Early Warning Strategy - NILU
Transcript of Early Warning Strategy - NILU
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Early Warning Strategy
Joel TaylorUCL Centre for Sustainable Heritage
Review of Preventive Conservation in EuropeEarly Warning System
Mitigation
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Review of Preventive
Conservation in Europe
Literature Review – published work on best practice and what should be done
Questionnaire – views on current practice and what is being done
Expert panel – end-user views on current practice and what could be done
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Literature Review
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Gaps in Preventive
Conservation StrategyTechniques often not integrated
Damage functions not known for many materials and difficulty in generalising from one material to a range of materials
Synergy of agents of deterioration
Effects of open display, e.g. historic houses and contemporary exhibitions
Data analysis deficit
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
(Problems with) Standards
Dosimeters cannot utilise existing standards – why?RH levels often mid-range, and concentrate on physical damage and mould growthTemperature incorporates human comfortLight levels incorporate visibilityFew standards on pollutants (concentration-based)No accepted standards for VOCs
The numbers, rather than the ways of achieving them, gained acceptance.
More recently, standards have evolved from single numbers to ranges
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Conclusions from the
Literature ReviewPreventive conservation strategies are drawing
upon risk management more and more
Acceptable change is being defined
The rate of deterioration for different locations can be compared, using techniques such as isoperms
An overall strategy can accommodate different aims
Interdisciplinary methodologies are replacing standards
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Questionnaire
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
QuestionnaireQuestionnaire on
monitoring
Sent out to determine how attitudes and practice reflect literature
31 heritage institutions of varying size from 18 European countries responded
Archive Museum Historic House
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Perceived causes of damage
02468
101214161820
RH
Tem
p
SO2
NOx
O3Lig
ht/U
VOr
g ac
idsPa
rticle
s
Pests
Hand
ling
Res
pons
esPrimary Cause Secondary causeNot important Unknown
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Is monitoring systematic?
Everyone that monitors at all, monitors temperature and RH. Very few people monitor pollutants
For many risks, there appeared to be very little correlation between perceived threat and amount of monitoring -except pests and organic acids (risks with visible impact)
All of the institutions that did not consider temperature or RH as important risk factors (or didn’t know), monitored them
Pollution monitoring is independent of location -proportionally, there is more pollution monitoring in rural locations than urban ones
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Questionnaire conclusions
Awareness of risks was related to standards, and therefore intellectual access to information
Pollution and organic acids were not frequently monitored
Rational decision-making is distorted by external factors, such as lack of resources
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Expert Panel
First end-user workshop in Krakow
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Outcomes of discussion
Light, temperature and RH commonly monitored. Pollutants, less so
Differences in monitoring are more related to the size and type of institution than differences in culture or country
Data overload is the most common problem
Most systems rely on single parameter monitoring, not considering synergistic effects
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Overall recommendations
‘Data overload’ a problem
Preventive conservationmethods often not integrated
EWO should help define relationship between risk and damage
EWO should be easy to use
Need for EWO to be relevant to existing methodologies
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Convergence of findings from literature
review, questionnaire and expert views -1Conservation assessments need to account for
synergy among different risks (all research strands revealed this)
Consistent way of comparing different risks needs to be carried out in practice (all strands)
More data analysis should be encouraged in practice (all stands)
Theory and practice emphasise some risks over others (literature review and questionnaire)
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Convergence of findings from literature
review, questionnaire and expert views - 2Understanding, and expression, of rate of change
for chemical deterioration is still required (literature review and workshop )
Data overload and lack of integration of methods within preventive conservation (questionnaire and workshop)
Communication between scientists and the general public is required (literature review and workshop)
Integration of existing preventive conservation methodologies is desirable (literature review and workshop)
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Early Warning System
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Developing a new preventive conservation strategy for the
MASTER sensorsIntegrate EWO sensor with existing
preventive conservation methodologies
Understand and express, of rate of change for chemical deterioration
Account for synergy between agents
Encourage more data analysis (simplicity)
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Existing Preventive
Conservation Methods
Environmental MonitoringCondition AssessmentLocation AssessmentRisk Assessment
Vantaa, 2000
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93FUTURE
PAST
DETERMINISTIC CATASTROPHICDeterioration
Location Assessment
Condition Survey Risk Assessment
Environmental Monitoring
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Properties of the EWO-G dosimeter
False PositiveResponse, no damage
False NegativeDamage, no response
True positiveResponse and damage
True negativeNo response or damage
Maritime vapours
Unusual VOCs
Inherent deterioration
Rate of reaction
Physical damage
Biological damage
Dose-response (Tetreault, Brimblecombe, Larsen)
Isoperms (Sebera, Michalski, TWPI)
JNFs (Ashley-Smith . et al.)
