ference in seat spacing (vertical projection) wasthe dominant factor in egress rates.

CAA also commissioned a study of humanfactors aspects of water spray system use dur-ing cabin evacuations. Using a 707 aircraftframe, CIT conducted eight full-scale evacu-ations, half in dry conditions. Mean evacuationtimes for the two conditions were virtuallyidentical, suggesting the operation of the CWSSdid not affect evacuat ion rates .53 C I T r e -searchers identified no significant visibilityproblems or hazards from wet cabin furnishingsand floor surfaces.

Human behavior in actual emergencyevacuations or even demonstrations for FAAcertification cannot be extrapolated from theresults of these series of CIT/CAA tests (e.g.,because of the differences in participant demo-graphics and small sample sizes). However, thedata have provided insight into the effects ofchanges in human motivation and the cabinenvironment on evacuation capability.

Computer Modeling and SimulationThe mathematical models used by aircraft

manufacturers to predict evacuation times aresimple calculations of total escape times basedon empirical relations for equipment prepara-tion and deployment times and the averagethroughput of exits. (These relations are de-rived from the results of research experimentsand demonstrations for evacuation certification,not from actual emergency evacuations. )

More complex network and queuing modelshave been used to represent the characteristicsof evacuation systems. w Network models,graphic representations of paths by which ob-jects may move from one point to another, areuseful for minimizing the time or distance ofpoint-to-point travel but can quickly grow toocomplex for efficient use on computers .55Queuing models describe the dynamics of .waiting lines, time-dependent processes thatobey the laws of probability.56 The initialpopulation distribution and the probability of aperson moving from one station in an evacu-ation system to another determine the waitingtimes and exit throughput.

Simulation relies on computer-generated ran-dom numbers to represent processes whosevalues cannot be approximated analytically.Parameter variability can be modeled withprobability distributions; step-by-step and item-by-item, the simulation predicts what is likelyto happen by running the model through severalconditions.57 For example, the influence ofvarious hesitation times in the face of a grow-ing fire threat could be observed using com-bined simulation models of aircraft evacuationand fire performance.

“A model is only as good as the parameterswhich describe the system . . . any evacuationmodels developed and used will need an exten-sive program of parameter determination andsensitivity analysis, and an equally extensivevalidation effort. “58 For example, if the pres-ence of passengers with disabilities is assumed,simulation results are of little use unless goodapproximations (distributions) of seat exit andaisle flow rates are incorporated. Both generalevacuation models and simulation effortsspecific to aircraft are discussed below.

52 Ibid., pp. 18-19.53 Researchers noted that the sample SIZeS were smalland that the test results are not as statistically reliable asthose derived from a larger sample. D.M. Bottomly andH.C. Muir, Cranfield Institute of Technology, AppliedPsychology Unit, “Aircraft Evacuations: The Effect of aCabin Water Spray System Upon Evacuation Rates andBehaviour, ” report prepared for the Civil AviationAuthority, February 1993, p. 5.

5 4 John M, watts, Jr., “Computer Models for Evacu-ation Analysis, ” Fire Safety Journul, vol. 12, 1987, p.241.55 Ibid., pp. 237-238.

56 Ibid., p. 240.57 Ibido, pp. 242-243.58 Marcus, op. cit., fOOtnOte 42”


General Models and AssessmentAssessment and modeling of flow problems

involving people began in the early 1980s.Several models were developed to estimate thetime required for groups of people to evacuatea given space or building. Building evacuationwas modeled for situations in which the numberof people inside a lobby affected the rate of exitfrom the lobby,59 and where inhabitants may ormay not be alerted before beginning egress.60

In each case, the network flow solution methodassumed egress occurred through well-definedpassageways.6l Other critical assumptionstypical of the general approaches to solvingrelated flow problems included:

• Any congestion will occur at doorways,and flow through vertical and horizontalpassageways will be relatively free flow-ing;

• Doors serve to meter flow to about oneperson per second per door.62

The building models do not consider damageto exits as flow obstructions.63 Implicit as-sumptions about nonvarying door and passage-way dimensions and stairwell and hallway flowrates do not apply to cabin evacuation, and themodels are inappropriate for conditions involv-ing aisle congestion. None of the models at-tempted to incorporate human decisionmakinginto the process, particularly in response to

