Risk Assessment using Scoring System Methodology (SSM) as holistic substitute for IBC Height

1
Risk Assessment using Scoring System Methodology
(SSM) as holistic substitute for IBC Height
Area Table

• Frank Noonan
• Professor of Industrial Engineering Fire
  Protection Engineering
• Worcester Polytechnic Institute
• For AFSCC, 8/3/07, Chicago, Il.

2
The challenge

• HA Table overly prescriptive?
• Incomplete with respect to assessing all relevant
  factors that make a given building size safe or
  unsafe.
• Goals
• Holistic approach.
• Acceptable level of fire safety for communities,
  occupants, emergency responders.

3
Scoring System Methodology

• Widely used throughout global economy for risk
  assessment.
• Pioneered by banking industry to assess
  likelihood for default with loan applications.
• Other applications
• IRS audit decisions
• Parole board decisions
• Mass mail and telemarketing decisions
• Insurance industry customer screening
• Hospital patient procedure decisions

4
Scoring Model Example
Risk Index
Explanatory Level
Factors Data
Very Poor Risk Poor Risk Medium Risk Relatively
Good Risk Good Risk Very Good Risk
-Personnel Experience -Material Handling
Processes -Managers Experience -Accident
History
Scoring Model to estimate a plants risk for a
toxic spill.
5
Scoring Models

• A type of Inference Tree which assesses a
  non-observable variable through other data which
  can be observed or measured
• (e.g. the probability of some negative event or
  the effectiveness of some complex system).
• Purpose Replace Individual Judgment with a more
  reliable method.
• Risk Index w(1)RF(1) ... w(n)RF(n)

6
Risk Factors
7
Risk Factors
8
(No Transcript)
9
Contract to Build a Scoring Model as alternative
to IBC Height Area Table 503.

• Frank Noonan reporting to Kate Dargan and Dave
  Collins
• Phase 1 Work with the BFP Features Study Group
  as SMEs to build a prototype model.
• Phase 2 Continuing with BFP Study Group open
  model to a larger group of building safety
  professionals to review and fine tune the model.
• Timeframe 3/8/07 2/28/08

10
SSM development for IBC

• 1. Establish SMEs and communication framework.
• 2. Define risk metric for building safety
  performance.
• 3. Define set of risk factors (and scales) for
  risk assessment,
• (e.g. occupancy classification, height,
  area, full menu of active and passive risk
  mitigation measures, expected emergency
  response).
• 4. Define relative importance of each risk
  factor.
• 5. Calibrate Scoring Model for Acceptable Risk on
  Building Safety Performance.

11
Establish SMEs and communication framework.

• SMEs Study Group Facilitator Frank Noonan
• MyWPI is internet forum
• Need name and email address of each SME
• Site Name BFP Features Study Group
• Discussion/ Voting/ Finding Consensus

12
2. Define risk metric for building safety
performance

• Narrative describing Building Safety (or risk
  to building safety)
• Sets transition for defining risk factors.
• Process 1st draft today (8/3/07) post for
  revisions complete by 8/24/07.
• Risk Metric w(1)RF(1)w(n)RF(n)
• Scale 0,10
• GAR GGGGGAAARRRRR

13
Risk Metric Narrative

• Building Safety is defined as the aggregate
  effectiveness of the mitigation features ((Active
  and Passive and Emergency Response) in a building
  that are provided to protect the structure,
  occupants, emergency responders, and property
  from losses associated with anticipated hazards
  primarily due to fire exposure and subsequent
  collapse.
• Building Safety Success can be defined as
  meeting the goals for the reduction
• of life and property loss that are
  acceptable and economically supportable. This
• is the core concept of acceptable risk.
  Acceptable building safety risk is that level of
  anticipated loss that each entity impacted can
  accept if a hazardous event occurs. It is based
  on the probabilities that the various mitigation
  strategies will perform as intended and it can be
  measured as a quantitative value, a qualitative
  value, or both.
• Identify and analyze a buildings hazards and the
  potential mitigations of those
• hazards that positively impact building
  safety and also considers their probability for
  successful performance

14
3. Define set of risk factors (and scales)
for risk assessment

• Set of RF complete, clear, avoid redundancy.
• Pursue Hierarchical Structure ( i.e. inference
  tree).
• See example on next slide.
• Process 1st draft today (8/3/07) post for
  revisions complete by 9/5/07.
• Each factor requires a narrative definition and a
  scale.
• Whenever possible use objective scale, defining
  best and worst.
• 0,1 factors are allowed.
• Risk factor constraints are allowed as necessary
  conditions for G or R.

15
Example from U.S. Navy Vessel Refueling
Decision

• Major categories
• Task Difficulty
• Environmental
• Equipment Readiness
• Manpower Readiness

16
Risk Factors

• 1. Exiting
• 2. Compartmentation
• 3. Smoke Management
• 4. Automatic Sprinklering
• 5. Fire-Resistive Construction
• 6. Structural Integrity
• 7. Better Inspection and Maintenance Compliance
• 8. Emergency Response
• 9. Occupancy Type

17
4. 4. Define relative importance of each
risk factor.

• Method of Pair-wise Comparisons
• Rank order the set and use least important as
  anchor to elicit relative importance repeat
  using most important as anchor.
• Make effort to exploit the hierarchy.
• Process
• Voting Average and post results
• Start voting on 9/07/07 post Round 1 results by
  9/18/07 post Round 2 results by 9/28/07.
• Eliminate factors if weight is less than 1 and
  more than 15 factors.

18
5. Calibrate Scoring Model for Acceptable
Risk on Building Safety Performance.

• Create a set of Bldg Design/Operating Scenarios
  where bldg safety performance is on margin with
  respect to Acceptable Risk.
• Each scenario has an SME creator who provides a
  narrative and set of risk factor scores.
• Once a scenario is created other SMEs may weigh
  in on scores and make a GAR vote scenario is
  eliminated if simple majority vote it as G or R.
  
  (cont.)

19
5. Calibrate Scoring Model for Acceptable
Risk on Building Safety Performance.

• Need Scenarios My worry is with too few rather
  than too many.
• Each scenario produces a risk index score and
  their range defines the Amber range in the GAR
  scale.
• Process Start 10/1/07 complete by 11/1/07.19

20
Review Fine Tune Prototype Model

• In Jan/Feb 08, post the BFP Features Study Group
  prototype model for input from larger group of
  building safety professionals.
• Revisions to
• Constraints on risk factors
• Relative weights of risk factors
• Ruling out any scenarios on acceptable risk margin