Passive layers on objects
Agents reducing reaction rates
Uncertainty (Ashley-Smith)
Stability (Bradley and Thickett)
Deterioration Dosimeter Response
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Classification in
Preventive Conservation
Building (ASHRAE, 1999; 2003)Condition (Keene, 1991; 2002)Environment (CIE, 1995; Larsen, 1996)Risk (Waller, 1994; 2003)Value (Delta plan, Van der Reyden 1996)
Different kinds of institution will have different aims and requirements in terms of assessment and control
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Expected,Acceptable
Very good control
Excellent control
Sensor is not
responding
Sensor is not
responding
External store
Could be better
Expected,Acceptable
Very good control
Excellent control
Sensor is not
responding
Open structure
Poor control
Could be better
Expected,Acceptable
Very good control
Excellent control
Historic housemuseum
Problem with control
Poor control
Could be better
Expected,Acceptable
Very good control
Purpose built museum
Serious problem
Problem with control
Poor control
Could be better
Expected,Acceptable
Archive
54321Calibrated levelsType of
building
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Types of institution -11 Archive
Climate controlled and air is purified Possibly more of a storage than exhibition spaceCollections very vulnerable or valuableA realistic optimum
2 Purpose Built MuseumEnvironmental control is possible - Thomson’s Class 1 museum Will probably meet the needs of human comfort and existing standards for organic objectsConsistent with existing ideas of good preservation
3 Historic House museumHistoric buildings and museums with limited resources -Thomson’s Class 2 museum. Most organic objects would be well preserved in this environment
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Types of institution - 2
4 Open structureOpen display when the environment is difficult to control May have open windows or little UV protectionFor robust organic objects but not valuable or vulnerable Just outside levels that might be recommended for a museum object
5 External store with no controlVery little control or protection. Environment is open to the outdoors in some respects and provides shelter rather than environmental control. Not to be aspired to but indicates location’s performance
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Existing Preventive
Conservation Methods
Environmental MonitoringCondition AssessmentLocation AssessmentRisk Assessment
Vantaa, 2000
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Strategy
Only required data is analysed
Methods are integrated
Diagnostic
Reduce uncertainty in each method
Does not assume what data exists (or ignore existing data)
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Interpreting the dosimeter
1
2
3
4
5Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.
Identified hazard causing damage?
Carry out risk assessment, location assessment, monitor environment look at collection.
Once value is assessed, prioritise mitigation.
Responding to identified hazard, possibly a long-standing problem. Diagnostic monitoring recommended.
Location Risk
EnvironmentCondition
Yes
No
Check existing data.
1
2
3
4
5
Expectation
Archive store
Purpose built museum
Historic house museum
Open display
Outside store with no control
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Trøndelag Folk Museum ExampleNew, purpose built museum. HVAC system. Category 2
(purpose built museum gallery)
Environment – ‘Class 1’ museum. Monitor temp, RH and light. No natural light (UV)
Object condition – No visible deteriorationLocation – Checked regularly, no problemsRisk – No risk assessment carried out but major
environmental threats perceived
1121112C
1232112B
O3SO2NO2TWPIUVLightT/RHSite
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
1
2
3
4
5
Identified hazard causing damage?
Carry out risk assessment, location assessment, monitor environment look at collection.
Once value is assessed, prioritise mitigation.
Responding to identified hazard, possibly a long-standing problem. Diagnostic monitoring recommended.
Location Risk
EnvironmentCondition
1
2
3
4
5
Expectation
No
Yes
Check existing data.
Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
1
2
3
4
5
Identified hazard causing damage?
Carry out risk assessment, location assessment, monitor environment look at collection.
Once value is assessed, prioritise mitigation.
Responding to identified hazard, possibly a long-standing problem. Diagnostic monitoring recommended.
Location Risk
EnvironmentCondition
1
2
3
4
5
PPO response (gallery)
Check existing data.