59 R.L. ~~a~ci~, A Negative fipone~la[ SOhUIOn to ~Evacuation Problem, Report NBS-GCR-84482 (Gaithers-burg, MD: Natioml Bureau of Standards, December1984).60 D*M. Alvord, ?7w Fire Emergency Evacuation Si~-lation for h4ultijhmily Buildings, Report NBS-GCR-84483(Gaithersburg, MD: Natioml Bureau of Standards,December 1984).61 T.M. Kisko and R.L. Francis, Network Models ofBuilding Evacuation: Development of So@are System,Report NBS-GCR-82-417 (Gaithersburg, MD: NatiomlBureau of Standards, December 1982).62 Haro]d E. Nelson, ‘Fireform’-A Computerized Col-lection of Convenient Fire S@ety Computations, ReportNBSIR 86-3308 (Gaithersburg, MD: National Bureau ofStandards, April 1986), p. 27.63 Matthew McCormick, chief, Survival Factors Divi-sion, National Transportation Safety Board, persomlcommunication, Nov. 16, 1992.

changing fire conditions. Neither panic, push-ing, nor falling was assumed.

Certain methodological problems limit thestudy of human behavior in fires: experimentalsubjects cannot be placed in real fire (or crash)situations; testimony obtained after the factfrom participants in fires may contain errors;and conclusions must be drawn cautiouslywhere sample data are limited or not represen-tative.64

In general, egress research (to fill models’data gaps) has fallen into three main categories:field studies of circulation facilities in non-emergency conditions; laboratory studies (e.g.,sign visibility in smoke); and post-incident sur-veys of human behavior in emergencies.65 Thenature of case studies has progressed frommainly descriptive to more complex, analyticalventures that attempt to identify typical behav-ior patterns or correlate behavior and fire de-velopment .66

Despite the frequent use of the term “panic”to describe human response to emergencysituations, particularly fire, researchers haveconcluded that “. . . people generally respondto emergencies in a ‘rational, ’ often altruisticmanner, in so far as is possible within the con-straints imposed on their knowledge, percep-tions, and actions by the effects of the fire. “67

Continued research into the reasoning and mo-tivation behind individuals’ actions, altruistic ornot, is necessary. Existing models typically donot represent the perception of cues, investiga-tive behavior (e.g., looking for the fire), andgeneral coping behaviors.68

These data have limited application to aircraftevacuation. For example, some of the indeci-

M R.L. Paulsen, “Human Behavior and Fire Emergen-cies: An Annotated Bibliography, ” Report NBSIR 81-2438(Gaithersburg, MD: National Bureau of Standards,December 1981), p. 3.65 Hamish A. MacLennan, “Towards an htegrakdEgress/Evacuation Model Using an Open Systems Ap-preach, ” in Proceedings of the First International Sympo-sium of Fire Safety Science, n.d., pp. 581-590.66 Paulsen, op. cit., footnote 647 P. 5“67 Ibid., p. 4.68 Mac~~n, op. cit., footnote 65.


sion attributable to not knowing for certain if afire has broken out is often absent, time scalesare different, and escape modes differ from thewell-defined hallway model typically used.However, data on how smoke and the sight offire affect decisionmaking skills, sex- or age-related effects, and so forth likely would betransferable.

Computer-based models that incorporate bothhuman performance parameters and systemcharacteristics are being developed. Simulationrequirements include: assessment of risk from,for example, crash-related injury, smoke/gastoxicity, and fire; typical human behavior un-der stress, darkness, and smoke; basic responsetimes under best conditions; and decision-making parameters. The generality and validityof these models must be determined. Modelsmust address performance in dynamic large-scale, multiperson systems, as well as the ef-fects of stress and emergencies.

In addition, the NIST Building and FireResearch Laboratory has developed theHAZARD1 model of evacuation from burningbuildings using fire survivor psychological datato construct human behavior parameters.69

HAZARD1 analyzes the fire environment forallowable egress time and demonstratesevacuation of building occupants based onbehavioral rules obtained from interviews withfire survivors. NIST staff acknowledge the dataare skewed in favor of successful behavior;those who did not survive cannot be inter-viewed. The software can be modified to doprobabilistic branching for non-universal be-havioral patterns; the current version uses onlya deterministic approach, Other work at NISTrelates to congestion in large buildings. Uni-versity of Florida/Gainesville researchers havedeveloped a version of HAZARD1 that allowsoptimization of building and fire safety designs.

