USA: +1(669)289-8683 
Sign in
Resource:
Railroad Accident Report(Attached PDF)
Write a 1,050- to 1,400-word paper analyzing the Graniteville, SC train accident(Attached PDF).
Include the following:
- Describe and summarize the critical incident at Graniteville, SC to include a discussion on manmade critical incidents and risks associated with the railroad operation.
- Discuss how a risk and vulnerability assessment may have prevented this incident.
- List and explain what actual prevention, preparedness and mitigation strategies could have been adopted to avert this disaster.
- Identify the strengths and weaknesses of actions taken by first responders to the scene to include communication processes. Why did the first responders take these actions?
- Provide recommendations for action plan development to prevent reoccurrences.
Format your assignment consistent with APA guidelines.
Collision of Norfolk Southern Freight Train 192 With
Standing Norfolk Southern Local Train P22 With
Subsequent Hazardous Materials Release a
t
Graniteville, South Carolina
January 6, 2005
Railroad Accident Report
NTSB/RAR-05/04
PB2005-916304
Notation 7710A
National
Transportation
Safety Boar
d
Washington, D.C.
National
Transportation
Safety Board
Washington, D.C.
National Transportation
Safety Board
Washington, D.C. 20594
OFFICIAL BUSINESS
Penalty for Private Use, $30
0
PRSRT STD
Postage & Fees Paid
NTSB
Permit No. G-200
RAILROAD ACCIDENT REPORT
NTSB/RAR-05/04 (PB2005-916304)
COLLISION OF NORFOLK SOUTHERN FREIGHT TRAIN 192 WITH
STANDING NORFOLK SOUTHERN LOCAL TRAIN P22 WITH
SUBSEQUENT HAZARDOUS MATERIALS RELEASE AT
GRANITEVILLE, SOUTH CAROLINA
JANUARY 6, 2005
• Page 50, fourth paragraph has been updated with the correct company name (Canadian
Pacific Railway) instead of Canadian National as originally printed. (15 JUNE 2006)
THE CORRECTIONS BELOW ARE INCLUDED
IN THIS VERSION OF THE PUBLISHED REPORT
Railroad Accident Report
Collision of Norfolk Southern Freight Train 192 With
Standing Norfolk Southern Local Train P22 With
Subsequent Hazardous Materials Release at
Graniteville, South Carolina
January 6, 2005
NTSB/RAR-05/04
PB2005-916304 National Transportation Safety Board
Notation 7710A 490 L�Enfant Plaza, S.W.
Adopted November 29, 2005 Washington, D.C. 20594
National Transportation Safety Board. 2005. Collision of Norfolk Southern Freight Train 192 With
Standing Norfolk Southern Local Train P22 With Subsequent Hazardous Materials Release at
Graniteville, South Carolina, January 6, 2005. Railroad Accident Report NTSB/RAR-05/04.
Washington, DC.
: About 2:39 a.m. eastern standard time on January 6, 2005, northbound Norfolk Southern
Railway Company (NS) freight train 192, while traveling about 47 mph through Graniteville, South
Carolina, encountered an improperly lined switch that diverted the train from the main line onto an
industry track, where it struck an unoccupied, parked train (NS train P22). The collision derailed both
locomotives and 16 of the 42 freight cars of train 192, as well as the locomotive and 1 of the 2 cars of train
P22. Among the derailed cars from train 192 were three tank cars containing chlorine, one of which was
breached, releasing chlorine gas. The train engineer and eight other people died as a result of chlorine gas
inhalation. About 554 people complaining of respiratory difficulties were taken to local hospitals. Of these,
75 were admitted for treatment. Because of the chlorine release, about 5,400 people within a 1-mile radius
of the derailment site were evacuated for several days. Total damages exceeded $6.9 million.
The safety issues addressed in the report are railroad accidents attributable to improperly lined switches
and the vulnerability, under current operating practices, of railroad tank cars carrying hazardous materials.
As a result of its investigation of this accident, the Safety Board makes recommendations to the Federal
Railroad Administration.
The National Transportation Safety Board is an independent Federal agency dedicated to promoting aviation, railroad, highway, marine,
pipeline, and hazardous materials safety. Established in 1967, the agency is mandated by Congress through the Independent Safety Board
Act of 1974 to investigate transportation accidents, determine the probable causes of the accidents, issue safety recommendations, study
transportation safety issues, and evaluate the safety effectiveness of government agencies involved in transportation. The Safety Board
makes public its actions and decisions through accident reports, safety studies, special investigation reports, safety recommendations, and
statistical reviews.
Recent publications are available in their entirety on the Web at
may be obtained from the Web site or by contacting:
National Transportation Safety Board
Records Management Division
490 L�Enfant Plaza, S.W.
Washington, D.C. 20594
(800) 877-6799 or (202) 314-6551
Safety Board publications may be purchased, by individual copy or by subscription, from the National Technical Information Service. To
purchase this publication, order report number PB2005-916304 from:
National Technical Information Servic
e
5285 Port Royal Ro
ad
Springfield, Virginia 22161
(800) 553-6847 or (703) 605-6000
The Independent Safety Board Act, as codified at 49 U.S.C. Section 1154(b), precludes the admission into evidence or use of Board reports
related to an incident or accident in a civil action for damages resulting from a matter mentioned in the report.
iii Railroad Accident Report
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Accident Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Site Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Events Preceding the Accident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
The Accident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
Emergency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
Fire Department Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
Sheriff�s Office Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
Hazardous Materials Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
Injuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
7
Damages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Hazardous Materials Cars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Other Railroad Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Emergency Response Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Wreckage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Personnel Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Train P22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Train 192 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Toxicological Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Meteorological Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Operations and Track Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Postaccident Inspections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Switch and Track . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Railcars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Tank Car and Hazardous Materials Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Tank Cars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Hazardous Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Tests and Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Sight Distance Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Event Recorder Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Laboratory Examination and Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Other Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Accidents Involving Improperly Lined Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Postaccident FRA Safety Advisory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
FRA Human Factors Working Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
FRA Emergency Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Switch Position Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Postaccident Action by NS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Hours of Service Act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
The Accident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Exclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Contents iv Railroad Accident Report
Emergency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Performance of Train P22 Crew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Train Speeds in Non-Signaled Territory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Hazardous Materials Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Tank Car Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Tank Car Crashworthiness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Reduction of Tank Car Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Train Crew Protection From Inhalation Hazards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Probable Cause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
v Railroad Accident Report
Executive Summary
About 2:39 a.m. eastern standard time on January 6, 2005, northbound Norfolk
Southern Railway Company (NS) freight train 192, while traveling about 47 mph through
Graniteville, South Carolina, encountered an improperly lined switch that diverted the
train from the main line onto an industry track, where it struck an unoccupied, parked train
(NS train P22). The collision derailed both locomotives and 16 of the 42 freight cars of
train 192, as well as the locomotive and 1 of the 2 cars of train P22. Among the derailed
cars from train 192 were three tank cars containing chlorine, one of which was breached,
releasing chlorine gas. The train engineer and eight other people died as a result of
chlorine gas inhalation. About 554 people complaining of respiratory difficulties were
taken to local hospitals. Of these, 75 were admitted for treatment. Because of the chlorine
release, about 5,400 people within a 1-mile radius of the derailment site were evacuated
for several days. Total damages exceeded $6.9 million.
The National Transportation Safety Board determines that the probable cause of
the January 6, 2005, collision and derailment of Norfolk Southern train 192 in
Graniteville, South Carolina, was the failure of the crew of Norfolk Southern train P22 to
return a main line switch to the normal position after the crew completed work at an
industry track. Contributing to the failure was the absence of any feature or mechanism
that would have reminded crewmembers of the switch position and thus would have
prompted them to complete this final critical task before departing the work site.
Contributing to the severity of the accident was the puncture of the ninth car in the train, a
tank car containing chlorine, which resulted in the release of poisonous chlorine gas.
The safety issues identified in this investigation are as follows:
� Railroad accidents attributable to improperly lined switches;
� The vulnerability, under current operating practices, of railroad tank cars
carrying hazardous materials.
As a result of its investigation of this accident, the National Transportation Safety
Board makes safety recommendations to the Federal Railroad Administration.
1 Railroad Accident Report
Factual Information
Accident Synopsis
About 2:39 a.m. eastern standard time on January 6, 2005, northbound Norfolk
Southern Railway Company (NS) freight train 192, while traveling about 47 mph through
Graniteville, South Carolina, encountered an improperly lined switch that diverted the
train from the main line onto an industry track, where it struck an unoccupied, parked train
(NS train P22). The collision derailed both locomotives and 16 of the 42 freight cars of
train 192, as well as the locomotive and 1 of the 2 cars of train P22. Among the derailed
cars from train 192 were three tank cars containing chlorine, one of which was breached,
releasing chlorine gas. The train engineer and eight other persons died as a result of
chlorine gas inhalation. About 554 people complaining of respiratory difficulties were
taken to local hospitals. Of these, 75 were admitted for treatment. Because of the chlorine
release, about 5,400 people within a 1-mile radius of the derailment site were evacuated
for several days. Total damages exceeded $6.9 million.
Site Description
Graniteville, South Carolina,1 is an unincorporated community of about 1,200
(1990 data) in a mixed rural and suburban area of Aiken County. Graniteville is about 4.5
miles west of the commercial/retail district of the city of Aiken, South Carolina, and about
9.8 miles northeast of the commercial/retail district of
Augusta
, Georgia. (See figure 1.)
Principals of local emergency response agencies estimated that 5,400 people live
within a 1-mile radius of the accident site; this includes peripheral areas of adjacent
communities. The fire district of the local responding fire and rescue agency (the
Graniteville, Vaucluse, and Warrenville Volunteer Fire Department) has an approximate
population of 22,000, which includes communities adjacent to Graniteville that were not
directly affected by the incident.
NS main line track runs in a north-south direction through the center of
Graniteville. To the east of the main line and extending in a southerly direction is an area
of primarily residential properties. To the west and extending in a northerly direction is an
area comprising several moderate- to large-sized industrial plant facilities, some of which
operate continuously. A small commercial/retail district is adjacent to and north of the
main line right-of-way (about 1,000 feet from the accident site). Two principal north-south
thoroughfares run parallel to and on either side of the NS main line. These thoroughfares
are Canal Street2 to the west and Trolley Line Road to the east.
1 Local maps identify a location proximate to Graniteville as New Hope, but local residents consider
the locality to be Graniteville.
2 Canal Street is also known locally as Main Street.
Factual Information 2 Railroad Accident Report
Graniteville lies in a shallow valley that encompasses a stream, known as Horse
Creek, which is about 1,000 feet west of the NS main line and somewhat parallels it. The
terrain adjacent to Horse Creek generally gains elevation as it extends to the east and to
the west of the streambed. The Horse Creek streambed is about 190 feet above sea level.
The main line track at the accident location is about 225 feet above sea level, with the
terrain moderately increasing in elevation as the track extends toward the north. The
industry track elevation is also about 225 feet at the turnout, with the elevation moderately
decreasing as the track extends north and west toward the plant.
The industry track where the accident occurred serves facilities of Avondale Mills,
Inc., a manufacturer of textiles. (See figure 2.) The facilities include the mill�s Woodhead,
Hickman, and Gregg Division plants as well as the Stevens steam plant. The Avondale
Mills industry track (hereinafter referred to as �industry track�), upon diverging from the
NS main line, immediately traverses Canal Street at a roadway at-grade crossing and
Figure 1. Accident location.
Columbia
Greenville
Anderson
Rock Hill
North Charleston
SOUTH CAROLINA
NORTH CAROLINA
GEORGIA
Summit
Aiken
Augusta
Graniteville
Warrenville
Railway System
Norfolk Southern Facilities
Graniteville
DIAGRAM NOT TO SCALE
Aiken
Warrenville
Augusta
Factual Information 3 Railroad Accident Report
continues in a northwesterly direction. The track then traverses Hickman Street at a
roadway at-grade crossing before continuing in a northwesterly direction toward several
industrial facilities and additional track turnouts. The portion of the industry track between
the Canal Street grade crossing (edge of pavement) and the Hickman Street grade crossing
(edge of pavement) measures about 328 feet in length.
Events Preceding the Accident
NS local train P22 (full designation P22P005) operated Monday through Friday
out of the NS yard at Aiken, South Carolina. The train crew (engineer, conductor, and
brakeman) normally went on duty at 7:00 a.m. and spent the day placing and picking up
cars at local industries along the main line.
On January 5, 2005, the day before the accident, the regularly assigned train P22
conductor was on vacation. During his absence, the train�s regularly assigned brakeman
was working as conductor, and the brakeman�s job was filled from the extra board.3 Both
the conductor and brakeman went on duty in Aiken at 7:00 a.m. on January 5. The
regularly assigned engineer for train P22 had taken the day off, and his job was also being
Figure 2. NS main line and Avondale Mills facility industry track in Graniteville. �P.S.�
identifies the point of switch from the main line onto the industry track.
3 The extra board is a list of qualified employees available to work in the absence of a regular
employee or to work unscheduled assignments. At the time of the accident, temporary job vacancies on train
P22 were filled from among extra board employees working out of Columbia, South Carolina, about 58
miles from Aiken.
Fence Line
Trolley Line Road
Canal Street
A
ik
en
Ro
ad
A
ik
en
R
oa
d
Hickman Street
65
’
65’
H
ar
d
S
tr
ee
t
C
hu
rc
h
0’ 50‘ 100’ 150’ 200’
Industry Gate
N
P.S.
NS Right of Way (Property Line
)
NS Right of Way
Parking LotHickman Mill
Building
Data Processing Offi
ce
Avondale Mills
Source: NTSB
Factual Information 4 Railroad Accident Report
filled from the extra board. Because of rest requirements,4 the only available extra board
engineer could not begin work until 8:32 a.m.
The NS trackage in this area is non-signaled, and authority to use various track
segments is via track warrants issued by the train dispatcher in Greenville, South Carolina.
At 8:11 a.m., the train P22 conductor was given a track warrant authorizing train P22 to
occupy and work on the east-west main line (designated the SA line) between milepost
(MP) SA51.0 and Warrenville, South Carolina (MP SA63.4). The SA main line connects
Aiken with the north-south main line (designated the R line) on which the accident
occurred. (See figure 3.)
The brakeman said that after the engineer arrived, the conductor held a job briefing
with the crew in which he advised them of specific movements that were to be made
during the shift.5 After the briefing, the crew boarded its locomotive, coupled to eight
freight cars, and departed the Aiken Yard.
4 Required off-duty time between assignments, generally referred to as �rest,� is governed by Federal
regulations and union agreements.
Figure 3. NS main lines in the accident area. Mileposts are referenced by letter and
number designations.
5 At the time of the accident, the Federal Railroad Administration had recommended, but had not
required, that railroads mandate such job briefings. The NS Safety and General Conduct Rules required that
job briefings be held �to review the planned itinerary, procedures, and necessary safeguards for the task to be
performed.� The rule specified that such a briefing �must always precede the task at the work site, be clearly
understood, and be updated or modified as conditions change.�
COLUMBIA, SC
Bridgestone
Vaucluse Siding
R 175
Warrenville Siding
R 179.3
SA 63.45
AIKEN
SA 57.29
Point of Switch
R 178.3
AB 17.3
AUGUSTA, GA
R 109.0
R 171.8
R 191.4
N
POINT OF
COLLISION
Avondale Mills /
Gregg Industry
AB 23.75
SA 51
SA-Line
R-Line
AB-Line
GRANITEVILLE, SC
Factual Information 5 Railroad Accident Report
The crew spent the first part of the day servicing industries along the SA line
before returning to the Aiken Yard about 1:00 p.m. for lunch. After lunch, the crew re-
boarded the train, and about 2:10 p.m., the conductor received a second track warrant
authorizing the train to occupy and operate on the R main line track between MP R185.0
and MP R171.0. Train P22 then proceeded east to the junction of the SA and R main lines.
