QRA Decision Support Model for Locating HazMat Teams

1
QRA Decision Support Model for Locating HazMat
Teams

• Ghada Hamouda, Frank Saccomanno, Liping Fu
• Department of Civil Engineering
• University of Waterloo

2
Background

• Transportation of HazMat poses special risks for
  population and the environment.
• Effective location of HazMat response teams can
  play an important role in reducing risks.

3
What is a HazMat team?

• Refers to responders who are specially trained
  and equipped to manage and control incidents
  involving different types of hazardous materials
  (US Dept of Labor).
• Involves a certain degree of multi-tasking, but
  central focus is on managing HazMat risk

4
Current practice

• HazMat teams housed in existing fire stations
  (multi-tasking). Not all fire stations.
• Decision is made at local level
• Focus is on covering high population
  concentrations (usually near towns and cities)
• HazMat transported in more remote areas (over
  90 of the regional highway network in the US
  and Canada is rural)
• Tendency not to cover marginal locations (poor
  coverage, high risk and cost)

5
Objectives

• Develop a risk-based DS model for locating HazMat
  teams on regional road network
• Illustrate the usefulness of the model through
  case study application in SW Ontario

6
Risk-based DS

• Risk defined
• Rkr Frqkr Csqkr
• Expected fatalities resulting from HazMat
  incidents at different segments of the highway
  network for a given period of time (accident
  induced).
• Csqkr output from time-dependent QRA
  analysis (US EPA ALOHA model)
• Frqkr output from accident/release prediction
  model or look-up table
• In study we do not consider long term health
  and env. impacts.

7
Schematic of HazMat team location

Network node
(
j
)
-
potential
demand for
Fire station (
i)

-

HazMat team

potential site for
HazMat team

area

Hazard
8
Location Heuristic
Find optimal location of np HazMat teams among nf
candidate nodes (fire stations), such
that Minimize total network risk (Expected
Fatalities/Yr)
Ensure that maximum response time and risk at all
nodes does not exceed some pre-set threshold
(Tmax Rmax)
9
Model framework

Choose number of
Hazmat
unites

Locate HazMat teams

Calculate response times from
Hazmat teams to nodes

I
Calculate time
-
dependant
societal risk on the region

No

Is location strategy
acceptable?

Yes

Stop

10
Case Study SW Ontario Region
11
Estimating HazMat flows
HazMat accounts for 9.9 of total truck traffic
in the region (DGAIS). Applied to AADT Trucks
to yield link HazMat flow on SW Ont
network. Estimate flows by HazMat type from
DGAIS percentages
12
Truck accident rates (accidents/MVKm)
13
Accident induced release probabilities by HazMat
type ()
 
14
Relevant Location Issues

• How many teams to locate region-wide
• Implications of current vs min risk strategy
• Implications of specific closure decisions for
  current strategy
• Implications of locating teams one at a time

15
Total risk versus no. of HazMat teams
Point of inflection (Four Teams)
Feasible Solutions
16
Current vs Optimum (4 HazMat Teams)

• Current strategy teams at nodes 6, 12, 15 and 22
• Risk minimum strategy at nodes 6, 14, 22 and 27
• (Nodes 12 and 15 replaced by 14 and 27)
• Small difference in total risk. Maximum resp.
  time and risk same.

Total risk

Max risk

Max
Fatality/year

Fatality/year

resp
onse
time (min)

Current
0.0710

0.0105

58.8

locations

Optimal
0.0666

0.0107

58.8

locations


17
Close one HazMat team (reduce costs from current
strategy)

• Best option to close Node 6 (Hamilton) results in
  lowest risk increase of 2.9 over the previous
  option
• Next best closure of Node 12 (Burlington). More
  practicable (risk increase of 3.9).

18
Location options for different network HazMat
team totals
19
Recommendations

• Extend analysis to consider cost of new HazMat
  teams
• Incorporating multi-tasking
• Permit incidents on links
• Incorporating personal injury and property
  damages (short and long term)