Yes
No
Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Expected,Acceptable
Very good control
Excellent control
Sensor is not
responding
Sensor is not
responding
External store
Could be better
Expected,Acceptable
Very good control
Excellent control
Sensor is not
responding
Open structure
Poor control
Could be better
Expected,Acceptable
Very good control
Excellent control
Historic housemuseum
Problem with control
Poor control
Could be better
Expected,Acceptable
Very good control
Purpose built museum
Serious problem
Problem with control
Poor control
Could be better
Expected,Acceptable
Archive
54321Calibrated levelsType of
building
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Disagreement between assessments
Risks are synergistic OR unmonitored ‘type 3’ risk OR present no loss of value
Risk
Problem is with control, not location, OR problem is very subtle
Location
Early warning of future condition by chemical deterioration
Condition
Response on EWO sensor is synergistic, OR a hazard that is not monitored
Environ
EWO-G
RiskLocationConditionEnvironEWO-G
Negative
Positive information
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Agreement between assessments
\\\ POSITIVE\\\
\\\\\\
NEGATIVE \\\
Risks are unpredictable, OR damage has occurred outside accepted levels
Deterioration is not visible and not expected
Risks are not environmental, or not measured. Check for pests etc.
Risk
\\\POSITIVE
\\\\\\
\\\NEGATIVE \\\
Location is not a cause of deterioration. Inspect data and services
Risk is not environment, or not monitored. Assess risk and condition
Location
\\\ POSITIVE\\\
\\\NEGATIVE \\\
Risk is not environment OR not a monitored hazard. Refer to risk assessment
Condition
\\\ POSITIVE\\\
\\\NEGATIVE \\\
Environ
\\\ POSITIVE\\\
\\\NEGATIVE \\\
EWO-GRiskLocationConditionEnvironEWO-G
Negative
Positive information
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Interpretation
Pollutants are the only hazards not monitored. None of the monitored hazards reveal a problem
We can deduce that pollutants are the problem
Pollutant gases are undetected by the museum and the EWO dosimeter reveals the problem
Diagnostic monitoring would reveal that NO2 and O3levels are higher inside than outside (during the field test)
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Pollutants identified
1
2
3
4
5
Identified hazard causing damage?
Carry out risk assessment, location assessment, monitor environment look at collection.
Once value is assessed, prioritise mitigation.
Responding to identified hazard, possibly a long-standing problem. Diagnostic monitoring recommended.
Location Risk
EnvironmentCondition
1
2
3
4
5
Check existing data.
Yes
No
Problem may be synergistic, or an unmonitored element of the environment. Assess unmonitored hazards.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Diagnostic monitoring
EWO-S dosimeter can be used for assessments of external pollution, as a diagnostic monitoring campaign
Shorter exposure time of twenty eight days suitable for investigating problems
Often based on hypothesis testing, monitoring different locations
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Mitigation
What to do when a hazard has been determined
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Risk chain
release
exposureattack consequence
Risk assessment
Dosimetry Condition assessment
Environmental monitoring
Part of this is determining which points are critical for hazards to have an effect on a collection.
There are various ways of assessing and mitigating damage which relate to different stages of the process.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Dependencies in risk
Outcomes are dependent on a series of prior events, which can be generalised.
What needs to take place for damage to happen?
What are the critical points and pathways in this chain of events?
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Dependency modelling
These are deductive, top-down methods of analyzing risks in system design.
It involves specifying a ‘top event’ to analyze (damage).
Followed by identifying all of the associated elements in the system that could cause that ‘top event’ to occur.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93Damage to object from
pollution
Exposed surfaces for deposition
Interaction of pollutant with
collection
Presence of pollutant in
building
No attractive deposition surfaces
Generation of pollutant internally
Infiltration through natural
ventilation
AND
AND
OR
Infiltration through HVAC
An example…
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Relationships in model
In a positively phrased dependency model, ANDdependencies are points of weakness because all events need to occur for the higher event to take place.
OR dependencies represent points of strength, because of alternatives options. Can turn these into ANDs through investigation.
Probabilities can be applied to each event, so cost effectiveness and efficiency can be determined for any action.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Modelling deterministic risk
Events often about extent of impact, not presence or absence of impact. Deterministic risks cannot be modelled this way.
The threshold levels can be used to create ‘steps’ for each event, so pathway has defined levels.
E.g. presence is NO2 at 10 ppb, rather than 5ppb. Steps based on object deterioration and MASTER calibration.
MA
STER
MA
STER
UCL Presentation for MASTER workshop, January 2006
EUK4
-CT-
2002
-000
93
Damage to objects e.g. 5ppb NO2 for one year
Presence of pollutant in
gallery
Presence of pollutant in building
Infiltration from outside
External presence of
pollutant
HVAC Filter
No scavenger in display area
Intake filtration
Intake position
Internal generation
Reaction from NO
Unfluedheating
appliance
Cellulose nitrate
break down
No attractive deposit surfaces
AND
AND
OR
OR
OR
OR
4ppb
6ppb
AND