During the afternoon, the crew worked various industries along the R line from
Bath, South Carolina, (MP R185.0) north to the Bridgestone industrial lead (MP R171.8).
The engineer said that about 4:00 p.m., the crewmembers realized that they would be
�pressed for time� to complete their planned work. The crewmembers completed their
duties at the Bridgestone lead at 5:50 p.m. After leaving Bridgestone, the crew intended to
place two cars of sodium hydroxide solution (caustic soda) at the Avondale Mills plant at
MP R178.3, after which they planned to proceed south to Warrenville (MP R179.3) and tie
down the train for the night. The conductor had arranged with a taxi service under contract
to the NS to have a taxi waiting at Warrenville at 6:15 p.m. to take the crew back to Aiken.
Train P22, now with 12 freight cars, reached the industry track about 6:10 p.m.
According to the engineer, although the conductor and brakeman had conducted job
briefings at other work sites during the day, no job briefing was held before the work
began at the Avondale Mills plant. When he was later asked why he believed no job
briefing was held, the engineer said, �It could have been they [the conductor and
brakeman] were in a hurry.�
The train stopped on the main track with its locomotive about 6 car lengths north
of the main track switch for the industry track.6 (See figure 4.) The conductor said that he
told the brakeman to bring in the train on the industry track, after which the conductor
dismounted the locomotive and walked west to open the facility gates and make sure the
tracks were clear. (See figure 5.) The conductor also lined the industry track switches for
the move. The train then pulled south, stopping just north of the switch, at which point the
brakeman dismounted. The train pulled south again and stopped when the last car was
clear of the industry track switch. The brakeman lined and locked the switch7 (see figure
6) to allow the train to move from the main track onto the industry track, then walked back
to flag the first road crossing. After the conductor had prepared the route, the engineer
shoved the train from the main line onto the industry track. Once the train occupied the
first crossing, the brakeman radioed the engineer to stop. The brakeman mounted the
leading end of the last freight car, and the train shoved back again, with the conductor
flagging the next road crossing.
6 The information in this section regarding crew actions and train movements is based on postaccident
interviews with the crewmembers.
7 Main line switches have locks that deter tampering by unauthorized persons. Railroad employees
have keys for the locks, and they are expected to relock a switch after it has been repositioned.
Factual Information 6 Railroad Accident Report
Figure 4. Looking north at the switch and turnout from the NS R main line onto the
Avondale Mills industry track at Graniteville. The switch banner, or target, shows white,
indicating that the switch is in the �normal� position, that is, aligned for movement along
the main line track. (The banner shows red when the switch is set for the industry track.)
North of the turnout and to the right of the main line is the wayside light indicating the
position of the switch for the Vaucluse Siding.
Figure 5. The Avondale Mills industry track and gates to the Avondale Mills facility at
Graniteville.
Factual Information 7 Railroad Accident Report
Train P22 had to pull out and set aside two empty cars before it could place the two
loaded caustic soda cars at the facility. (According to the engineer, these moves did not
require that the train pull out as far as the main line, and no part of the train moved over
the main line switch after the train was backed onto the industry track.) The conductor said
that he was aware that the Federal hours-of-service limitation of 12 on-duty hours8 would
occur for himself and the brakeman at 7:00 p.m., and he told the engineer and the
brakeman that they would not have time to continue to Warrenville as planned. At
6:20 p.m., he radioed the taxi driver waiting at Warrenville and told him to come to the
industry track (about 1 mile north of Warrenville) to pick up the crew.9
After placing the two caustic soda cars at the Avondale Mills plant, the crew
prepared to leave the train on the industry track for the night. Because of the length of the
train and the need to avoid blocking the road crossings or the plant gate, the crew had to
uncouple 8 cars from the 10-car train and place them on two tracks within the plant gates.
The engineer then moved the locomotive and the remaining two empty freight cars onto
the industry track between the main line and the industry gate, stopping the locomotive in
the clear about 5 to 6 car lengths from the main line switch.10 About this time, the taxi
arrived and parked on the grass on the east side of the industry track near the locomotive.
Figure 6. The switch stand for the Avondale Mills turnout switch at Graniteville showing
the switch locked in position.
8 Title 49 Code of Federal Regulations Part 228.
9 The vehicles used by the contracted taxi service have radios that allow them to communicate with
train crews.
10 The length of train P22 as parked on the siding was about 172 feet. Based on this measurement, crew
statements, and other documentation obtained during the investigation, the head end of train P22 was
estimated to be about 342 feet from the switch stand for the main line track switch.
Factual Information 8 Railroad Accident Report
The engineer applied the locomotive handbrake and began shutting down and
securing the locomotive. The brakeman said that he closed one of the two sides of the
swinging gate at the entrance to the plant and walked toward the locomotive to tie down
the train and retrieve his personal belongings. He said they had been �pushed for the
[hours-of-service] law� as they worked, and he noted the time and was �in a hurry to get
everything done that I needed to do.� The brakeman said he looked at his watch and noted
that it was about 6:59 p.m. when he reached the locomotive. The brakeman set handbrakes
on the train and retrieved his bag from the locomotive.
Meanwhile, the conductor noted that it was about 6:57 p.m. when he closed the
other side of the plant gate and began walking toward the locomotive. He said he
remembered thinking to himself, �Lord, mission accomplished. Everybody�s happy. Let�s
get our stuff and go in.� He said he saw the brakeman at the locomotive when he arrived.
He said he retrieved his clipboard and bag and placed them in the back of the taxi, then
went back to help the engineer with his bag. The three crewmembers then got in the taxi,
and the taxi departed.
During the time the crew was working at the Avondale Mills facility, the main line
switch remained lined for the industry track. NS Operating Rule 104, in effect at the time,
required that the train P22 crew return the switch to its �normal� position (lined and
locked for the main line) when the work was complete. Rule 104(a) states that, while the
position of a switch is the responsibility of the employee handling it, �this�does not
relieve other crewmembers of the responsibility if they are in place to observe the position
of switches.�
In postaccident interviews, the brakeman said he was aware that whenever a job is
completed, any main line switch that was used must be lined and locked for the main
track. He said,
in my mind, when I left [the industry track], everything was properly lined back to
the main line. I had no doubts in my mind when I left there.
He also said, �I am not 100 percent sure that I did [reline the main line switch]. I
would say I might have made a mistake.�
The engineer said that he was on the �off side� of the locomotive and could not see
the main line switch while working on the industry track. He said that he did not know
�how the brakeman handled it.�
The conductor stated:
While we [were] working in the plant, I was off in the field and I never got near
the switch�. We usually have job briefings, but I never told my brakeman to
make sure that the switch was lined and locked for the main line movement. I
never told him, and I never touched the switch myself. It was in my mind when I
arrived at the mill. I should have done a briefing to ensure that the switch was
lined back.
Factual Information 9 Railroad Accident Report
The conductor also stated that because the industry track was curved, the main line
switch banner11 was not illuminated by the locomotive headlight while the train was in the
industry track, nor was it otherwise visible to the crewmembers while they worked.
The conductor said he had no conversations with the engineer or the brakeman
about the way the switch was lined. The conductor also said, and the taxi driver
confirmed, that the subject of the switch did not come up during the ride back to the
terminal. The conductor said, �Once we got in that taxi, we went back toward the depot;
and in my mind, everything was all right.� As it left the area, the taxi traveled south along
Canal Street and over a grade crossing within 21 feet of the main track switch, but neither
the driver nor his passengers noted its position.
The taxi arrived at the Aiken Yard office at 7:15 p.m. After the crewmembers
gathered their belongings, the engineer departed the yard while the conductor and
brakeman proceeded to the yard office. About 7:50 p.m., as the conductor was completing
his paperwork, he asked the brakeman to contact the train dispatcher and clear the two
track warrants that had been issued to train P22. The brakeman cleared the two track
warrants with the dispatcher at 7:53 and 7:54 p.m., respectively. The brakeman said later
that he would not have cleared the track warrants for the conductor if he had thought he
had not left everything lined properly at the industry track.12 The conductor and brakeman
completed the paperwork at 8:11 p.m.13
The Accident
The train dispatcher on duty when local train P22 finished its work said that no
trains occupied the main line track in the Graniteville area from the time the local�s track
warrants were cleared at 7:54 p.m. until he went off duty at 11:00 p.m. The dispatcher who
came on duty at 11:00 p.m. said that he gave no authority for train operations over that
section of track until after 2:00 a.m. on January 6, 2005, when he issued a track warrant
for NS freight train 192 (full designation 192P005) to operate from Augusta, Georgia (MP
R191.4), to Summit, South Carolina (MP R132.8).
Train 192 originated in Macon, Georgia, on January 5, 2005, and was destined for
Columbia, South Carolina. The NS records indicate that the train received an initial
terminal air brake test in Macon at 10:50 a.m. The train departed Macon at 1:30 p.m. on
January 5 with 2 locomotives, 16 loads, and 14 empties.
11 A switch banner, sometimes termed a switch indicator or switch target, is typically a reflectorized
metal flag-like device connected to the switch stand. (See figure 4.) It consists of one or two flags of
different colors that indicate the position of the switch.
12 NS Operating Rule 181, regarding track warrants, states, in part, �When clearing [a track warrant] at
a point where a switch must be returned to the normal [main line] position, �clear� must not be given until
such switch has been locked in the normal position�.�
13 Although the crew completed train operations within the 12-hour Federal hours-of-service limit, the time
required to complete paperwork for the train caused the on-duty time of the conductor and brakeman to exceed 12
hours. The NS subsequently filed excess service reports with the Federal Railroad Administration for this time.
Factual Information 10 Railroad Accident Report
While en route, the head-end train telemetry device on the lead locomotive failed
to communicate properly with the end-of-train telemetry device.14 At McBeam, Georgia,
the train crew switched the two locomotives, after which reliable communications were
established. No other problems were encountered en route. The train arrived at 10:50 p.m.
at the NS Nixon Yard in Augusta, Georgia, where the inbound crew went off duty.
The outbound train 192 crew (engineer and conductor), who would continue the
trip toward Columbia, South Carolina, went on duty at 12:30 a.m. on January 6, at Nixon
Yard. The conductor said that after taking control of train 192 at Nixon Yard and receiving
the necessary track warrant, he and the engineer took the train about 4 miles across town
to the NS Augusta Yard, where they performed switching duties and added cars to the
consist. After a successful brake test, train 192 departed Augusta Yard about 2:10 a.m. on
January 6. The train consisted of two locomotive units (operating short hood forward)
pulling 25 loads and 17 empties. The train had a trailing tonnage of 3,520 tons and a
length of 2,553 feet. The NS consist list for the train showed 14 cars containing hazardous
materials or hazardous materials residue.15
About 30 minutes and 13 miles after departing Augusta, train 192 approached the
industry track at Graniteville. According to event recorder data, the train was traveling 44
mph at this time (recorder time 2:36:30 a.m.). The engineer moved the throttle handle to
notch 8 (maximum throttle), and the throttle remained in this position until 2:38:11, when
the train speed indicated 47 mph. As the speed continued to increase to 48 mph, the
engineer decreased the throttle to notch 6. At recorder time 02:38:37, the throttle handle
was placed in notch 4, with the speed remaining at 48 mph.
The conductor stated that about this time he heard the train�s emergency brakes
activate. He said he recalled the engineer saying, �The target [switch banner] is wrong,�16
but the conductor said he did not observe the switch target himself. He said that his train
was diverted onto the industry track, where it struck another train, throwing him to the
floor of the locomotive. He said he recalled smelling chemicals after the impact. Event
recorder data indicated that the speed of the train at 2:39:00 a.m., approximately 467 feet
from the final resting point, was 47 mph. At 2:39:20, the speed registered 0 mph.
According to dispatcher radio transcripts, at 2:40:11 a.m., a radio emergency tone
sounded at the desk of the NS dispatcher on duty in Greenville. When the dispatcher
responded to the emergency tone, a caller (believed by the dispatcher to be the train 192
engineer) reported:
14 The end-of-train telemetry device (EOTD) transmits pertinent information, including brake pipe
pressure at the rear of the train, to the head-end telemetry device (HOTD) on the lead locomotive.
15 These included three tank cars of chlorine (the 6th, 7th, and 9th cars from the locomotives); two tank
cars of sodium hydroxide (the 8th and 31st cars); two tank cars containing residue of elevated temperature
rosin (the 16th and 17th cars); one tank car of cresols (the 18th car); four tank cars containing methanol
residue (the 24th through 27th cars); and two tank cars of aniline (the 34th and 35th cars).
16 The investigation revealed that the switch target was on the engineer�s side (the west side) of the
locomotive as the train proceeded north.
Factual Information 11 Railroad Accident Report
Dispatcher, this is train�192�. We need emergency assistance in Graniteville.
We just�hit a switch�at Graniteville�. It was lined off the main line and we
went around there and�hit an engine that was in the track. �We went through
the switch at 45 mph. We need help.
Asked by the dispatcher if the crew was all right, the caller said, �It just
happened�. We need an ambulance. I think I�m bleeding.�
According to NS records, the dispatcher reported the collision and possible injury
to the NS police command center. An employee of the command center subsequently
reported the accident to the Aiken County Sheriff�s Department and requested emergency
medical services.
In a Safety Board interview, the conductor said that he attempted to get off the
locomotive from his side (the east side) of the locomotive, but was unable to do so because
the window and door were jammed. He said that he exited the locomotive from the other
side, followed by the engineer. The conductor said that he and the engineer then walked
about 100 yards, at which point they �met up with some people.� He recalled the engineer
saying that they needed to get �downwind� of the area. He also recalled seeing white or
gray smoke, but no fire. He said they walked a little farther, and then lay on the ground.17
Emergency Response
When train 192 struck train P22, both locomotives and the first 16 cars of train 192
derailed. (See figure 7.) The ninth car from the locomotive units, containing 90 tons of
chlorine, was punctured during the derailment and released chlorine gas. Winds were light
at the time of the accident, and the chlorine vapor cloud settled in the low-lying valley
along the tracks.18 Based on emergency responder observations and the locations of those
receiving fatal injuries, the cloud extended at least 2,500 feet to the north; 1,000 feet to the
east; 900 feet to the south; and 1,000 feet to the west. The sudden release and expansion of
the escaping gas caused the product remaining in the tank to auto-refrigerate and remain in
the liquid state, slowing the release of additional gas.
17 The conductor said that he was subsequently informed in a hospital that an Avondale Mills employee
placed him on the back of his truck and transported him to a hospital after he and the engineer walked away
from the collision. A motorist reported picking up the engineer and transporting him to a hospital.
18 Because chlorine gas is heavier than air with a vapor density of 2.5 at 32° F, it will seek the lowest
point in the immediate area.