Aviation-Specific ModelsIn the early 1970s, FAA developed a com-

puter model of aircraft evacuation usingGeneral Purpose Simulation System (GPSS)

69 Bukowski, op. cit., footnote 1.

language, developed by IBM.70 To estimate andanalyze the escape process, the model usedstatistical functions to control passenger move-ments and to advance time related to eachevent.71 First applied to evacuation tests in twosingle-aisle, narrow-body aircraft configu-rations,72 FAA further developed the model torepresent evacuations from wide-body air-craft.73 The model provided for passengerreassignment to equalize the length of queuesbefore exits. An average of 20 runs was used toevaluate each scenario.

FAA executed simulations of evacuationsfrom DC-10, L-101 1, and B-747 aircraft duringthe same period the wide-body aircraft wereundergoing evacuation certification tests. Thesimulation results correlated well with full-scaledemonstration times.74 However, the simulationmodel could not assess a priori the effects ofhuman behavior. Although FAA’s modelpredicted the total evacuation time for 527 pas-sengers aboard a 747 would be 84 seconds, in ademonstration for certification participants ex-ited the aircraft in under 67 seconds.75 FAAattributed the difference to the motivation ofpassengers and crew.

Recent and Continuing Efforts. Under aFAA/CAMI-sponsored contract initiated in1987, Gourary and Associates developed aclock-driven simulation model of the aircraftevacuation process for use on a computer. Eachcycle, the model recalculates the position ofeach passenger subject to initialized variables:exit preference; endurance, or probability ofsurviving heat or smoke; agility; and “wake-up

TO J*DO Gamer e t al., GPSS Computer Simulation ofAircrafi Passenger Emergency Evacuations, FAA-AM-78-23 (Washington, DC: U.S. Department of Transporta-tion, Federal Aviation Administration, June 1978), p. 1.71 ~rl D. Fo]k et al., GPSS/360 Computer Models TO

Simulate Aircraft Passenger Emergency Evacuation, FAA-AM-72-30 (Washington, DC: U.S. Department of Trans-portation, Federal Aviation Administration, September1972), p. 1.72 Ibjdo, p. 8. Cornpmkon of test rfNIkS fOr 134- and

234-passenger loads showed that larger exits used in thelatter case allowed higher flow rates through the doors,73 Garner et al., op. cit., footnote 70, p. 1.

74 Ibid., p. 5.75 Ibid., p. 6.


time, ” or the time it takes for a passenger tobegin to move purposefully (reflects shock andthe capability of opening one’s lap-belt).76

Increases in heat or smoke, passenger“fatalities,” and disabled exits affect the flowrates through aisles and doorways (i.e., transi-tional probabilities).

The Gourary model is not comprehensive interms of the human behavior assumptions, butit does portray passenger evacuation under avariety of crash/fire scenarios described by 40crash characteristics and narrow-body aircraftcabin layouts.

None of the simulation models describedabove directly addressed psychological factors.Manufacturers and researchers lack the data todetermine how much of a role these factorshave in the overall success of evacuation. Withdevelopment and validation of adequate pa-rameters, the simulation may closely approxi-mate an emergency evacuation.77

ATA-sponsored research by the SouthwestResearch Institute seeks to simulate passengeregress under a variety of evacuation conditionsand passenger characteristics. ATA hopes todevelop safety requirements that are sufficientfor all passengers, including persons with dis-abilities, in all evacuation circumstances. TheSwRI four-phase effort aims to create an air-craft evacuation (AIREVAC) computer modelto simulate passenger behavior during emer-gency evacuations. Phase 1, completed in Sep-tember 1991, entailed a literature search toidentify variables, mathematical relations, andother information necessary to construct themodel, scheduled for Phase 2 of the project.The model validation will be based on eitherarchival evacuation data or on data from a newevacuation exercise .78

76 GouraV Ass~iates, Inc., “Evacuation of passengersFrom Transport Aircraft: A Microcomputer-BasedModel, ” User’s Manual for Evacuation Simubtor Dem-onstration Diskette, Version 7.03, May 31, 1991, pp. 7-9.77 Marcus, op. cit., footnote 42.

