NONACCIDENT RELEASE RISK INDEX

1
NON-ACCIDENT RELEASE RISK INDEX

• Tom Mitchell
• AMEC Earth Environmental, Inc
• Houston, Texas

2
Non-Accident Release Risk Index

• A Product of the
• AAR Hazardous Materials (BOE)
• Working Committee
• and the
• Non-Accident Release Risk Index
• Task Force
• NAR Task Force Meeting
• Houston, Texas
• April 29, 2003

3
Genesis of the NARRI
TOTAL NON-ACCIDENT RELEASES - TANK CARS
4
NARRI Task Force

• Working group established to develop risk index
  to be used by the rail industry
• Participants AAR, BNSF, CN, CP, CR, CSXT, IC,
  KCS, NS, PWRR, UPRR, AMEC and University of
  Illinois
• Representatives from both BOE and NAR
  Committees

5
NARRI Development Criteria

• Task force agreed that NAR Risk Index must meet
  five criteria to maximize its effectiveness
• 1. Must be meaningful and focus on viable safety
  concerns
• 2. Must be easy to calculate with readily
  available information
• 3. Must be statistically proven as accurate and
  reliable
• 4. Must be easily transmitted to AAR for
  calculation, analysis and dissemination
• 5. Must provide a means of providing uniform
  communications with shippers on NAR issues

6
NARRI Development Results

• Final product designed to measure seriousness and
  potential exposure to human life and environment
• Similar to measurement used by chemical industry
  (e.g., Dow) to evaluate severity of rail accidents

7
NARRI Structure

• Preventability
• Measure of human error involved in creating NAR
• Potential Consequences
• Measure of potential damage to environment had
  the NAR not been detected and fixed
• Actual Consequences
• Measure of the actual damage to environment and
  adverse impact to human activities

8
(No Transcript)
9
NARRI Formula

• A (B C D) (E F)
• -or-
• Prevention (Packaging Product Hazard
  Extenuating Hazard) (Environmental Impact
  Human Impact)

10
NARRI Example
On May 14, 2001, on track 5 in Alta Yard, tank
car BTRX-975680, load of methanol (PGII) was
observed by a switch crew with liquid product on
the top and sides of the car. The car was moved
to an isolation track and the shipper notified
through Chemtrec. The shipper responded to the
yard and upon inspection discovered that product
had leaked from between the manway nozzle and
manway cover plate due to a missing gasket. A
new gasket was installed, the manway cover plate
secured and car authorized for transportation.
Approximately 3 gallons of product was lost.

• 1. Preventative Factor 5 Cause Code 207
• 2. Shipping Packaging Factor 5
• 3. Product Hazard Factor 3 5 ( 5 3 0 ) X
  ( 2 0 )
• 4. Extenuating Product Hazard Factor 0
  
• 5. Environmental Impact Factor 2 NARRI
  80
• 6. Human Impact Factor 0

11
NARRI Risk Evaluation

• Grouped the NARRI Scores according to severity
• Minor Less than the annual median score
• Moderate Between the annual median and one
  standard deviation above the median
• Severe Greater than one standard deviation
  above the annual median

12
NARRI Validation

• Validation process incorporated 6? components
• Conducted sample calculations to validate
  measurement system
• Checked the reproducibility of measurement system
  (Gage RR)
• Cause codes developed to reduce subjectivity
• Developing future validation process to ensure
  the statistical integrity of the program

13
NARRI Program

• AAR manages and distributes industry-wide data
• Program began 1 January 2001 with some recovery
  of historical data
• Shipper NARRI information made available to
  carriers, shippers, AAR and FRA

14
NARRI Data Analysis

• Data Review

15
Tank Car NARs by Hazard Class
12 months ending 9/30/02
16
Average NARRI by Hazard Class
12 months ending 9/30/02
17
Top Commodities for NAR
12 months ending 9/30/02
18
Top Causes Non-Pressure Car
12 months ending 9/30/02
19
Top Causes Pressure Car
12 months ending 9/30/02
20
Top Causes Flammable Liquids
12 months ending 9/30/02
21
Top Causes Corrosive Liquids
12 months ending 9/30/02
22
Total NARRI of Top Commodities
12 months ending 9/30/02
23
Causes with Highest Average NARRI
12 months ending 9/30/02
24
NARRI Data Analysis

• Comparative Analysis

25
NARRI Comparison
26
NARRI Comparison
Nominal NAR involving a loaded tank car. The
cause was judged to be not due to human error.
Less than 10 gallons of product were released.
However, the local emergency responders
over-reacted and ordered an evacuation of the
area.
27
NARRI Comparison
Nominal NAR involving a loaded tank car. The
cause was judged to be not due to human error.
No human impact was involved. Chlorine - lt10
gallons released Nitric Acid lt100 gallons
released