Factual Information 12 Railroad Accident Report
Figure 7. Distribution of wreckage after the derailment and collision. �P.C.� indicates esti-
mated point of collision
F
en
ce
L
in
e
Tr
ol
le
y
Li
ne
R
oa
d
C
an
al
S
tr
ee
t
Ai
ke
n
Ro
ad
Aiken
Road
H
ic
km
an
S
tr
ee
t
65
’
65
’
Hard Street
Church
1 2
3
S2
S1
17
0’
5
0‘
1
00
’
1
50
’
2
00
’
16
7
6
54
In
du
st
ry
G
at
e
N
K
E
Y
8
9
10
15
P.S
.
65
5’
77
1’
N
S
R
ig
ht
o
f W
ay
(
P
ro
pe
rt
y
Li
ne
)
N
S
R
ig
ht
o
f W
ay
P
23
6’
34
2’
P1
P2
13
6’
0.
0
(R
ef
.)
H
H
H
14
11
12
13
A
sc
en
di
ng
G
ra
de
P
ar
ki
ng
L
ot
H
ic
km
an
M
ill
B
ui
ld
in
g
D
at
a
P
ro
ce
ss
in
g
O
ffi
ce
P.
L
oc
o
m
ot
iv
e
46
22
P
1.
P
P
G
X
1
21
19
P
2.
W
IT
X
4
76
0
S
1.
Lo
co
m
ot
iv
e
66
53
S
2.
L
oc
om
ot
iv
e
65
93
1.
N
S
4
73
15
3
2.
N
S
2
69
18
2
3.
S
O
U
6
77
4
4.
A
C
F
X
0
51
21
3
5.
N
A
H
X
5
51
30
7
6.
S
B
L
X
1
41
46
7.
G
AT
X
1
71
05
8.
G
AT
X
5
83
26
9.
U
T
LX
9
00
27
0
10
. N
A
H
X
5
51
59
9
11
.
C
S
X
T
49
64
30
12
.
C
S
X
T
49
69
04
13
.
C
S
X
T
49
55
73
14
.
N
S
4
11
17
8
15
.
N
S
4
03
13
0
16
.
G
AT
X
3
19
41
(
O
ne
e
nd
o
ff)
H
.
C
oi
l C
ar
H
oo
ds
P.
S
.
P
oi
nt
o
f S
w
itc
h
&
S
w
itc
h
S
ta
nd
L
oc
at
io
n
P.
C
.
P
oi
nt
o
f C
ol
lis
io
n
S
ou
rc
e:
N
T
S
B
56
4’
A
vo
nd
al
e
M
ill
s
Av
er
ag
e
1.
14
%
D
es
ce
nd
in
g
G
ra
de
71
9’
P.C
.
Factual Information 13 Railroad Accident Report
Immediately after the collision, at 2:39:43 a.m. on January 6, a female employee
on duty at one of the Avondale Mills facilities (about 200 feet from the collision site)
placed a 911 call to the Aiken County Sheriff�s Office 911 Emergency Call Center. The
caller identified herself and reported, �I think there�s been a train wreck�at
Graniteville�at Hickman Mills.� She said that she was alone and that when she went
outside to investigate, she could �see smoke� but could not discern exactly what had
happened. She further indicated that the accident was at the Hickman Street railroad grade
crossing. Near the end of the 48-second call, the caller appeared to become increasingly
agitated, saying, �I smell smoke.� The caller then exclaimed, �I got to get out of here,� at
which point the call abruptly ended.
Over the next 10 minutes, about a dozen additional calls were made to 911, with
callers reporting that there had been a train wreck. Some callers reported a low-lying
yellow haze that smelled like bleach. Within about an hour after the accident, more than
100 additional 911 calls were received. By 6:00 a.m., more than 200 calls had been
received. Also commencing about 2:40 a.m. were calls on the non-emergency telephone
lines of the emergency call center. About 80 calls were received on these lines within the
first hour, and by 5:40 a.m., about 200 calls had been received.
Fire Department Response
At 2:40:40 a.m., resources of the Graniteville, Vaucluse, and Warrenville Volunteer
Fire Department were dispatched to the scene, with the first responding unit reported to be
en route less than a minute later.19 At 2:42 a.m., upon hearing a report from the dispatcher
that a smell of chemicals was reported in the area, the initially responding fire department
senior officer (the fire chief) advised further responding fire department personnel to stand
by at their locations away from the scene until the situation could be further assessed.
At 2:45 a.m., emergency dispatch advised the fire department that it had confirmed
the possibility that two trains had collided head on. Additional confirmation came from
the NS about 3 minutes later. The fire chief, upon approaching the scene, smelled an
intense chemical odor and experienced difficulty breathing. At 2:46 a.m., a hazardous
materials team was requested. At 2:48 a.m., the fire chief advised dispatch that he could
not breathe and was withdrawing from the area. At 2:49 a.m., the fire department asked
dispatch to initiate the Aiken County Reverse 911 Emergency Notification System, with
instructions for residents to shelter indoors.
Commencing about 2:50 a.m., additional resources⎯ambulances, hazardous materials
personnel and equipment, Aiken city and other mutual aid services⎯were asked to respond. At
2:52 a.m., dispatch advised responders that three persons were trapped inside the Hickman
plant. At 2:54 a.m., fire department personnel asked dispatch to advise Aiken Hospital that
persons overcome by chemical fumes were en route from the scene. At 2:57 a.m., the fire
department asked that approach roads be blocked (which effectively initiated a 1-mile-radius
buffer around the accident site) and reiterated the earlier reverse 911 request to shelter indoors.
19 Information in this section is based on incident response data and communications information
provided by the responding agencies, as well as on interviews with key emergency response personnel.
Factual Information 14 Railroad Accident Report
At 3:05 a.m., while awaiting delivery of train consist information that had been
faxed from NS, the fire chief directed that an incident command center be established; the
fire chief would become incident commander for the search and rescue effort. About this
time, firefighters also asked dispatch to obtain wind direction information (from Bush
Field in Augusta), which was received about 3 minutes later.
At 3:06 a.m., firefighters were informed that the sixth through ninth cars on the
train contained chlorine and sodium hydroxide and that additional information would be
forthcoming. At 3:08 a.m., the fire department began staging equipment and personnel at
the incident command center, which had been set up at a nearby car dealership. At
3:10 a.m., firefighters asked that all Aiken County emergency medical services
ambulances be placed on standby.
At 3:13 a.m., an Avondale Mills employee told firefighters that workers on duty at the
Stevens steam plant could not be contacted. The employee expressed concern about a possible
explosion if the workers had departed the plant without properly shutting down the boilers.
At 3:21 a.m., personnel from the Aiken County hazardous materials team arrived
at the incident command center. At 3:24 a.m., a copy of the faxed train 192 consist list was
delivered, and fire department authorities advised all responding personnel to report with
their equipment to the command center. At 3:30 a.m., the fire department requested all
available self-contained breathing apparatus.
At 3:33 a.m., a report was received of a �steady stream� of individuals departing
the Ascauga Lake Road area. At 3:35 a.m., authorities established the first of four
decontamination stations to treat individuals exposed to the chlorine gas. At 3:37 a.m.,
dispatch advised firefighters that it had received reports of people �down� inside the
Avondale Mills facility. At the same time, a firefighter and a mill supervisor entered the
steam plant to prevent a possible boiler explosion. A decision was made to shut down and
evacuate the plant.
At 3:38 a.m., a second decontamination station was set up. At 3:39 a.m., the
incident commander, concerned that the incident command center was too close to the
accident site, directed that it be moved from the car dealership to a location about a mile
away. Some emergency response apparatus remained at the initial site as a �forward
command� location. At 3:40 a.m., a firefighter wearing personal protective equipment got
close enough to the accident site to note the number on the tank car that had been
breached. He also encountered an individual suffering from gas inhalation and discovered
another individual trapped in an automobile beneath a fallen tree near the derailment site.
(Both these individuals were later successfully rescued.)
Beginning at about 3:50 a.m., several entry teams, consisting of firefighters
wearing personal protective equipment and riding in privately owned pickup trucks, were
organized and dispatched to the accident site for search and rescue operations. Upon
locating individuals or groups affected by the chlorine gas, the teams transported them to
one of the decontamination sites. The entry teams then returned to the site to repeat the
search and rescue cycle for several hours.
Factual Information 15 Railroad Accident Report
At 3:53 a.m., a third decontamination station was established at a local high
school. At 4:10 a.m., an entry team reported a downed electrical power line near the
wreckage site, and a request was made to the utility company to respond to the scene to
disconnect the power feed. Also at 4:10 a.m., an entry team reported no visible fire in the
derailment wreckage. This report was revised at 4:24 a.m., when responders reported
seeing �bright orange smoke�20 emanating from one railcar and �green smoke�21 from
another.
At 4:55 a.m., an entry team entered the steam plant to shut down the facility.
Moments later, the mission was revised when �five or six� individuals were reported to be
trapped in a room at the plant. One person was found and rescued, after which the team
completed the shutdown of the steam plant at 5:07 a.m. and performed a final sweep of the
plant. Several additional entry teams began missions to assess the condition of the railroad
equipment.
The immediate area around the accident site remained relatively stable until
1:00 p.m., when a fire was reported at the steam plant. A fire department entry team
entered the plant and found that a fire had ignited in coal chutes feeding several of the
boilers. A pumper truck supplied water to an unmanned waterline-fed monitor nozzle that
discharged a spray on all the coal feeders. At this point, the fire was under control but not
extinguished. The discharging monitor nozzles were left in place while workers evacuated
the area. The scene remained relatively stable for the balance of the day.
For the next several days, fire department entry teams monitored and contained the
fire in the steam plant coal feeders while a cleanup of the railroad wreckage continued. At
8:00 a.m. on January 14, with the report that the hazardous materials had been removed
from the unbreached railcars, support operations were concluded, and hazardous materials
response personnel were released. At about noon, fire department personnel returned to
the steam plant to extinguish the remaining fire in the coal feeders. The fire was reported
extinguished at about 4:00 p.m., which concluded the fire department�s operations for the
incident.
Sheriff�s Office Response
After the initial 911 call, patrol units of the Aiken County Sheriff�s Office were
dispatched to the scene. The first sheriff�s office responder arrived on the scene about
2:42 a.m., followed shortly thereafter by several other officers. When they approached the
scene, the officers began experiencing respiratory difficulties because of the chlorine gas.
They immediately withdrew to a safe distance and awaited further instruction. Several of
the responding officers were taken to a local medical facility for treatment.
20 According to technical experts, the appearance of �orange smoke� is consistent with entrained ferric
chloride in the released liquid chlorine. Light appears red as it passes through the ferric chloride that forms
as chlorine reacts with steel in an oxygen-deprived environment. The absence of fire damage in proximity to
the chlorine tank puncture indicates that the derailment did not result in a fire.
21 The appearance of �green smoke� is consistent with a discharge of chlorine gas.
Factual Information 16 Railroad Accident Report
Workers and residents in the immediate vicinity of the gas release began to
evacuate the area within moments of the accident. They were assisted in this evacuation
by fire department and sheriff�s office personnel on scene.
Later in the morning, the sheriff�s office directed the evacuation of those who had
been sheltering in place within a 1-mile radius of the site. This mandatory evacuation
affected about 5,400 people. Sheriff�s office personnel, assisted by mutual aid from a
number of other area law enforcement agencies, conducted a house-to-house evacuation.
Approach roads were closed at key intersections. The sheriff�s office also instituted a
300-meter buffer zone around the site and restricted access to only those individuals
wearing the appropriate personal protective equipment. The evacuation was lifted in
phases, beginning on January 13, with the last evacuees permitted to return to their homes
on January 19.
Hazardous Materials Response
About 11:00 p.m. on January 6, responders inserted a temporary polymer patch in
the opening of the punctured ninth tank car. At 7:00 p.m. on January 8, responders began
unloading sodium hydroxide from the eighth tank car. At 8:50 p.m., the temporary
polymer patch in the ninth tank car failed, releasing chlorine vapors and causing the
unloading of the eighth tank car to be temporarily discontinued. By 8:37 a.m. on
January 9, responders had inserted a second polymer patch in the opening in the punctured
ninth tank car. Chlorine vapor was then drawn from the car to create a vacuum that would
reduce the amount of chlorine gas escaping. The chlorine vapor removed from the tank
was transferred into a sodium hydroxide solution to neutralize it. Following these
measures, the unloading of the eighth tank car was resumed and completed by 3:30 p.m.
While the eighth car was being unloaded, construction began on a permanent lead patch
for the punctured ninth car.
At 9:30 a.m. on January 10, the punctured chlorine tank car was rotated so the
puncture was at the highest elevation on the tank car. This rotation disturbed the liquid
chlorine in the tank and caused a delay in efforts to unload other tank cars.
At 12:10 p.m. on January 10, responders began unloading the chlorine from the
derailed sixth car in the train. On the morning of January 11, responders rejected the plan
for a lead patch on the punctured ninth tank car and decided to use a steel patch instead.
Unloading of the sixth car was completed by 2:00 p.m. on January 11.
About 1:10 a.m. on January 12, responders began unloading chlorine from the
derailed seventh car. By 9:30 a.m., the steel patch was in place on the punctured tank car,
and the unloading was started. Because the punctured tank car had extensive damage, the
remaining chlorine could not be removed as it had been from the unbreached cars. The
chlorine in this car had to be vaporized and transferred from the tank as a gas, after which
it was bubbled through a sodium hydroxide solution held in a separate tank. This process,
which converted the chlorine into a relatively safe and easily transportable liquid bleach
and salt solution, required several days to complete.
Factual Information 17 Railroad Accident Report
By 2:10 p.m. on January 12, the unloading of the seventh car was complete. By the
early morning hours of January 13, the two unbreached chlorine tank cars had been
unloaded, placed on railroad flat cars, and moved from the site.
By midnight on January 18, the unloading of the punctured tank car was complete.
(See figure 8.) On January 19, by 9:00 a.m., the tank car was cleaned and purged on site. It
was then loaded on a flat car and moved to the Augusta Yard the following morning,
January 20.
Injuries
Nine persons, including the train 192 engineer, died from chlorine gas inhalation
as a result of the accident.22 Of the eight civilians who received fatal injuries, six were
employees of Avondale Mills facilities to the west and north of the accident site, one was a
truckdriver at one of the plant facilities to the west of the site, and one was in a residence
south of the site.
Figure 8. Ruptured chlorine tank car during unloading. The tank has been rotated so
that the patched puncture is at the top. Hoses attached to the patch are drawing off
chlorine vapors that are then bubbled through a sodium hydroxide solution to reduce
their hazard.
22 The death of another Graniteville resident on April 19, 2005, was initially reported by media sources
to have occurred as a result of exposure to the chlorine gas released in the accident. The final autopsy report
for this individual listed the death as natural due to pulmonary thromboemboli. Regardless of the
circumstances, this death would not have met the criteria for death reporting under 49 Code of Federal
Regulations 830.2, because it did not occur within 30 days of the accident.
Factual Information 18 Railroad Accident Report
About 554 people, among them civilians, railroad crew, and emergency
responders, were taken to local hospitals complaining of respiratory difficulties. Of these,
75 were admitted for treatment. Six firefighters were treated and released; one firefighter
was admitted to the hospital and remained there for several days. Two sheriff�s department
officers were also treated and released.
The coroner�s investigation reports for the eight deceased civilians all listed the
probable cause and mechanism of death as asphyxia, which occurred within �minutes� of
exposure, with secondary/contributing factors that included exposure to chlorine gas. The
locomotive engineer survived the collision but died several hours later. The coroner�s
investigation report listed the probable cause and mechanism of death of the engineer as
lactic acidosis with secondary/contributing factors that included exposure to chlorine gas.