78 James E. Schroeder and Megan Tuttle, Developmentof an Aircraft Evacuation (AIREVAC) Computer Model,SWRI Project No. 12-4099 (San Antonio, TX: SouthwestResearch Institute, Sept. 30, 1991).

SwRI’s literature search yielded no system-atic overview of the evacuation process; rather,the effort produced references to work on spe-cific issues and concerns.79 The SwRI modelvariables address situational characteristics;passengers’ physical and psychosocial charac-teristics, including motivational variables; andbehavioral outcome variables. The latter in-cludes initial response, helping another passen-ger, panic, and competitive behavior.80

Data Issues. Although FAA’s simulationmodel provided for variations in passengermix, seating and exit configuration, door-opening delay, time on the escape slide, andslide capacity, insufficient data were availableto establish appropriate variables representingthe different influences on evacuation rate.8l

Also, the lack of data on the effects of adverseconditions (e.g., smoke and debris) preventedtheir simulation.82

Boeing said that its own simulation effort inthe 1980s was dropped in the belief it wouldnot significantly improve the evaluation ofevacuation systems and procedures, given thelack of evacuation data to substantiate thesimulation model and the reliability of its exist-ing mathematical models,83

Today, as in the 1970s, no central clearing-house for evacuation data exists.84 The largestcollection of data published to date, the Aero-space Industry Association’s report on its year-long evacuation system study, was completedprior to the conduct of mostcertification tests.85 Phase

wide-body aircraft3 of the SwRI

79 Ibid., p. 8.80 Ibid., pp. 9-11.81 Gamer, et al., op. cit., footnote 70, p. 2.82 Ibid., p. 11.83 George Vewiogiou, senior mamger, 747/767 payloadSystems, Boeing Commercial Airplane Group, personalcommunication, Dec. 12, 1991.M Gamer et al., op. cit., fOOtnOte 70, P. 4.85 ~id., p. 4. See also Aerospace Industries Associationof America, Inc., “Evacuation, ” Technical Group ReportAIA CDP-4, hdy 1968.


simulation task, designated for filling data“holes,” is yet unfunded by ATA.86

Studying events related to single-aisle, nar-row-body aircraft is possible with test beds inthe United States and elsewhere. However,there is no research facility in the world thatcan be used for investigating wide-body, dual-aisle aircraft evacuation issues .87 The fiscalyear 1995 FAA capital budget contains fundingfor such an evacuation facility. (A 747-100 hasbeen offered to CAMI--the difficulty lies ingetting it to Oklahoma City.) Just as certifica-tion of the 747, with its dual-aisle configura-tion, introduced more complexity into analyti-cal methods, the proposed super jumbo aircraft(seating 550 to 800) will also stretch the capa-bility of existing models and facilities.

The need for more data to extend the utilityand reliability of the simulation technique isapparent. Improved accident data analysis, pas-senger demographics information, thermo-toxic

86 James Schroeder, president, Applied Human Factors,Inc., personal communication, Oct. 28, 1992.87 Marcus, op. cit., fOOtnOte 42.

environment information, parameterization offlight attendant and passenger behaviors, and atest bed for evaluating wide-body aircraft sce-narios are required to validate evacuationsimulations.

Even augmented, validated simulations mayhave their detractors. One passenger advocacygroup has expressed alarm at the possibility thatthe SwRI computer simulation models spon-sored by ATA will be used to rationalize limit-ing the number of passengers with disabilitiesallowed on board transport aircraft.88 One canexpect that this or any other test or analysis ofevacuation performance is likely to produceslower egress times as the percentage of olderpassengers, children, or persons with disabili-ties on board aircraft increases. This fact oflife, along with equity and other issues, willaffect those finally making policy decisions.

88 Fred Cowell, executive director, Paralyzed Veteransof America, statement before the Emergency EvacuationSubcommittee of the Aviation Rulemaking AdvisoryCommittee, May 14, 1992.

