Incident Response Mechanism for Chemical Facilities

By Stephen Fortier and Greg ShawGeorge Washington University, Institute for

Crisis, Disaster and Risk Management

20 September 2010

6th Dresden SymposiumHazards – Detection and Management

September 20-24, 2010 Dresden, Germany

Research Problem

• Chemical facilities are an integral part of the critical infrastructure of the USA and remain vulnerable to natural and man-made threats

• Response to incidents that could lead to catastrophic losses run the gamut from pre-defined to ad-hoc.

• Related Research in emergency response (Turoff, et.al, 2002), incident command systems (Choi et al, 2004), data modeling (Chen, et.al, 2008)

• There is no standard or defined incident response mechanisms for the chemical industry• Vetted by members of the ChemITC and discussions

with DHS Critical Infrastructure Protection personnel

Goal of the Research Project

• Examine the current state-of-the-art of response mechanisms for chemical facilities and provide a “normalized” view of current practices

• Propose a “to be” view of an efficient response mechanism for a chemical facility. There are existing and evolving technologies that could be used for the response mechanism when emergencies occur at chemical facilities and this research will determine what is being used in the industry

• Model and simulate the utilization of various technology to determine the potential benefit and utility for the response mechanism

Research Questions• Research Question 1. What are the essential

elements of and effective and efficient chemical facility emergency response mechanism?

• Research Question 2. Does the size of a chemical facility influence its ability to provide an effective response to a threat?

• Research Question 3. What technologies, specifically decision support systems, could be utilized to improve the chemical facility emergency response mechanism?

• Research Question 4. How do you optimize the technologies used in emergency response to provide an range of improved solutions?

Methodology

Collect RMS* Data

Collect CFATS

Data

Analyze Data

Interview Experts in Chemical

Facility Emergency Response (CFER)

Mechanisms

Refine Research on SOTA IT for Insertion into

CFER Mechanism

Create System Process Models, “as is” view of a set of CFERs

Develop “Normalized” and

“to be” Models of a Standard CFER

Insert SOTA IT into Model and run

Simulations

Analyze Output to CFER IT trade-offs

and Finalize Dissertation

Develop Simulation Model

• Collect Data• Interview Experts• Review

Government Databases

• Create Process Models

• Refine SOTA Research

• Normalize Models• Develop

Simulation Model• Test and Validate

Research

Question 1

Research

Question 2 Research

Question 3

Research

Question 4

Elements of Incident Response

• What is the identification of the chemical? Is the release a result of a chemical interaction?

• Is the chemical an EHS or CERCLA hazardous substance (CFATS, 2007)?

• What quantity was released?• What was (or is) the time and duration of the release?• What was the release mode? Waterborne or airborne?• Is there any acute or chronic health risks associated

with this release?• What medical information is available for exposed

individuals?• What part of the local community could be impacted by

the release?• Should the effected area shelter in place or evacuate?

Develop IDEF0 Models

Modeling, Simulation and Incident Management System Tools

• Hazard Prediction and Assessment Capability (HPAC)

• Consequence Assessment Tool Set-Joint Assessment of Catastrophic Events (CATS)

• Computer-Aided Management of Emergency Operations (CAMEO)

• Area Locations of Hazardous Atmospheres (ALOHA)

• National Atmospheric Release Advisory Center (NARAC)

• Chemical Biological Response Aide (CoBRA®)

Building the Simulation Model• Problem Analysis and Information Collection• Data Collection• Model Construction• Model Verification• Model Validation• Designing and Conducting Simulation

Experiments• Output Analysis• Final Recommendations

Draft Simulation Model

Input Analysis• If system exists, then requisite empirical data

from field measurements will be used• Otherwise, will use expert judgment including

intuition, past experience with other systems or educated guess

• Use heuristics including:• Random variables modeled as deterministic

quantities• Unknown distributions with partial know functional

form that incorporate available partial information• Past experience that provides information in the

functional form of distributions

Data Analysis• Conduct preliminary analysis to determine

model fit• From the collected data, compute various

empirical statistics related to:• Moments (mean, standard deviation, coefficient of

variations, etc.)• Distributions (histograms)• Temporal dependence (Autocorrelations within an

empirical time series, or cross correlations among two or more distinct time series)

Output Analysis• Replication Design. Goal to obtain statistical

information form simulation runs in an efficient method

• Estimation of Performance Metrics. Provide the data for computing point estimates and confidence intervals for system parameters of interest.

• System Analysis and Experimentation. Statistical estimates will be used to understand system behavior and generate performance predictions under various scenarios

Preliminary Findings

• The effects of 9/11 and protection of propriety data makes it difficult for chemical facilities to share data. This contrasts with the US laws that guarantee the “citizen’s right to know”

• Although required to have emergency response plans, a number of chemical facilities handle incidents ad hoc

• Business process modeling is assisting in formalizing ad hoc incident response procedures

• Reliance on Local Emergency Planning Committees (LEPCs)

Control is Key

• The goal of emergency responders at chemical facilities is to provide control of an emergency or disaster situation. As control is lost during a disaster situation, the consequences of the event escalate rapidly

Time

0%

Control Severity

Loss of Control

100%

Conclusion• This research project will analyze the current

practices for emergency response for chemical facilities and provide a model that could be utilized by small or large entities

• Provide statistical estimates that will predict the range for an improved solution for an emergency response mechanism

• Determine the impact or measures of effectiveness of a response to a harmful event, depending on the size of the chemical facility

• Determine what technologies, specifically decision support systems, could be utilized to improve the chemical facility emergency response mechanism