Damages
Hazardous Materials Cars
Of the 16 derailed cars, 5 contained hazardous materials or hazardous materials
residue. The hazardous materials cars included three tank car loads of chlorine (the 6th, 7th,
and 9th cars from the locomotives), one tank car load of sodium hydroxide (the 8th car), and
one empty tank car containing residue of elevated temperature rosin (the 16th car). The ninth
car was the only derailed hazardous materials tank car that was breached and released its
cargo. The other four tank cars were damaged, but their tanks were not breached.
The sixth car (chlorine) sustained a severe dent just to the right of center in the
lower half of the B-end head. The appearance and location of the dent was consistent with
its having been made by a coupler from another car. The bottom of the tank sustained
some flattening, and the jacket was substantially damaged.
The seventh car (chlorine) sustained a severe dent on the left side (from the B-end)
between the center of the tank and the A-end, just below the midline of the tank. A severe dent
was also found on the A-end to the left of the coupler, and the coupler had been torn out. The
jacket had several dents, with the most severe occurring between the B-end and the top fittings.
The eighth car (sodium hydroxide) sustained extensive jacket damage, and the
jacket was torn from almost one-third of the tank on the A-end. Two smaller dents were
observed in the shell near the A-end on the right side (from the B-end). The coupler and
draft gear were missing from the B-end, and the stub sill was bent upwards.
The ninth car (chlorine) had a puncture and a tear on the right side near the middle
and slightly toward the A-end of the tank. The opening was 34 inches long and 5 inches
wide at its widest point. The area around the puncture was crushed inward, and there were
severe dents on either side of it. These dents ranged in depth from 15 to more than 20
inches. The car came to rest angled slightly to one side with the puncture opening
extending below the midline of the resting tank. The left side of the tank had some
Factual Information 19 Railroad Accident Report
flattening, and there was a dent on the shell near the A-end head weld at about the
2 o�clock position. Emergency responders on the day of the accident measured the liquid
level of the chlorine remaining in the tank and found that about one-third of the original
load, or 30 tons (4,609 gallons), remained in the tank.23 The responders also noted that the
B-end coupler24 of a steel coil car (the 11th car in the train) was in contact with the
damaged tank jacket near the puncture and that it had frost on its surface. Metallurgical
examination of the puncture area revealed that the dents in the puncture area mated with
damaged protrusions on the 11th-car coupler.
Other Railroad Equipment
The lead locomotive unit of train 192 derailed but remained upright (leaning to the
east at about a 12° angle) and positioned somewhat aligned with and about 4 feet from the
centerline of the industry track. (See figure 9.) The locomotive unit appeared not to have
lost overall structural integrity or to have sustained significant loss of occupant survival
space. The cab windows were not broken. The unit sustained severe impact damage to the
front end, with the front coupler, pilot plate structure, and stepwell elements severely
mangled and with various components bent and displaced in an upward and aft direction
by about 2 feet. Except for the front impact damage, the carbody of this unit did not
exhibit substantial collision damage.
23 The NS estimated the amount remaining in the tank at 44 tons, or about 1/2 the original load of 90 tons.
24 A standard E-type bottom-shelf coupler.
Figure 9. The lead locomotive of train 192 (left) and the lead locomotive of train P22
(right) showing collision damage.
Factual Information 20 Railroad Accident Report
The second locomotive unit on train 192 derailed but remained coupled to the lead
locomotive, upright and positioned with the aft end skewed from alignment with the track
by about 15 feet. The unit showed impact damage at the aft end on the right side of the
frame, including a segment of missing handrail. The aft right corner pilot plate structure
and stepwell elements were severely bent and displaced in an aft direction. The aft left
corner pilot plate structure sustained impact damage. Except for the aft end frame and
pilot plate structure damage and several carbody panel separations, this unit did not show
substantial collision damage. There was visible evidence of a diesel fuel spill beneath the
fuel tank, but there was no apparent fire.
Train P22 consisted of one locomotive and two empty freight cars at the time of
the accident. It was propelled about 217 feet northward (along the track) by the impact.
The locomotive derailed but came to rest upright and positioned somewhat longitudinally
aligned with the track about 3 feet to the east of the track centerline. The unit sustained
impact damage to the front end, with the front coupler, pilot plate structure, and stepwell
elements severely mangled. Various components were bent and displaced in an upward
and aft direction by about 5 feet.
The P22 locomotive unit sustained severe impact damage to the front carbody
cowl (short hood), which was displaced in an aft direction by about 2 feet. The right side-
sill was buckled downward several inches, proximate to the area above the fuel tank. The
left side-sill was also buckled, but to a lesser degree than the right. Except for the front
impact damage and the damaged side-sills, the carbody of this unit did not exhibit
substantial collision damage. There was visible evidence of a diesel fuel spill beneath the
fuel tank but no evidence of fire. The cab compartment appeared to be relatively intact and
did not lose overall structural integrity or sustain significant loss of survival space. One
window on the left side of the cab was shattered.
The first railcar on train P22 was a covered hopper. It did not derail and remained
coupled to the locomotive. The second railcar, also a covered hopper, uncoupled from the
first car and came to rest about 81 feet away along the industry track. Both truck
assemblies separated from the carbody. The leading truck assembly was found on the
ground adjacent to the car�s front coupler; the aft assembly was found wedged against the
underside discharge door beneath the center of the car. The leading end of the car was on
the ground; the aft end had overridden and was wedged against and resting upon the
leading end-sill of a tank car that had been on the industry track west of train P22.
Some main line and industry track was damaged in the derailment or removed to
facilitate cleanup. A total of 14 track panels (39 feet each) were installed: 6 panels on the
main track and 8 panels on the industry track. In addition, 10 carloads of ballast stone
were spread. Total damage to the rolling stock and track was estimated in excess of $2.19
million. The NS reported total damages to exceed $6.9 million.
Emergency Response Equipment
The Graniteville, Vaucluse, and Warrenville Volunteer Fire Department reported
that two pumper trucks, one medical unit vehicle, and one service truck, all of which had
Factual Information 21 Railroad Accident Report
been parked at the department�s Station No. 1 for the duration of the gas release, had been
destroyed or damaged beyond economical repair as a result of chlorine gas contamination.
Total cost for the damaged or destroyed equipment was estimated in excess of $630,000.
Wreckage
Wreckage and debris of the derailed railroad equipment extended more than 500
feet, beginning at the industrial turnout switch and continuing in a northerly direction
along the industry track. Investigators were not able to perform a detailed documentation
of the wreckage because of the continuing potential of a hazardous gas release and
because of the wreckage cleanup efforts that prevented access to the site.
Personnel Information
Train P22
Engineer. The train P22 engineer was originally hired on July 25, 1979, as a car
retarder operator.25 On September 3, 1979, he was promoted to yard foreman, helper, and
trainman. On July 7, 1986, he was promoted to conductor. He became an engineer on
September 12, 1990. His most recent engineer re-certification was on August 16, 2004.26
The engineer stated that he was not preoccupied or distracted during his shift. He
also said he was not aware that either his brakeman or conductor was preoccupied or
distracted during their shift. He said that other than intermittent problems with his
locomotive radio, he experienced no problems with any of his equipment.
The engineer characterized his health as good. His most recent physical
examination before the accident was in August 2004. The engineer explained that he
underwent two physicals at that time: one conducted by his personal physician and one as
part of his engineer re-certification. He said that he had no problem with his hearing or
vision, although he wore prescription reading glasses. He said he had not used any over-
the-counter or prescription medications, but said he periodically took vitamins. He also
said he had not used alcohol or illicit drugs on the day before the accident.
As an extra-board employee, the engineer was called to work as needed. He said
he was called at 1:00 a.m. on Tuesday, January 4, for a 2:30 a.m. on-duty time. He worked
from that time until about 12:30 p.m. on January 4. He said he went to bed about 7:00 p.m.
on January 4. At 10:32 p.m., the engineer was called to work as the engineer on train P22
the next day, January 5. He was authorized to deadhead (by taxi or other means) from his
25 A car retarder operator remotely operates a brake device built into the rails to reduce the speed of
cars being switched into a classification yard (where cars are grouped before being made up into trains).
26 Engineer certifications are valid for 3 years at time of issuance and include an evaluation of vision
and hearing.
Factual Information 22 Railroad Accident Report
home in Columbia to Aiken to arrive at 12:32 a.m. so that he would be in place and ready
for work the next morning. Because the engineer was required to be given 8 hours rest
time after arriving in Aiken, he could not report for train P22 until 8:32 a.m., more than
1 1/2 hours after the conductor and brakeman had gone on duty.27 He went off duty at
8:11 p.m. and had been off duty for about 6 1/2 hours when the accident occurred.
Conductor. The train P22 conductor (who was the regularly assigned P22
brakeman but who was working in place of the vacationing conductor on the day before
the accident) was originally hired as a trainman on April 4, 1978. On May 18, 1978, he
was promoted to yard foreman, helper, and trainman. On January 2, 1992, he was
promoted to conductor.
The conductor said he had no problems with any equipment as he worked the day
before the accident. Similarly, he had experienced no problems related to the weather, nor
was he distracted or preoccupied while he performed his duties. He stated his overall
health was �in good shape.� He said he had not used prescription medications but that he
sometimes took medicine for allergies. The conductor said that he believed he had taken
two Benadryl28 tablets about 1:00 p.m. on the day before the accident. He said he took no
additional over-the-counter medications, nor did he consume any illicit drugs or alcohol
that day. The conductor recalled that his most recent physical examination before the
accident was in January or February 2004, adding that he undergoes a physical
examination annually.
The conductor traveled about 50 miles from his home to the work site Monday
through Friday, arriving in time to go on duty at 7:00 a.m. He said that on the Monday
before the accident, he went off duty about 7:40 p.m. and retired about 9:50 p.m. On
Tuesday, he awoke at 5:15 a.m., went on duty at 7:00 a.m., went off duty about 8:00 p.m.,
and retired about 9:45 p.m. He said he arose on Wednesday, January 5, the day before the
accident, at about 5:00 a.m. and reported for work at 7:00 a.m. He went off duty about
8:11 p.m. and had been off duty for about 6 1/2 hours when the accident occurred.
Brakeman. Records revealed that the brakeman was originally hired on
October 18, 1978. On November 26, 1978, he was promoted to yard foreman, helper, and
trainman. On January 2, 1992, he was promoted to conductor.
The brakeman said he had not been distracted or preoccupied during his shift.
Likewise, he said he was not aware that his engineer or conductor had been distracted or
preoccupied while working.
The brakeman reported that he was �in good health� and that his most recent
physical examination was conducted by his personal physician in December 2004. The
27 The engineer would have had to have been called by 9:00 p.m. on January 4 to be able to start work
at 7:00 a.m. on January 5 in Aiken. The NS representatives said that the job vacancy did not occur in time to
allow a call by 9:00 p.m.
28 Benadryl, an over-the-counter antihistamine containing diphenhydramine, relieves red, irritated,
itchy, watery eyes; sneezing; and runny nose caused by hay fever, allergies, and the common cold. The
medication can have sedating effects.
Factual Information 23 Railroad Accident Report
brakeman said he took multivitamins and a vitamin B complex. He also said he took a
green tea supplement called MSM29 as well as a supplement known as Laye.30 He said
both supplements were used to help alleviate pain in his joints. He said he was not sure
whether he had taken either product the day before the accident. He said he had not used
over-the-counter or prescription medications or alcohol or illicit drugs on the day before
the accident.
The brakeman was an extra-board employee who was called to work as needed. He
had not worked the Saturday or Sunday before the accident. He said that on Monday, he
awoke at 4:25 a.m., reported for duty in Newberry, South Carolina, at 7:00 a.m., and went
off duty at 1:30 p.m. He said he retired at 9:45 p.m. that evening. The brakeman worked
train P22 on Tuesday. He said he awoke about 4:25 a.m. and reported for work at Aiken at
7:00 a.m. He went off duty at 7:31 p.m. and retired about 10:30 p.m. On Wednesday, the
day before the accident, he said he awoke at 4:25 a.m., went on duty in Aiken at 7:00 a.m.,
and went off duty at 8:11 p.m. At the time of the accident, he had been off duty for about
6 1/2 hours.
Train 192
Engineer. Records revealed that the engineer was originally hired on April 7,
1997. On August 25, 1997, he was simultaneously promoted to yard foreman, helper,
trainman, and conductor. On December 12, 2001, he was promoted to engineer. His most
recent engineer re-certification was issued on March 20, 2002. The engineer�s most recent
evaluation was on August 16, 2004. In the comments section, the examining officer noted,
�Did well, no exceptions.�
On Monday, January 3, 2005, he had been called for duty at 1:24 a.m. for an
on-duty 3:00 a.m. start time. He went off duty later that morning at 8:30 a.m. He was
called later that day, at 4:02 p.m., for an on-duty start time of 5:30 p.m. He went off duty
the following morning, Tuesday, January 4, at 1:25 a.m. At 10:49 p.m., he was called for
an on-duty start time of 11:59 p.m. He went off duty the following day, Wednesday,
January 5, at 11:45 a.m. He was called at 11:07 p.m. that evening for an on-duty start time
of 12:30 a.m., Thursday, January 6, for train 192. At the time of the accident, he had been
on duty for about 2 hours 10 minutes.
Conductor. Records revealed that the conductor was originally hired on
February 22, 1999. On August 30, 1999, he was simultaneously promoted to yard
foreman, trainman, helper, and conductor.
On Sunday, January 2, the conductor had been called at 9:45 p.m. for an 11:10 p.m.
on-duty start time. He went off duty the following day, Monday, January 3, at 4:34 a.m. He
29 Methylsulfonylmethane (MSM) is also known as methyl sulfone or dimethylsulfone. It is advertised
for relief of pain from arthritis, back pain, or muscle pain. No published research studies link MSM to any of
the health claims made for it.
30 Subsequent contact with the brakeman determined this product to be L-Lysine. L-Lysine is used by
the body in building new tissue. It also promotes the body’s protective substances, such as enzymes and
antibodies. L-Lysine is a natural constituent in foods, with no known toxic effects.
Factual Information 24 Railroad Accident Report
was called the following day, Tuesday, January 4, at 1:00 a.m. for an on-duty start time of
2:30 a.m. He went off duty later that day at 12:33 p.m. He was called later that evening at
10:49 p.m. for an on-duty start time of 11:59 p.m. He went off duty the following day,
Wednesday, January 5, at 11:46 a.m. He was called at 11:07 p.m. that evening for an
on-duty start time of 12:30 a.m. on Thursday, January 6, for train 192. At the time of the
accident, he had been on duty for about 2 hours 10 minutes.
Toxicological Testing
In accordance with Federal requirements,31 specimens were obtained from the
engineer and conductor of train 192 and submitted for toxicological testing. The results
were negative for alcohol and the screened substances.32 Based on the circumstances of
the accident and the fact that the crew of train P22 had been off duty for several hours
when the accident occurred, the NS determined, in consultation with representatives of the
Federal Railroad Administration (FRA), that neither the P22 crew nor dispatching
employees were required by Federal regulations to undergo required postaccident
chemical testing.
Meteorological Information
The nearest National Weather Service reporting station to Graniteville, South
Carolina, was about 17 miles away, in Augusta, Georgia. Sunset was at 5:34 p.m. Augusta
weather at 6:53 p.m. on January 5, 2005, (the approximate time the P22 crew was
preparing to depart Graniteville and return to Aiken) was as follows: winds from the south
at 5 mph; visibility unrestricted at 10 miles; skies mostly cloudy; temperature 60°
Fahrenheit (F); dew point 57° F. No precipitation was reported.