Chapter 4

FINDINGSAND

CONCLUSIONS

CHAPTER 4

Findings and Conclusions

An aircraft’s evacuation capability is one ofmany safety issues the Federal Aviation Ad-ministration (FAA) reviews before the aircraftis permitted to enter service. Evacuationequipment is one of a long line of measures in-tended to improve passenger safety in the eventan emergency occurs aboard an aircraft. How-ever, technology alone does not ensure that apassenger escapes the cabin under adverse cir-cumstances. The abilities and actions of flightattendants and the passengers themselves factorgreatly into the success of an evacuation.

BENEFITS AND LIMITATIONS OFEVACUATION DEMONSTRATIONS• Full-scale demonstrations are costly and

expose participants to significant hazards.The cost of conducting full-scale demonstra-tions can exceed $1 million each. For exam-ple, the cost for the first attempt to certificatethe MD-1 1 aircraft with 410 passenger seatswas approximately $1.3 million. 1

On av-

erage, approximately 6 percent of partici-pants are injured during full-scale tests.Participant injuries were the basis for the1978 rule change allowing analysis and par-tial/component tests to replace full-scaledemonstrations when conditions warranted.While most injuries have been minor, brokenbones and paralysis have occurred. Lesssevere injuries than expected for traditionaldemonstration formats occurred in the FAA-approved December 1992 MD-11 certifica-tion test in which slides were replaced withramps and the exterior of the aircraft waslighted.

• A full-scale demonstration simulatesevacuation for only a narrow, optimisticrange of emergency conditions. The certifi-cation requirements represent an abortedtakeoff at night (i.e., no structural damage,cabin fire, or smoke) with a distinct subset of

1 Webster C . Heath, manager, Technical Liaison,

Industry Regulatory Affairs, Douglas Aircraft Co.,personal communication, July 8, 1992.

potential passengers (i.e., no children, per-sons with disabilities, or non-Englishspeaking passengers).

There are major weaknesses with usingFAA full-scale demonstrations as measuresof evacuation performance: 1) only one datapoint is provided for a measurement thatcould have a broad probability distribution;2) the selection criteria for test “passengers”do not reflect actual passenger demograph-ics; and 3) tests do not encompass many ofthe conditions in actual accidents.

• Successful evacuation in an actual emer-gency depends on more than the flow ratesdemonstrated for certification. Factors inthe outcome of a real emergency evacuationinclude: cabin and flight crew capabilities;aircraft integrity and seating technologies;passenger and baggage characteristics; andactual accident conditions, such as fire andsmoke.

Adding “realism” (e.g., smoke and fire) tofull-scale demonstrations as they are cur-rently configured increases the risk of injuryto test participants without guaranteeing re-duced risk of injury to the flying public. Thevariability of actual accident conditions can-not be represented with only a fewapproximations of emergency settings.

The Performance Standards WorkingGroup of FAA’s Aviation Rulemaking Advi-sory Committee, charged with developingrevised emergency evacuation requirementsand compliance methods for reducing therisk of injury to full-scale demonstration par-ticipants, was not able to reach consensus onits proposal to modify the certification test torely more heavily on analysis. Recommenda-tions and a final report were submitted inJanuary 1993. The working group has sincebegun to consider whether design standardsfor emergency evacuation can and should beconverted to performance standards.

43


• FAA acknowledges that demonstrationsprovide only a benchmark for consistentevaluation of various seating and exit con-figurations; the requirement to demonstratecomplete evacuation within 90 seconds is notan adequate performance standard for meas-uring evacuation capabilities.2 Nor is theone-time demonstration useful for system op-timization; it provides manufacturers a singleopportunity to observe flaws and take cor-rective actions. After a second attempt to at-tain certification, one cannot be confidentthat differences in test results are statisticallysignificant.

Because compliance with some of the testconditions is subjectively determined, vari-ability in the test conditions can occur. Dur-ing the October 1991 certification test for theMD-11 aircraft, the simulated “dark ofnight” and crew mix requirements werethought to be more rigorous than for priortests.

• Present evacuation certification rules donot encourage development of new tech-nology for extending the period of surviv-ability in postcrash fires. FAA has empha-sized the development of technologies thatcan speed evacuation rates and reduce totalevacuation time (e.g., faster deployingslides, floor-level path lighting). But FAArules give no credit for technologies that, inreal life, could extend the period of surviv-ability within the cabin.