Augusta weather at 2:32 a.m. on January 6, 2005, (minutes before the collision and
derailment) was as follows: winds from the south-southwest at 7 mph; visibility 5 miles;
skies clear; temperature 55° F; dew point 54° F. No precipitation was recorded between
7:00 p.m. on January 5 and 3:00 a.m. on January 6.
Operations and Track Information
The NS, which owns and operates the track in the accident area, is a subsidiary of
the Norfolk Southern Corporation, headquartered in Norfolk, Virginia. The NS operates
approximately 21,300 route miles of track in 22 eastern and southeastern States, the
District of Columbia, and Ontario, Canada.
31 Title 49 Code of Federal Regulations Part 219, Subpart C.
32 These substances included cannabinoids, cocaine, opiates, amphetamines, methamphetamines,
phencyclidine, barbiturates, benzodiazepines, and ethyl alcohol.
Factual Information 25 Railroad Accident Report
The NS is the product of various railroad combinations, reorganizations, and
consolidations. The largest NS predecessors were Norfolk & Western Railway and
Southern Railway. The NS subsequently acquired a portion of the assets of Conrail, and
the holdings of these former carriers make up the majority of the NS today.
The area in which the accident occurred is within the NS Piedmont Division. The
R main line is a 190.52-mile track segment that extends from a junction to a main line near
Charlotte, North Carolina, south through Columbia, South Carolina, to Augusta, Georgia.
Numerous small towns and industries are near this track segment.
The segment of the R line at Graniteville is a non-signaled single main track
controlled by a train dispatcher using a track warrant control system from Greenville,
South Carolina. According to the terminal superintendent in charge of the territory, at the
time of the accident, five trains were being operated daily over the track at Graniteville.
These included two through freight trains in each direction operating between Augusta,
Georgia, and Columbia, South Carolina, in addition to local train P22. The NS timetable
speed from Columbia to Augusta at the time of the accident was 49 mph, in accordance
with FRA limits for the territory, except in areas with reduced speeds because of curves.
The closest curve to the accident site is a left hand, 1º curve beginning at MP
R178.45 and ending at MP R178.65. At MP R178.3 is the turnout to the industry track.
The turnout uses a No. 10 switch33 that is of similar construction and material to the main
track. When lined for the diverging track, the maximum authorized speed through the
switch is 15 mph. The switch was equipped with a 14-inch-diameter red and white
reflective banner atop a 7-foot long banner shaft secured to the switch stand.
Postaccident Inspections
Switch and Track
The chlorine release that occurred as a result of the accident prevented Safety
Board investigators from immediately entering the accident site. Because the Federal
Bureau of Investigation (FBI) had trained hazardous materials teams with the equipment
necessary to enter the site safely, the Safety Board requested the assistance of the FBI to
document switch alignment, equipment, and other information. After being advised as to
what to look for by Safety Board investigators, FBI representatives inspected the switch at
the industry track turnout on January 7, 2005, the day after the accident. The switch was
found to be lined and locked for the industry track. It showed no evidence of having been
moved or tampered with. The switch points were in the proper position for the turnout,
and the switch banner was in the red position, indicating that the switch was lined for the
diverging track. No defects were noted with the switch points, switch throw rods, switch
33 The number designation of a turnout switch refers to the angle at which the turnout track diverges
from the main track, with higher numbers representing a more gradual angle. The investigation determined
that the turnout into the Avondale Mills industry track led a train from the main line into an approximate 8º
curve.
Factual Information 26 Railroad Accident Report
lugs, heel blocks, switch stand (base), or the switch lock. FBI investigators removed the
switch handle and lock and the mast and banner and assisted the Safety Board in its
investigation of these items.
On January 14, 2005, Safety Board investigators were able to inspect the accident
site. A track inspection was performed on the industry track turnout and the track north
and south of the turnout for approximately 1/10 mile. No main line track damage was
incurred due to the accident, although a portion of the track north of the turnout had been
removed to allow for cleanup.
The industry track turnout and the track south of it were intact and had no defects
under Federal track standards. A section of the industry track near the switch had been
damaged in the accident and had also been removed to allow for cleanup. The remainder
of the industry track was intact. The NS maintains the track in the turnout to a limit of 295
feet, beginning at the switch points and extending into the industry track. An FRA
inspection on January 14, 2005, identified no defects in this section of track.
Investigators reviewed track inspection records for the R line between Columbia,
South Carolina, and Augusta, Georgia, for the time period between November 9, 2004,
and January 5, 2005. NS inspection records for the track segment where the accident
occurred, between MP R178 and MP R179, showed no defects for the 2-month period.
The FRA requires that this class of track (class 4) be inspected twice weekly, and the
records showed that the NS was in compliance with this requirement. The FRA also
requires a monthly switch inspection. NS records showed the switch to have been last
inspected on December 7, 2004. Records of a postaccident (January 8, 2005) FRA
inspection of this track segment showed no defects.
Railcars
After the derailment, the 26 cars from train 192 that had not derailed were
decontaminated and moved to the Augusta Yard. On January 8, 2005, two locomotive
units were coupled to that train, and an air brake test was performed. Both automatic and
emergency brake applications were performed successfully.
After the brake tests, the cars were inspected by Safety Board investigators along
with representatives of the NS, the State of South Carolina, and the FRA. A number of
FRA safety appliance defects were found. In addition, inspectors found two broken brake
shoes, one knuckle pin (coupler) defect, one leaking roller bearing seal, two piston travel
defects, and one brake shoe that failed to center. The brakes also cut out on one car.34 The
identified defects would not have affected the performance of the train.
Train 192 was equipped with an automatic two-way end-of-train device (EOTD).
Movement by the engineer of the automatic brake handle to the emergency position would
automatically (via telemetry) cause the EOTD to activate emergency braking from the rear
34 Freight car brakes are �cut out� when the branch pipe from the train air brake line to the car�s air
brake system is closed off, nullifying the brakes on that car without affecting the brakes of adjacent cars.
Factual Information 27 Railroad Accident Report
of the train, as well as from the head end. This system is designed to provide a more rapid
brake application throughout the train and to ensure a complete brake application even in the
event of a blockage in the air brake line. The postaccident inspection of the control stand in
the cab of the lead locomotive of train 192 showed the automatic brake handle in the
emergency position. The EOTD from train 192 was tested and found to function as designed.
Tank Car and Hazardous Materials Information
Tank Cars
A review of Association of American Railroads (AAR) certificates of construction
for the derailed tank cars, as well as the postaccident inspection, revealed that each tank
car was equipped with double-shelf couplers. These couplers are designed to resist vertical
movement between joined couplers and thereby remain engaged during switching
operations or accidents.
The sixth car in the train (SBLX 14146), containing chlorine, was built in 1997 by
Trinity Industries, Inc., as a U.S. Department of Transportation (DOT) specification
105J500W tank car. The last periodic qualification of the car was in February 2004. The
tank heads (the curved ends of the tank) were manufactured from 13/16-inch plate, and the
shell was manufactured from 0.7874-inch carbon steel plate. Head and shell plates were
specified as AAR specification TC-128-B normalized steel.35 This car had a thermal
protection system consisting of 2 inches of ceramic fiber covered with 2 inches of
fiberglass. The AAR certificate of construction indicated that the 13/16-inch tank head, in
combination with the insulation and the 0.1196-inch steel jacket, met the DOT tank-head
puncture-resistance performance standards of 49 Code of Federal Regulations 179.16.
The seventh car in the train (GATX 17105), also containing chlorine, was built in
1979 by General American Transportation Corporation. It was a DOT specification
105J500W tank car that was originally built as a specification 105A500W tank car. The
last periodic hydrostatic test was in January 2000. The tank heads were manufactured
from 13/16-inch plate, and the shell was manufactured from 0.7874-inch carbon steel
plate. Head and shell plates were non-normalized AAR specification TC-128-B steel. This
car had a thermal protection system consisting of 4 inches of foam insulation. The tank car
also met the DOT tank-head puncture-resistance performance standards of 49 Code of
Federal Regulations 179.16.
The eighth car in the train (GATX 58326), containing sodium hydroxide solution,
was built in June 1980 by ACF Industries, Inc., as a DOT specification 111A100W1 tank
car. The last periodic hydrostatic test was in January 1998. The tank heads and shell were
manufactured from 7/16-inch carbon steel plate. Head and shell plates were non-
35 Normalized steel has undergone a heat treatment process that lowers the temperature at which the
material transitions from ductile to brittle. The process also increases the amount of energy required to cause
fracture. Since 1989, pressure tank car shells have been required to be fabricated from normalized steel.
Factual Information 28 Railroad Accident Report
normalized ASTM36 specification A-515 grade 70, except for the center section of the
shell, which was non-normalized AAR specification TC-128-B steel. This car had a total
of 6 inches of fiberglass insulation.
The ninth car in the train (UTLX 900270), the punctured chlorine car, was built in
1993 by Union Tank Car Company. It was a DOT specification 105J500W tank car that
was originally built as a specification 105S500W tank car. The last periodic qualification
was in July 2004. The tank heads were manufactured from 53/64-inch plate, and the shell
was manufactured from 0.777-inch carbon steel plate. Head and shell plates were
specified as AAR specification TC-128-B normalized steel. This car had a thermal
protection system consisting of 2 inches of ceramic fiber covered with 2 inches of
fiberglass. This car was equipped with 1/2-inch full head shields.37
Chlorine tank cars such as the punctured car are pressure tested to 500 pounds per
square inch, gauge (psig), compared to 300 psig for tank cars used to transport anhydrous
ammonia and liquefied petroleum gas. Because of the higher test pressure, chlorine tank
car walls are thicker than those of lower-rated pressure tank cars.
Hazardous Materials
Four of the five tank cars that derailed (three car loads of chlorine and one car load of
sodium hydroxide) were shipped by Olin Chlor Alkali Products (Olin) of Augusta, Georgia.
The three tank cars containing chlorine were loaded between December 31, 2004, and
January 4, 2005. Loading of the tank car that was punctured in the derailment had been
completed at 3:10 a.m. on January 4. The recorded loading pressure for each car was 22
psig, which equates to a chlorine temperature of about 12º F. The estimated temperature of
the chlorine (and the tank car steel) at the time of the derailment was 26º F.
The cars each contained 180,000 pounds (about 13,830 gallons)38 of chlorine.
According to Olin�s material safety data sheet, chlorine is a poisonous gas, an oxidizer,
and a marine pollutant. It has an IDLH value39 of 10 parts per million (ppm). If inhaled,
chlorine will react with moisture in the respiratory tract and lungs to form hydrochloric
acid, resulting in inflammation of these tissues. Severe exposure can result in pulmonary
edema, suffocation, and death. Chlorine has a vapor pressure of 31 psig at 26º F and a
vapor density of 2.5 (heavier than air). At atmospheric pressure, chlorine changes from a
36 ASTM International is a voluntary standards organization originally known as the American Society
for Testing and Materials.
37 A head shield is a supplemental heavy steel plate required by Federal regulation on the ends of some
hazardous materials tank cars to reduce the likelihood that a tank head will be punctured by the coupler of an
adjacent car in the event of excessive end impact or derailment.
38 Calculations made using a specific gravity for chlorine of 1.56.
39 The IDLH (immediately dangerous to life and health) value is an atmospheric concentration of any
toxic, corrosive, or asphyxiate substance that poses an immediate threat to life, or would cause irreversible
or delayed adverse health effects, or would interfere with an individual�s ability to escape from a dangerous
atmosphere. The Environmental Protection Agency uses 10 percent of the IDLH value when determining
that a release has reached a level of concern for public exposure.
Factual Information 29 Railroad Accident Report
liquid to a gas at �29º F. It is miscible40 in water. It is considered corrosive because it
forms hypochlorous acid and/or hydrochloric acid when combined with water.
The sodium hydroxide car contained 191,750 pounds (15,340 gallons)41 of a
50-percent solution of sodium hydroxide in water. The solution is corrosive, and
prolonged contact can cause permanent skin damage.
The fifth derailed tank car was shipped by Westvaco Corporation from De Ridder,
Louisiana. The car was listed as empty but contained a residue of rosin, which is
transported as an elevated temperature liquid. Rosin is normally a solid, and in this form is
not regulated by the DOT as a hazardous material. The material is heated to 212° F or
more for loading and unloading and at such temperatures is considered a burn hazard.
Tests and Research
Sight Distance Tests
On March 29, 2005, Safety Board investigators conducted sight distance tests at
the accident scene using locomotive equipment similar to that in use on train 192 at the
time of the accident. The tests were conducted at the same time of day as the accident and
under similar weather and lighting conditions. During the tests, the test engineer was
seated on the left (west) side of the locomotive, as had been the accident engineer. A test
conductor was seated on the right (east) side of the locomotive.
The test determined that the red reflection from the switch banner could be first
observed from the engineer�s seat at a distance of 1,461 feet from the switch as the train
transitioned from curved to tangent (straight) track. The red reflection was first visible on the
conductor�s side of the locomotive at 1,339 feet from the switch. The test also determined,
however, that the red reflection could not be identified by either crewmember as a switch
banner until the train was within 566 feet of the switch when the train headlight illuminated the
banner. The test revealed that the position of the switch points (which indicate which way the
switch is lined) could not be seen by the crew until the train was 220 feet from the switch stand.
The tests showed that a number of warning and signal lights were visible to the
crew of a train approaching the industry track switch. Canal Street and Trolley Line Road
intersected with Aiken Road about 393 feet north of the switch on either side of the NS
main line. The grade crossing at these intersections was protected with flashing lights and
bells. A total of eight pairs of red flashing lights at the intersections activated when a train
was 1,927 feet south of the crossing, or 1,534 feet from the switch. Additionally, about
281 feet north of the switch and just east of the main track was a wayside light indicating
the switch position at the Vaucluse Siding.42
40 Miscibility refers to the ability of a liquid or gas to dissolve uniformly in another liquid or gas.
41 Calculations made using a specific gravity for a 50-percent solution of sodium hydroxide of 1.5.
42 See the �Other Information� section of this report for more information on switch position indicator lights.
Factual Information 30 Railroad Accident Report
The driver of the taxi that transported the P22 crew on January 5 said that as he left
the area with the crew, he drove along Canal Street, on the west side of the main line. The
distance from the centerline of that roadway to the industry track switch stand measured
21 feet. Investigators determined that the position of the switch would have been clearly
visible to an occupant of a vehicle proceeding along the road.
Event Recorder Data
Each of the three locomotive units involved in the accident was equipped with an
event data recorder. The lead locomotive of train 192 used magnetic tape as a recording
medium; the trailing unit and the single locomotive of train P22 were equipped with solid-
state devices. Magnetic tape data were read on scene; the solid-state units were shipped to
Safety Board headquarters in Washington, D.C., where the information was downloaded
and analyzed by the Safety Board�s Vehicle Recorders Division.
The clock from each event recorder was correlated to the approximate reported
time of the collision. Using the Safety Board�s data analysis software, the data were
verified for accuracy by comparing the data recorded before the collision against the
operating range and limitations of the locomotives and against the known preaccident
operation of the train. The data recorded for traction motor current from the trailing
locomotive of train 192 was found to be inconsistent with the operation of the train. No
evidence was found of damage or recorder malfunction. Because damage from the
accident prevented the testing of the locomotive, the reason for the anomalous data
readings could not be determined. The remaining data from the trailing locomotive were
found to be consistent with the operating range and limitation of the equipment.
Train P22. The crew of local train P22 stated that the train arrived at the industry
track about 6:10 p.m. on January 5. Event recorder data indicate that the train came to a
stop at approximate recorder time 6:08:51 p.m. The train remained at a stop until recorder
time 6:09:57, after which the train made a series of forward and reverse movements with
the throttle handle and automatic brake handle being manipulated. The speed during these
movements remained below 11 mph. At recorder time 6:52:57 p.m., the speed indicated 0
mph; the throttle was in idle; and the automatic brake was in the 26-pound application
position. No further operational control changes were recorded.
Train 192. Event recorder data from the magnetic tape recorder on the lead
locomotive indicated that at approximate recorder time 2:10:41 a.m., consistent with the
reported time of departure from Augusta, the train brake was released and the speed began
to increase. Train speed continued increasing with the throttle handle in the idle position
until 2:12:36, when the throttle handle was moved to notch 3 and the speed indicated 9
mph. The train continued north with the throttle manipulated between the idle and notch 3
positions while the speed fluctuated between 9 and 12 mph.
At recorder time 2:22:30 a.m., the throttle handle was placed in notch 4, and the
speed began increasing. The speed continued increasing while the throttle handle was
manipulated between notch 3 and notch 8, and at 02:29:47, the data indicated a speed of 49
mph, which was the maximum speed attained during the trip from Augusta to Graniteville.
Factual Information 31 Railroad Accident Report
The train continued northbound with the speed fluctuating between 44 and 47 mph.
At recorder time 2:36:30 a.m., the throttle handle was increased to notch 8 while the train
traveled at 44 mph. Data indicated that the throttle handle remained in notch 8 until 2:38:11
with the train speed indicating 47 mph. The throttle was decreased to notch 6 with the speed
continuing to increase to 48 mph. At recorder time 2:38:37, the throttle handle was placed in
notch 4 with the speed remaining at 48 mph. At 2:39:00, approximately 467 feet from the
final resting point of the train,43 the speed indicated 47 mph. At 2:39:03, approximately 268
feet from the final resting point, traction motor current rose sharply, indicating an increase in
load on the motors. The indicated speed at that point was 42 mph, and the train was
decelerating rapidly. At 2:39:06, approximately 117 feet from the final resting point, the
automatic brake (train brakes) indicated an emergency; the independent (locomotive) brake
toggled on; the throttle was in the idle position; and the speed was 26 mph. A second, larger
spike in traction motor current occurred at 2:39:15. At 2:39:20 a.m., speed indicated 0 mph.
Laboratory Examination and Testing
The Safety Board�s Materials Laboratory examined a portion of the tank from the
ruptured ninth car on the train. This examination revealed that a 2 1/2-inch-long fracture
area in the center of the tank shell impact area, including the origin of the crack, separated
in a ductile mode. The remaining portions of the crack on either side of the ductile fracture
area separated in a brittle mode.
The Safety Board performed Charpy impact testing of the steel from the tank car
shell. The Charpy testing showed that the ductile-to-brittle transition temperature of the steel
plate from the shell of the ruptured tank was 40º F for specimens oriented transverse to the
rolling direction of the steel plate and 0º F for specimens oriented parallel to the rolling
direction of the steel plate. The estimated temperature of the chlorine (and the tank car steel)
at the time of the derailment was 26º F. The chemical composition and tensile properties of
the head and shell material met the specifications for AAR TC-128-B normalized steel.
Other Information
Accidents Involving Improperly Lined Switches
According to FRA safety data, �Human factors constitute the largest category of
train accident causes, accounting for 38 percent of all train accidents over the last 5 years.�
The data show that the leading cause [of human factor accidents] for 2004 was improperly
lined switches, which alone accounted for 16 percent of human factor accidents in the last
4 years.�44
43 Distance estimates are based on time and speed calculations; they do not account for any wheel
sliding or skidding that may have occurred during the accident sequence.
44 Douglas Taylor (Staff Director Operating Practices, FRA Office of Safety), �FRA�s Operating Rules
Working Group Spurs Action to Address Critical Safety Issues,� American Short Line Railroad Association
Safety Bulletin No. 9, August 2005, p. 2.
Factual Information 32 Railroad Accident Report
In 1999, the P22 conductor in the Graniteville accident was dismissed from service
for his failure to �properly restore the main track switch�resulting in your endangering
the safety and lives of other employees.�� Documentation pertaining to this incident,
which occurred on March 10, 1999, disclosed that a train, traveling about 30 mph,
encountered the misaligned switch at 3:05 a.m. the following morning. The engineer saw
the red switch target and placed the train in emergency braking, coming to a stop about 5
car lengths from the switch. The NS suspended the conductor from service on the day of
the incident and dismissed him from service after an investigation on March 31, 1999. The
conductor was reinstated on April 11, 1999.
On January 8, 2005, two days after the Graniteville accident, a Burlington
Northern Santa Fe Railway Company (BNSF) freight train was unexpectedly diverted
onto an industrial track in Bieber, California. The BNSF train struck two loaded grain
cars, derailing 7 locomotives and 14 cars. Two railroad employees were injured.
On September 15, 2005, during the preparation of this accident report, a southbound
Union Pacific (UP) freight train was unexpectedly diverted into a passing siding at Shepherd,
Texas, where it struck a northbound local train that had been parked on the siding track to
allow the southbound train to pass. The conductor of the parked train had earlier lined the
northernmost siding switch to allow the local to back from the main line into the siding where
it was to await the southbound train. To avoid violating Federal hours-of-service regulations,
the local train crew secured the train in the siding and, without relining the switch at the north
end of the siding, departed in a contracted taxi. When the taxi was about 20 miles from
Shepherd, the conductor realized he had left his keys in the switch lock at the siding and
insisted that the driver return to Shepherd. Meanwhile, a 3-person relief crew had boarded the
local train. As the southbound train approached, the local train�s engineer remained aboard the
locomotive while the other two crewmembers dismounted to inspect the passing train from
the ground. Instead of proceeding down the main line, the southbound train entered the siding
via the improperly lined switch and struck the standing train. The engineer of the local train
was killed, and four other crewmembers of the two trains were injured.
Postaccident FRA Safety Advisory
In response to the Graniteville accident, the FRA, on January 13, 2005, issued
Safety Advisory 2005-01, �Position of Switches in Non-Signaled Territory,� to
advise all railroads to review their operating rules and take certain other action
necessary to ensure that train crews who operate manual (hand-operated) main
track switches in non-signaled territory restore the switches to their normal
position after use.
The advisory informed railroads of the circumstances of the Graniteville accident
as well as of the January 8, 2005, BNSF accident.
The safety advisory referenced rules promulgated by the UP and the BNSF
railroads regarding the relining of switches. On October 1, 2004, the UP adopted a
requirement that before reporting clear of the limits of a track warrant, the crewmember
releasing the track warrant must first inform the train dispatcher that main track switches
Factual Information 33 Railroad Accident Report
have been restored to their normal positions. If the crew does not provide this information,
the train dispatching system prompts the train dispatcher to ask for it. The change was
made because of a September 29, 2004, collision that occurred at Thomaston, Texas, in
which a Texas Mexican Railway Company (TM) crew released a track warrant without
verifying that a siding switch was properly lined for the main track. A southbound UP
train entered the siding and collided with an unattended TM train.
On October 31, 2004, the BNSF adopted a requirement that before releasing a
track warrant, a train crew must report to the train dispatcher the position of any switch the
train has used. The dispatching system will not allow a track warrant to be cleared until the
dispatcher confirms the switch position through a job briefing with the crew. According to
the FRA advisory, this change was made as the result of a recommendation from a BNSF
division safety team. The safety team was concerned about incidents in which crews in a
particular subdivision had forgotten to reline main line switches. The BNSF rules
department issued this change across the BNSF system.
In the safety advisory, the FRA �strongly urged� railroads to (1) ensure that their
operating rules contain a provision similar to those of the UP and the BNSF that require train
crews who manually operate main line switches in non-signaled territory to report to the
dispatcher that the switches have been restored to the normal position before the crews
report clear of the main track, (2) require that the conductor of a crew working in non-
signaled territory, before reporting clear of main track, sign a switch position awareness
form that lists the name and location of any switch operated by any member of the crew and
the time each switch was relined for the main line, (3) require that the switch position
awareness form be submitted to a designated railroad official at the completion of each tour
of duty, (4) require that railroad officials review the forms for accuracy and to use the results
in the railroads� testing and proficiency programs, and (5) ensure that the revised rules,
procedures, and forms are immediately disseminated to all affected operating personnel.
The advisory stated that the FRA is considering the need for additional measures,
such as regulatory action or further advisories.
FRA Human Factors Working Group
On May 18, 2005, the FRA�s Railroad Safety Advisory Committee45 established an
operating rules working group to review the primary human factor causes of train
accidents and incidents and to recommend methods of reducing those accidents/incidents
and the resulting employee injuries. The working group was tasked with reporting its
findings and recommendations to the full committee by February 10, 2006.
FRA Emergency Order
In follow-up to Safety Advisory 2005-01, the FRA, on October 20, 2005, issued
Emergency Order No. 24, �Emergency Order Requiring Special Handling, Instruction and
45 The Railroad Safety Advisory Committee is made up of representatives of government, industry, and
other entities having an interest in railroad safety. The FRA established the committee in 1996 to advise the
FRA and develop consensus recommendations on safety issues.
Factual Information 34 Railroad Accident Report
Testing of Railroad Operating Rules Pertaining to Hand-Operated Main Track Switches.�
The order states that:
�public safety compels issuance of this Emergency Order (EO) requiring
railroads to modify their operating rules and take certain other actions necessary
to ensure that railroad employees who dispatch non-signaled territory or who
operate hand-operated main track switches�in non-signaled territory ensure the
switches are restored to their proper (normal) position after use.
In the �Background� section of the order, the FRA notes that the year 2004 saw a
marked increase in the �frequency and severity of collisions resulting from improperly
lined main track switches�.� The order states that after the issuance of the January safety
advisory, and with the exception of two accidents that occurred shortly after the advisory
was promulgated, �there was a respite of nearly six months in accidents resulting from
improperly lined switches in non-signaled territory.� However, in July 2005, two such
accidents occurred, and within a 28-day period in August and September, according to the
emergency order, three additional, more serious accidents occurred:
On August 19, 2005, a Kansas and Oklahoma Railroad freight train, operating at
26 mph in Nickerson, Kansas, encountered an improperly lined switch, entered a siding,
and struck a standing cut of cars. The train�s engineer was severely injured.
On August 21, 2005, a UP freight train, operating at 30 mph, was unexpectedly
diverted into a siding in Heber, California, where it struck a standing cut of cars. The control
compartment of the lead locomotive was destroyed. The three crewmembers �survived only by
quickly throwing themselves on the floor of the locomotive immediately before impact.�
The third serious accident was the previously described September 15, 2005,
accident on the UP railroad in Shepherd, Texas.
The July-through-September accidents, according to the FRA in its emergency
order, were a �clear indication that the Safety Advisory ha[d] lost its effectiveness.� The
emergency order was thus issued:
�to accomplish what the Safety Advisory could not: implement safety practices
that will abate the emergency until [the] FRA can complete rulemaking�.
The FRA states that, using the advice contained in the February 2006 report of the
previously referenced Railroad Safety Advisory Committee human factors working group,
its �goal is to publish a proposed rule in 2006, and a final rule soon thereafter.�
The emergency order identifies certain actions that must be taken by railroad
employees who operate hand-operated main line switches in non-signaled territory or who
dispatch in non-signaled territory. Those actions may be summarized as follows:
� All employees subject to the emergency order must receive both initial and
periodic instruction on the emergency order and on the railroads� operating
rules related to hand-operated switches in non-signaled territory. Railroads
must maintain records of the training.
Factual Information 35 Railroad Accident Report
� Employees who operate the switches are responsible for restoring them to their
normal positions.
� Employees releasing track authority must report to the dispatcher that all main
line switches have been restored to their appropriate positions. Additionally,
the dispatcher must confirm that the conductor and engineer have initialed the
switch position awareness form (which the FRA recommended in its safety
advisory that railroads adopt).
� Employees operating hand-operated switches must complete a switch position
awareness form, and all the information required by the form must be entered
before an employee reports clear of the limits of track authority.
� Job briefings must be held by employees in connection with the operation of
hand-operated main track switches in non-signaled territory. A briefing must
be conducted before the work is done, each time the work plan is changed, and
when the work is complete.
� Each time a train crewmember changes the position of a hand-operated main
line switch in non-signaled territory, the crewmember, while at the switch, shall
inform the engineer by radio of the switch name, location, and position. The
engineer must acknowledge the information before any movement can occur.
� Operational tests and inspections must incorporate the requirements of the
emergency order.
� Every affected employee will be provided a copy of the emergency order, and
receipt must be acknowledged in writing.
The emergency order provides relief for railroads that provide a level of safety the
FRA considers equivalent to the emergency order. As outlined in the order, relief is
automatically granted when (in reference to hand-operated switches in non-signaled
territory):
� A railroad�s operating rules require that trains approaching the switches be
prepared to stop;
� The switches are protected by distant switch indicators; or
� The switches are automatic or self-restoring and are protected by switch point
indicators [described in the next section of this report], unless those switches
are operated by hand.
The emergency order provides for a civil penalty of up to $27,000 for any person
(including individuals or corporate entities) violating the requirements of the order. Steps
to implement the order must begin immediately, with full implementation required by
November 22, 2005.
Switch Position Indicators
Some railroad switches, even in non-signaled territory, are equipped with switch
circuit controllers that can detect the switch position and alert the crew of an approaching
Factual Information 36 Railroad Accident Report
train if the switch is improperly lined. In signaled territory, the position of the switch is
conveyed to the train crew through a corresponding wayside signal. In non-signaled
territory, depending on the type of switch system in use, the crew can be alerted by
wayside indicator lights or by a radio message broadcast on the train crew�s radio channel.
The switches for the siding at Vaucluse,46 about 1.1 miles north of the industry track,
are equipped with switch position indicator lights. One wayside light at the switch signifies
the switch position, and if the switch is not lined for the main line, another wayside light
farther along the track alerts an oncoming train in time for it to slow and stop short of the
switch. Unlike the hand-operated switch at the industry track, the switches at Vaucluse are
�spring� switches.47 If a train is to use one of these switches to enter the siding from the main
line, a crewmember must still line the switch by hand and reline it for the main line after the
train is in the siding. But while the switch is lined for the main line, a train may move from
the siding onto the main line without the need to reposition the switch. As the train moves
through the switch, the train wheels force open the switch points for the movement. Once the
last wheels of the train have cleared the switch, a powerful spring, hydraulically assisted,
automatically returns the switch points to their normal position (lined for the main line).
Because a crew can move their train through a spring switch onto the main line
without manually manipulating the switch, no crewmember tends the switch after the train
has passed through. If the spring switch is equipped with a switch circuit controller, the
controller will detect the switch position and if, for example, an obstruction has prevented
the switch from returning to the normal position or if the switch has been left lined for the
siding, the lights at and in advance of the switch will display the appropriate aspect to the
next train approaching that location.