MODELS AND SIMULATIONS FOREVACUATION CERTIFICATION• At present, neither certification by fill-scale

demonstration nor by purely analytical meth-ods is acceptable to all segments of the avia-tion community. Manufacturers feel existingdata from prior evacuation certification testshave validated their mathematical modelssuch that full-scale demonstrations can oftenbe replaced by combinations of compo-nent/system tests and analysis, but statisticalanalysis of data and model sensitivity have

z M Federal Register 26692 (June 23, 1989).

not been explored. Passenger, flight atten-dant, and pilot groups have expressed con-cern with reliance on analysis to demonstratecompliance with evacuation standards.

• The aircraft certification process will likelycontinue to rely on human test subjects inthe foreseeable future. However, a combi-nation of analysis and partial demonstrationsor component tests can be developed tominimize the risk of injury and provide morecomprehensive data on aircraft performancethan fill-scale demonstrations.

• Human behavior in certification tests maybe empirically modeled using data fromprior demonstrations, but cannot yet bereliably “simulated.” Estimates for averagereaction times and egress rates are known forevacuation during controlled conditions. Be-cause few reliable data exist on human be-havior during accidents, the variations inhuman judgment and decisionmaking thatmight be expected for changing hazardousconditions cannot be predicted. These datacannot be obtained from current demonstra-tion requirements, which do not address mo-tivational effects or other behavioral factorsthat often exist in a real emergency. Manu-facturers’ mathematical models are insensi-tive to age, sex, and other characteristics ofdemonstration participants.

Evacuation models developed for buildingsinclude some human behavioral factors butare not fully transferable to aircraft. For ex-ample, there are many configurational differ-ences. The psychological data used in thesemodels are limited to that obtained in inter-views of building fire survivors.

Computer simulations, creating repeatedand varied evacuation trials, may be morevalid as measures of an aircraft’s evacuationperformance than a single full-scale demon-stration would be. At the very least, thesimulations can suggest a range of outcomesfor given test conditions. Recent computersimulation efforts may provide the technol-ogy base for an improved simulation capabil-ity. The additional psychological data

Chapter 4—Findings and Conclusions . 45

required for validating behavioral assump-tions will be difficult to attain.

• Evacuation rates for existing aircraft com-ponents are predictable. The results of in-dustry analyses typically correlate well withobserved rates through doors, aisles, slides,and other components under consistent testconditions. What is not known or predictableis the performance of the emergency evacu-ation system in an actual emergency (i.e.,how many doors/slides will be inoperable orblocked by fire/smoke, how quickly willsmoke or flames enter the cabin if a post-crash fire occurs, what interactions will takeplace between passengers to speed or slowthe evacuation?). OTA notes that evacuationtrials with human test subjects measure noneof these events. Factors that greatly affect theoutcome of a real emergency evacuation in-clude: cabin and flight crew capabilities; theintegrity of the aircraft and cabin furnishings;passenger and baggage characteristics; andhazardous conditions.

DATA ISSUES• Additional experimental data are required to

validate models/simulations. Earlier industrysimulation efforts stalled due to the lack ofhuman factors data. The data collectionphase of a current simulation developmentproject has been delayed for lack of funding.Although FAA’s present test fuselage is ade-quate for studying issues related to single-aisle, narrow-body airliners, there is no facil-ity, worldwide, that can be used to analyzeegress from double-aisle, wide-body trans-ports.

• Data on injuries related to evacuation testingare not readily available, nor are they classi-f ied by severi ty . Nei ther FAA nor theNational Transportation Safety Board(NTSB) collect information on precaution-a r y3 evacuat ions. Data from actual emer-gency evacuations are unevenly collected andanalyzed. With current database structures,

evacuation performance data must be pains-takingly gleaned from accident reports.

• Recommended changes to FAA fire safetytest objectives in the 1980s helped FAA todevelop improved test procedures and obtainmore comprehensive data for rulemaking.The National Institute of Standards andTechnology and the United Kingdom’s CivilAviation Authority are two excellent sourcesfor complementary evacuation and fire data.

AIRCRAFT EVACUATIONPERFORMANCE AND SAFETY• Survivability in commercial air transports

is improving, largely through the introduc-tion of technologies that mitigate impactforces and delay incapacitation from smoke,heat, and toxic gases.4 Though still a sig-nificant threat, fire has become less of a riskin survivable accidents. In the early 1980s,FAA attributed 40 percent of fatalities insurvivable accidents to fire effects. A reviewof U.S. airline accidents that occurred be-tween 1985 and 1991 showed that approxi-mately 10 percent of fatalities were related tofire. Two central elements of evacuation sys-tems are heat-resistant slides that inflate anddeploy automatically and floor-level pathlighting.

• Crew training and passenger actions are ascrucial to successful evacuations as are theaircraft’s design and equipment. Accordingto NTSB, as the crashworthiness of aircraftimproves, flight attendants assume a morecritical role in ensuring passenger safety.Flight attendant training, done in cabinmockups without passengers, may not pro-vide crew members with sufficient skills formotivating passengers to evacuate more effi-ciently and assessing flow control problems.Because simulation could rapidly show thepotential results of different commands andcrew actions on the outcomes of emergencyevacuations, simulation technologies could

3 Evacuations in cases where the threat of fire or harmlater disappeared.

4 By delaying the onset of smoke, the presence of fire

retardant materials augments passenger vision as well asimproves the potential for survival.


enhance training in passenger managementand use of emergency equipment.

• There is wide variation in the mobility,strength, and perceptive capabilities of air-craft passengers. Under existing aviationregulations, airline operators restrict seatingin exit rows to those persons willing and ableto read, hear, and understand emergency in-structions and to operate evacuation equip-ment.5

• Another factor is the presence and type ofcarry-on baggage. Some increase in cabinsafety could be expected from further restric-tions on carry-on baggage, which in an acci-dent can impede passenger movement fromseat to aisle and aisle to exit. In addition,passengers often stop to retrieve carry-onitems, which flight attendants must removebefore the passengers use the slides.

• Technology efforts to suppress or mitigatethermo-toxic conditions will likely aid pas-senger survivability more than efforts tofurther speed evacuations. Changingdemographics suggest passenger evacuationrates will be, on average, slower in the fu-ture. To achieve significant reductions inevacuation times from typical seating con-figurations would require more doors (whichadd weight and reduce seating capacity, re-sulting in revenue losses) or fewer passen-gers (more lost revenue). Speeding evacu-ation through small changes in technologywould be difficult. British analysis of firedeaths in international accidents during the1980s indicated that new technologies in-tended to delay deadly heat and toxicity lev-els after a crash are likely to save many morelives than would efforts to further speedevacuation. 6

5 55 Federal Register 8072 (Mar. 6, 1990).

6 Ronald Ashford, “Air Safety Regulation and ItsCommercial Impact, n Aeronautical Journal, vol. 95, No.943, March 1991, p. 85.

APPENDIX

Figure A-1 -- Comparison of Hypothetical Test Results Using “Real World” vs. Certification Test Passenger Mixes

1 KEY

Hypothetical probability distribution forpassenger mix used in certification tests (e.g.,70%male, 30% female, passengers in “normalhealth,” 570 over 60 years of age, no children).

Hypothetical probability distribution forpassenger mix on actual airline flights (e.g.,children, persons with disabilities, andelderly).

“real world”

>

Egress time

NOTE: This figure shows a rough estimate of sample distributions that might be produced if repeated evacuations were conducted: 1)under existing FAA criteria; and 2) with passenger loads representative of the real world. The estimated distribution of egress timesunder the real world conditions is much broader and its mean is greater.


Figure A-2 -- Comparison of Hypothetical Sample Distributions for Aircraft Evacuation Times

● ✝ Observed evacuation timefrom single certification test.

“faster” “slower”aircraft aircraft

1

90 seconds Egress time

NOTE: This figure illustrates the problem of assuming a single test will indicate which is the better or more acceptable of two aircraft testedfor compliance with evacuation certification requirements. It represents two theoretical sample distributions of egress times for a“slower” and a “faster” aircraft, as might be obtained from repeated evacuation tests of the two configurations. As shown in thefigure, it is quite possible for the faster aircraft (having a shorter average egress time) to fail and the slower aircraft to pass, basedon single tests.


Chapter 4—Findings and Conclusions . 51

I

I

I

Aircraft Evacuation Testing: Research and Technology Issues

Documents

Transcript of Aircraft Evacuation Testing: Research and Technology Issues