Postaccident Action by NS
Immediately after the accident, the NS amended its Operating Rule 181 to address
manually operated main track switches in non-signaled territory. Under the new rule
(Operating Rule 181a), train and engine crews, when reporting clear of track authority
limits in non-signaled territory, must advise the dispatcher or control operator of:
� The total number of hand-throw main track switches operated within the track
authority;
� The name and location of each main track switch operated;
� The restoration and securement of main track switches in their normal (main
line) position.
The rule states that train dispatchers or control operators within non-signaled
territory are not to clear a track authority until notified by the train or engine crew that each
hand-throw main track switch that has been operated has been locked in normal position.
46 This is a passing siding paralleling the main line with a switch at the north and south ends.
47 The NS has not equipped switches other than spring switches with switch position indicator lights or
other methods of alerting train crews of improperly lined switches.
Factual Information 37 Railroad Accident Report
Hours of Service Act
Congress enacted the original Hours of Service Act in 1907. The intent of the act
was to promote railroad safety by limiting the number of hours train crewmembers may
remain on duty and by requiring railroads to provide them with a minimum rest period
between shifts. The act, which has been revised and amended since its enactment, is
codified in FRA regulations at 49 Code of Federal Regulations Part 228 and applies to any
railroad employee engaged in or connected with the movement of a train. The act provides
that train employees may not remain on duty for more than 12 consecutive hours or 16
hours in an emergency.48 The Hours of Service Act further provides that a covered
employee must be given at least 8 consecutive hours off after a tour of duty of less than 12
duty hours, and 10 consecutive hours off after a tour of duty of 12 hours or more.
When a train crew cannot reach a scheduled or convenient crew change point
within 12 duty hours, the train must stop so that a replacement crew can take over. The
crewmembers being relieved are paid for any time they spend �deadheading� back to the
terminal; however, this time is referred to as �limbo time,� as it is not considered to be
time on duty or time off duty.49
The hours-of-service limits for the conductor and brakeman of train P22 expired at
7:00 p.m. After the crew arrived at the Aiken Yard, the conductor and brakeman finished
the required paperwork, and the entire crew was shown off duty at 8:11 p.m. Records
showed that on 10 of the 30 working days before the accident, train P22 crews had spent
time on paperwork after having been on duty for 12 hours in train or engine service. An
FRA representative responsible for enforcing the Hours of Service Act told the Safety
Board that any work, including paperwork, done on behalf of the railroad beyond the
allotted 12 hours is considered a violation of the act.
In Safety Board interviews, the P22 crew working the day before the accident said
that they were not aware that doing paperwork after 12 hours of operating a train could be
considered a violation of the Hours of Service Act. They indicated that they understood
the law covered only the movement of railroad equipment, not subsequent time spent in
administrative duties. NS managers told the Safety Board that they were unaware that
train crews were working outside the time limits. The computer time submissions by train
crews were used primarily for accounting and pay purposes and were not used by
operations managers to audit train crew activities. The FRA requires railroads to file
excess-service reports when covered employees exceed the 12-hour limit. The NS had
submitted no such reports for service performed by the P22 crew before the accident.
After the accident, the FRA cited the NS for violations of the Hours of Service Act.
48 The Hours of Service Act does not specifically refer to �emergency� or �emergencies,� but to
�casualties, unavoidable accidents, Acts of God,� and is interpreted by the FRA as �a cause not known to the
carrier or its officer or agent in charge of the employee at the time the said employee left the terminal, and
which could not possibly be foreseen.�
49 The amount of time that a train crew spends deadheading to (as opposed to from) an assignment is
considered on-duty time and is included when calculating total duty time.
38 Railroad Accident Report
Analysis
The Accident
NS train 192 was operating under track warrant authority on January 6, 2005,
when it departed Augusta, Georgia, bound for Columbia, South Carolina. The crew had
every reason to believe that no other trains would be occupying their track and that all
switches would be set for the main line. The 30-minute trip was uneventful from Augusta
to Graniteville.50
Train 192 continued operating at near the maximum authorized speed of 49 mph as
it approached the industry track switch, which was improperly lined for the industry track
instead of the main line. Event recorder data indicated that the train was traveling about 47
mph just before it encountered the switch and was diverted into the industry track. At a
point north of the turnout switch, as the train entered the 8º left-hand curve of the industry
track, the first car aft of the locomotives rolled toward the right and derailed. As the car
toppled to the right, it became uncoupled from the locomotives, and the train�s emergency
brakes applied. The movement of the first car off the rails also caused the aft end of the
second locomotive to derail.
Almost simultaneously with the separation of the first railcar, the front of the lead
locomotive of train 192 struck the front end of the locomotive of standing train P22, which
had been left on the industry track the previous evening. The derailment and impact likely
accounted for the spike in traction motor current recorded at 2:39:03 by the event recorder
in the lead locomotive of train 192. The speed of the locomotives, which were rapidly
decelerating, was about 42 mph at that point.
With the collision and derailment, the cars coupled behind the first railcar were
carried onward toward the derailed first railcar as a result of �run-in� momentum. The 2nd
through 14th cars in the consist then progressively derailed as they made contact with the
preceding derailed cars, with the individual cars coming to rest in a wreckage pileup. The
15th and 16th cars in the consist derailed but did not become entangled in the pileup.
The collision of the two locomotives derailed the standing train P22 locomotive
and propelled it about 217 feet northward (along the track). The P22 locomotive came to
rest about 5 feet south of the Hickman Street pavement edge, remaining upright and
somewhat aligned with the track.
The lead locomotive of train 192 traveled about 145 feet after the collision and
came to rest an estimated 77 feet south of the Hickman Street pavement edge, which is
about 487 feet to the north of the switch for the industry track turnout.
50 The accident scenario outlined in this section is based on postaccident inspections, physical evidence,
and event recorder data.
Analysis 39 Railroad Accident Report
The investigation revealed that the improperly lined switch had most recently been
used by the crew of local train P22 about 8 hours before the accident. The crew had lined
the switch for the industry track in order to place two cars at the Avondale Mills plant. No
crewmember remembered relining the switch for the main line before they boarded a taxi
and returned to the terminal.
Exclusions
Inspection records and postaccident examination revealed that the track and switch
in the derailment area were in good condition on the day of the accident. The non-derailed
cars from train 192 were tested, and no defects were found that would have affected the
operation of the train or altered its performance. The train 192 crewmembers were
qualified for their duties; they had received adequate time to rest before being called for
work; and they had been on duty for only about 2 hours 10 minutes when the accident
occurred. Postaccident toxicological testing was negative for the train 192 engineer and
conductor. The Safety Board therefore concludes that neither train equipment defects nor
track condition were causal or contributory to this accident. The Safety Board further
concludes that, in regard to the crew of train 192, fatigue, crew training and qualifications,
and drugs and alcohol were not factors in this accident.
The train P22 crew was experienced and qualified for the position each
crewmember held on the day before the accident. Each crewmember had had sufficient
off-duty time to have been well-rested before reporting for duty on January 5, 2005.
Because about 6 1/2 hours elapsed from the time the P22 crew went off duty until the
accident occurred, toxicological testing of the train P22 crew was not required or
performed. Crewmembers said they had not used alcohol or illicit drugs before or during
duty on January 5, and no evidence was found to suggest such use. The Safety Board
therefore concludes that, in regard to the crew of train P22, neither crew qualifications and
training nor fatigue were causal or contributory to this accident, and no evidence was
found to suggest drug or alcohol use.
Emergency Response
Local emergency response agencies were notified via 911 calls within about 1
minute of the accident. The fire chief and other resources of the Graniteville, Vaucluse, and
Warrenville Volunteer Fire Department were en route to the scene about 1 minute thereafter.
When the first-arriving responders reported breathing difficulties, the fire chief ordered his
resources to stand by and not approach the scene, which proved to be a prudent action. When
the fire chief arrived and was himself almost overcome by the toxic fumes, he directed a
mass evacuation of the area. Within about 13 minutes of dispatch, the fire chief began
marshaling his firefighting personnel and equipment at a staging area upwind of the toxic
gas release site. The fire chief also began establishing incident command, requested mutual
aid from nearby communities, requested weather information, and asked that the reverse 911
Analysis 40 Railroad Accident Report
system be activated. Within about 27 minutes of dispatch, equipment and personnel were
being successfully staged at what was later identified as a �forward command� site. Within
about 34 minutes of dispatch, all Aiken County fire resources were placed on standby.
Within 53 minutes of the initial dispatch, the first of four decontamination stations
was being organized. Within 69 minutes of dispatch, several firefighter entry teams,
wearing personal protective equipment and riding in privately owned pickup trucks, were
dispatched to the accident site. The teams transported individuals or groups of people who
had been exposed to the gas vapor to one of the decontamination stations before returning
to the scene to repeat the search and rescue cycle. The technique used by the entry teams,
which allowed them to rapidly cycle into and out of the �hot zone,� proved to be a
particularly efficient and expeditious means of evacuating those individuals who were in
the most danger because of their proximity to the accident site.
A review of the causes of death of the fatalities occurring in the field revealed that
the mechanism of death for all the fatalities was asphyxia that occurred within minutes of
exposure to the chlorine gas. This finding suggests that many, if not all, of the civilian
fatalities in this accident occurred within the minutes that elapsed before emergency
responders arrived on the scene or were able, because of the toxic fumes, to begin a safe
search and rescue effort.
Based on the promptness of the dispatch of emergency response resources to the
scene; the immediate implementation of the incident command system; the timely request
that additional mutual aid emergency response resources be dispatched to the scene; and
the relatively prompt search and rescue, evacuation, and decontamination efforts, the
Safety Board concludes that the execution of the emergency response to this accident was
timely, appropriate, and effective.
Performance of Train P22 Crew
The crew of local train P22, working out of Aiken, South Carolina, completed their
work along the NS main line in the Graniteville area about 7:00 p.m. on January 5, 2005.
No other trains or crews occupied or worked along the main line between the time the P22
crew left the industry track and the arrival of train 192 about 7 1/2 hours later.
The P22 conductor stated that his crew arrived at the industry track about 6:10
p.m. the day before the accident. Because the brakeman and conductor had been on duty
since 7:00 a.m., they had only 50 minutes to complete their work at the mill and safely
secure their train before reaching the maximum 12-hour limit imposed by Federal hours-
of-service regulations. The brakeman recalled being �pushed for the [hours of service]
law� as he worked in the Avondale Mills area. The engineer said that throughout the
workday, both the brakeman and conductor held job briefings at various places they
worked, but no such job briefing was held once the crew arrived at the industry track. The
engineer later speculated that the reason no job briefing was held might have been that the
conductor and brakeman were �in a hurry.�
Analysis 41 Railroad Accident Report
The brakeman told investigators he was sure �in his mind� that everything was
lined properly when he left the Avondale Mills area the evening before the accident. But
he also said he was not �100 percent� sure that he had relined the main line switch before
departing. Postaccident inspection revealed that the switch was lined and locked for the
industry track, as it had been when train P22 used the switch on the evening of January 5.
The switch showed no evidence of tampering, and no other trains used the track in the area
from the time the train P22 crew left until the accident the next morning. The Safety Board
therefore concludes that the crew of train P22 failed to reline a main line switch after using
it, leading to the subsequent and unexpected diversion of train 192 into an industry track
where it struck train P22 and derailed.
Perhaps the most vexing question when considering the circumstances of this
accident is how an experienced train crew could fail to execute a simple action⎯relining a
switch⎯that they had performed many times before and that, in fact, was a routine part of
their jobs. As the crew wrapped up their work at the industry track and departed the area,
none of them realized that a critical final task had been omitted.
One expert who studies the nature of errors is James Reason, professor emeritus of
psychology at the University of Manchester in Manchester, UK. Reason has characterized
errors of omission as a �particularly worthwhile target since the failure to carry out
necessary steps in the performance of a task is probably the single most common human
error type.�51
In a paper,52 Reason discusses the nature of omissions and how they occur. He
identifies a number of task features that are likely to increase the probability that a
particular step in a task will be omitted. The task features Reason identified are as follows:
� The greater the informational loading of a particular task step�that is, the
higher the demands imposed upon short-term memory�the more likely it is
that items within that step will be omitted.53
� Procedural steps that are functionally isolated, that is, ones that are not
obviously cued by preceding actions nor follow in a direct linear succession
from them, are more likely to be left out.
� Recursive or repeated procedural steps are particularly prone to omission. In
cases where two similar steps are required to achieve a particular goal, the
second step is the one most likely to be neglected.54
51 J. Reason, �How Necessary Steps in a Task Get Omitted: Revising Old Ideas to Combat a Persistent
Problem,� Cognitive Technology. 1998; 3:24-32.
52 J. Reason, �Combating Omission Errors Through Task Analysis and Good Reminders,� Quality and
Safety in Health Care. 2002; 11:40-44.
53 D. A. Norman, The Psychology of Everyday Things. New York: Basic Books, 1988.
54 C. Baber and N. A. Stanton, �Task Analysis for Error Identification: A Methodology for Designing
Error-Tolerant Consumer Products,� Ergonomics. 1994; 11:1923�41.
Analysis 42 Railroad Accident Report
� Necessary steps that follow the achievement of the main goal of a task are
likely to be omitted. This is an instance of a general principle: Steps near the
end of a task sequence are more prone to omission. Such �premature exits� are
due in part to the actor�s preoccupation with the next task, particularly when the
current activity involves largely routine actions.55
� Steps in which the item to be acted upon is concealed or lacking in conspicuity
are liable to omission.
� Steps following unexpected interruptions are especially prone to omission.
This can occur because the person loses his or her place in the action sequence
or because some unrelated action is unconsciously �counted in� as part of the
task sequence.56
� In tasks that involve departures from standard operating procedures or from
habitual action sequences, the intended actions may be supplanted by the more
frequently used routine in that context and thus omitted.
� Actions that are triggered by weak, noisy, or ambiguous signals are likely to be
omitted.
Reason discussed a study involving frequent users of office copiers that illustrated
the potential for omission errors created by the above task features. Based on these
features, researchers correctly predicted that the most common error reported by copier
users who manually copied multi-page documents would be the failure to remove the last
page of the original after copying was complete. This failure was predictable because:
� The emergence of the last copy page is a strong, but false, signal that the job is
complete. That is, the main goal of the activity (copying) is achieved before
completion of all the necessary steps.
� This false completion signal gains influence because of its proximity to the
presumed end of the activity. As the last page is copied, attention to the
subsequent task increases.
� As the last sheet is copied, it is no longer necessary to insert another original.
This leaves removing the last original page a functionally isolated act. Up until
this point, removal of an original page has been cued by the need to place the
next one.
� The closed copier lid conceals the last sheet of the original, so there is no
visible reminder to remove it.
The primary purpose of the exercise was to demonstrate that task features that are
likely to provoke frequent omissions can be identified and, once they are identified, steps
can be taken to reduce the frequency of such errors.
55 J. Reason, Human Error. New York: Cambridge University Press, 1990.
56 J. Reason, Human Error.
Analysis 43 Railroad Accident Report
Some of the task features that have been identified as leading to omission errors
were evident in the Graniteville accident. Some of the more important of these features are
as follows, illustrated by events from the accident sequence:
Procedural steps that are functionally isolated, that is, ones that are not obviously
cued by preceding actions nor follow in a direct linear succession from them, are
more likely to be left out.
When the crew arrived at the Avondale Mills plant, they immediately undertook a
series of steps that would permit them to achieve their objective of placing two cars at the
industry track. This included the conductor�s getting off the locomotive and leaving the
train to unlock gates and prepare switches near the industry track, the brakeman and
engineer�s continuing on toward the switch, the brakeman�s dismounting the train and
lining the switch for the industry track, the engineer�s backing the train into the industry
track, and the brakeman and conductor�s continuing with their switching and placement of
cars.
In this sequence of actions, each step proceeded in a direct linear fashion, one step
leading naturally to the next. Had the initial plan called for moving the train onto the
industry track and leaving it there without returning to the main line, relining the switch
after the train was on the industry track would have been a natural next step. Similarly, had
the crew been able to finish their work at the industry track and take the train to
Warrenville as they planned, relining the switch once the train returned to the main line
would have followed naturally in an organized sequence.
As it was, the 12-hour duty limit interrupted the process of planned, deliberate,
sequential steps. Now, instead of completing their job at the mill, continuing southbound,
and securing their train, the crew had to secure the train in place. This required that the
crew switch and place cars to get the train clear of the main line and road crossings and do
it within a strict time limit. They were just able to complete this work before their
maximum on-duty time expired.
The last-minute change in plans created a new sequence of events that left one
critical step functionally isolated⎯that of relining the main line switch. Because the train
did not return to the main line, relining the main line switch was no longer part of an
organized process of sequential steps and was thus neglected.
Necessary steps that follow the achievement of the main goal of a task are likely
to be omitted. This is an instance of a general principle: Steps located near the end
of a task sequence are more prone to omission.
The main objective for the crew once they arrived at Avondale Mills was to place
two cars within the plant. This objective was realized once the last car had been placed.
The conductor expressed relief that they had completed the job within their allowable
hours of service. In the minds of the crewmembers, the only task remaining was to secure
the equipment, retrieve their gear, and depart, which they did. The task of relining the
main line switch was a necessary step in terms of safety and efficient train operations, but
Analysis 44 Railroad Accident Report
it was not part of the immediate job at hand. In other words, the crew forgot to reposition
the switch, something that had been done many times before, because the primary
objective (switching cars) had been achieved, and they had already begun to focus on the
next task (securing the train and departing).
Steps in which the item to be acted upon is concealed or lacking in
conspicuity are liable to omission.
When the crew completed securing the train, the head end was about 342 feet from
the main line switch. Both the engineer and the conductor said the switch banner was not
visible at that point. Had the switch banner been conspicuous, it might have been detected
by a crewmember who would likely have realized that the switch was not properly lined.
And even though the switch position could have been detected as the crew passed along
the adjacent road on the way to the terminal, they had no reason to observe it and
apparently did not.
The Safety Board concludes that the crew of train P22 failed to reline the main line
switch for one or more of the following reasons: (1) the task of relining the switch was
functionally isolated from other tasks the crew was performing, (2) the crewmembers
were rushing to complete their work and secure their train before reaching their hours-of-
service limits, (3) the crew had achieved their main objective of switching cars and were
focused on the next task of securing their equipment and going off duty, and (4) the switch
was not visible to the crew as they worked, leaving them without a visual reminder to
reline the switch.
The conductor stated that he did not hold a job briefing at the industry track that
specifically addressed the switches at that location. Furthermore, he said that he did not
tell the brakeman to ensure that the switch was lined and locked for the main track. NS
rules require that a job briefing be held in this circumstance. Had a job briefing been held,
it would likely have included a discussion of the switches and specifically who was
responsible for ensuring that they were properly positioned. Had such a briefing taken
place, the relining of the switch might not have been overlooked. The Safety Board
concludes that had the conductor of train P22 held a comprehensive job briefing at the
Avondale Mills industry track, as required by NS operating rules, the crew may have
attended to the main line switch, and the accident may not have occurred.
After the accident, the FRA issued Safety Advisory 2005-01, which urged
railroads to review their operating rules and take certain steps to ensure that crews using
manually operated switches leave those switches in the proper position when their work is
complete. The advisory referenced rules already implemented by the BNSF and UP
railroads requiring that crews inform dispatchers of switch positions or inform them that
switches had been properly relined before reporting clear of main line track. These rules
were developed because of accidents similar to the one at Graniteville. The FRA also
urged the use of a switch position reporting form to be filled out by the conductor before
reporting clear of main line track.
Analysis 45 Railroad Accident Report
While any operating rule change designed to enhance safety is welcomed, the
Safety Board does not believe that rule changes or the use of forms is sufficient to prevent
recurrences of accidents such as the one at Graniteville. The Safety Board notes that only
2 days after the Graniteville accident, a BNSF freight train was unexpectedly diverted into
an industrial siding in California where it struck two loaded cars and derailed. This
accident occurred less than 3 months after the BNSF implemented the rule referenced in
the FRA advisory, a rule similar to those the FRA is urging other railroads to adopt and to
the rules adopted by the NS after the accident. The Safety Board further notes that the UP
had also adopted such a rule before the issuance of the advisory, but this did not prevent
the September 15, 2005, collision of a southbound UP freight train with a standing local
train in Shepherd, Texas, that resulted in a fatality and several injuries.
At Graniteville, the brakeman whose job it was to reline the switch said that he
believed everything was correct when he left the scene, and there is no reason to believe
that, even in his haste to return to the terminal, he would knowingly have left the switch
improperly lined. While it is possible that a discussion with the dispatcher specifically
regarding switches would have caused him to think through his actions and remember that
he had neglected the switch, it is also possible that during such a discussion he would
simply have confirmed his belief that he had left the site properly secured. He was
certainly aware that when he cleared the track warrants with the dispatcher he was
certifying that the main line was ready for use by other trains. He would not likely have
done this if he had any doubt about how he had left the track. Finally, under normal
conditions, the conductor would have cleared the track warrants with the dispatcher. He
likely would have assumed that the brakeman had relined the switch and would have
reported it to the dispatcher accordingly, especially if the brakeman had already departed.
Similarly, the use of forms, such as the switch position awareness form, has not
been shown to be particularly effective in preventing railroad accidents. For example,
some railroads, in order to lessen the chance that a traffic control signal will be missed or
misinterpreted by a crew, require that conductors record signal indications as they are
encountered en route. But the Safety Board has investigated a number of accidents in
which such forms, although required and used, did not prevent crews from missing signals
and causing accidents.
The FRA itself acknowledged the ineffectiveness of the safety advisory when, in
October 2005, it issued Emergency Order 24 in response to a number of accidents
involving improperly lined switches that occurred after promulgation of the advisory.
While the Safety Board acknowledges the timeliness with which the FRA has addressed
this safety issue, the Board is concerned about the effectiveness of the emergency order in
preventing future accidents. The primary concern of the Board is that the emergency order
largely requires what the previous safety advisory had recommended, which has been
shown to be of questionable effectiveness.
For example, it is not likely that a railroad employee qualified and authorized to
operate a hand-operated switch is unaware of the rules requiring that the switch be
returned to its proper position after work is complete. It is therefore unclear how
additional instruction on rules will improve employee performance. The emergency order
Analysis 46 Railroad Accident Report
also directs that an employee who operates a switch is responsible for returning it to its
normal position; however, NS operating rules placed responsibility accordingly, and this
did not prevent the Graniteville accident. The shortcomings of requiring the completion of
a switch position awareness form and of requiring that switch positions be confirmed with
the dispatcher have already been discussed. The additional operational tests required by
the emergency order may be expected merely to confirm that employees know their
responsibilities and that they fulfill those responsibilities most of the time. Such tests can
perhaps prevent errors of ignorance, but not of omission.
The emergency order goes beyond the safety advisory recommendations and
current regulations in at least two respects: by directing that job briefings be held at the
completion of work and by requiring that a train crewmember who repositions a main line
switch in non-signaled territory communicate with the engineer regarding the switch
position. The Safety Board welcomes these requirements as worthy additions to existing
requirements that could provide an additional layer of safety. As previously noted, a
comprehensive safety briefing was not held before the work at Graniteville. Had such a
briefing been held before and, more importantly, after the work (as required by the FRA
emergency order), the accident might have been avoided.
A significant element of the emergency order is the provision for a civil penalty of
up to $27,000 for violations of the order. The penalty may apply to the individual at fault
and/or to the company or other corporate entity. The magnitude of this penalty reflects the
seriousness with which the FRA views violations of this kind; however, it does not, in the
view of the Safety Board, address the cause of the violations. That is, the Safety Board
does not believe that employees forget to reline switches because the existing penalties are
inadequate. Employees are acutely aware that an improperly lined switch, in addition to
being a rule violation that could lead to removal from service, is likely to result in
significant property damage or the injury or death of fellow employees or innocent
bystanders. A substantial financial penalty is unlikely to be more effective than this
sobering prospect in preventing these types of accidents.
Moreover, the Safety Board is concerned that the significant civil penalty may
have an unintended impact on safety under some circumstances. That is, an employee
who, after leaving a work site, realizes that a switch has been left improperly lined may be
made more reluctant than in the past to immediately report the error to train dispatchers.
The threat of the severe fine may prompt the employee to attempt a remedy (such as
returning later to reline the switch) before the mistake can become known. As happened in
the September 2005 fatal collision in Shepherd, Texas, such action on the part of the
employee could contribute to an accident that might otherwise have been avoidable.
Clearly, measures beyond added or enhanced operating rules or additional forms,
or even severe penalties, are needed to ensure that accidents such as the one at Graniteville
do not recur. For example, a conspicuous visual stimulus associated with the switch at
Graniteville might have alerted the P22 crew to the position of the main line switch
despite any distractions.
Analysis 47 Railroad Accident Report
A conspicuous visual stimulus could take one of many forms. It could be a steady
or flashing strobe light (such as those used on some school buses and traffic signals) of a
color that would not be confused with other railroad signals. This would be analogous to
the �blue flag� procedures mandated by the FRA to draw particular attention when
personnel are working on, under, or between rail cars. The crew would probably have seen
a highly conspicuous light before leaving and would have relined the switch. Assuming
they had tied down the train out of sight of the switch (and had not traveled past it in
leaving) and had therefore left the switch improperly lined despite its conspicuity, a
unique flashing strobe or other obvious light might have alerted the train 192 crew to the
switch position in time to slow the train.
Alternatively, a device could be installed that would use electronic technology to
draw the crew�s attention to an improperly lined switch. Once an employee moved a
switch to a non-normal position, the device could monitor the employee�s proximity to the
switch. Should the employee leave the vicinity without relining the switch, a notification
could be sent to the employee�s pager or cell phone. If the employee failed to respond
within a specified time, the system could alert the railroad dispatcher or other designated
railroad employee.
The foregoing examples represent two possible means, one visual and one
electronic, of capturing an employee�s attention, but the Safety Board recognizes that
there are likely additional ways by which this objective could be achieved. The Safety
Board therefore believes that the FRA should require that, along main lines in non-
signaled territory, railroads install an automatically activated device, independent of the
switch banner, that will, visually or electronically, compellingly capture the attention of
employees involved with switch operations and clearly convey the status of the switch
both in daylight and in darkness.
Train Speeds in Non-Signaled Territory
The maximum authorized speed along the NS main line through Graniteville was
49 mph, and according to all available evidence, train 192 did not exceed this speed from
the time it left Augusta until it reached Graniteville. However, sight distance tests
demonstrated that the banner indicating the misaligned switch was not identifiable (by
investigators who were specifically looking for it) until the train was within about 566 feet
of the switch. To the crew of train 192, this distance might have been even less because of
the other lights and signals within the train crew�s visible range that may have created a
perceptual conflict. The Safety Board concludes that at the speed train 192 was traveling
as it entered Graniteville, the distance required for the train crew to perceive the banner of
the misaligned switch, react to it, and bring the train to a safe stop was greater than the
distance available.
Analysis 48 Railroad Accident Report
The Safety Board was concerned as early as 1974 about the issue of train speeds in
areas not under a form of centralized traffic control. As a result of its investigation of a
fatal accident in Cotulla, Texas, involving a misaligned switch in non-signaled territory,57
the Safety Board made the following safety recommendation to the Missouri Pacific
Railroad (now part of the Union Pacific Railroad):
R-74-22
Review your operation on main tracks that are not equipped with automatic
block signals and take appropriate action to ensure the capability of
engineers to stop trains in advance of misaligned switches. This action
could include reducing the size or speed of trains, installing automatic
block signals or advance-position indicators, or improving the visibility of
switch stand targets.
This recommendation was classified �Closed⎯No Longer Applicable� after the
Safety Board was provided with information indicating that the Missouri Pacific Railroad
would continue to evaluate territories for the possible installation of automatic block
signals or centralized traffic control.
At the time of the Cotulla accident, Interstate Commerce Commission (ICC) Order
29543 was in effect, which established a speed limit of �less than 50 mph� for freight
trains operating in non-signaled territory.58 The Safety Board�s investigation of the Cotulla
accident revealed that Order 29543 was inadequate in that the maximum allowable speed
was established without consideration of factors, such as visibility and stopping distances,
that at times may require lower speeds for safe operation. Therefore, the Safety Board
made the following safety recommendation to the FRA:
R-74-26
Determine and assess the current risks of train accidents involving
misaligned switches, collisions, broken rail, and other route obstructions on
main track where automatic block signal systems do not exist. Promulgate
regulations to replace Interstate Commerce Commission Order 29543.
These regulations should detail the major risks and controls assumed, set
guidelines for safe operations below the maximum operating speed, and
assign responsibility to the carrier for safe operations.
When the FRA issued regulations for signal and train control systems in January
1984, the wording of ICC Order 29543 was incorporated, unchanged, into the new
regulations. The Safety Board had intended that the new regulations specify circumstances
57 National Transportation Safety Board, Collision of Missouri Pacific Railroad Company Freight
Train Extra 615 South With a Standing Locomotive, Cotulla, Texas, December 1, 1973, Railroad Accident
Report NTSB/RAR-74/03 (Washington, D.C.: NTSB, 1974).
58 This speed limit does not apply along non-signaled track where train movements are governed by a
manual block system permanently in effect. See 49 Code of Federal Regulations 236.0(c).
Analysis 49 Railroad Accident Report
that required that trains be operated below the allowable maximum speed. Because the
FRA�s actions did not satisfy the Safety Board�s intent, Safety Recommendation R-74-26
was classified �Closed⎯Unacceptable Action.�
As acknowledged by the FRA, the frequency and severity of accidents involving
misaligned switches in non-signaled territory appear to be increasing. While at least some
of the measures the FRA has directed through its emergency order may aid in reducing the
number of switching mistakes, they are unlikely to eliminate such mistakes entirely.
Additional measures are therefore needed to help ensure that such mistakes, when they do
occur, do not result in accidents.
The Safety Board therefore believes that the FRA should require railroads, in non-
signaled territory and in the absence of switch position indicator lights or other automated
systems that provide train crews with advance notice of switch positions, to operate those
trains at speeds that will allow them to be safely stopped in advance of misaligned switches.
Hazardous Materials Release
During the derailment, the ninth car in the train, a tank car loaded with chlorine, was
punctured. Emergency responders observed that the B-end coupler of the 11th car in the
train, a car transporting steel coils, was in contact with the damaged tank jacket near the
puncture in the tank shell and was covered with frost. Leaking chlorine, which boils at
-29º F at atmospheric pressure, vaporizes rapidly from a liquid to a gas as it escapes
through an opening such as a puncture, thereby freezing water vapor in the air and causing
frost to form on nearby objects. Metallurgical examination of the damage on the shell
around the puncture documented several impression marks on the shell that matched
damage found on projecting surfaces of the coupler. The Safety Board therefore concludes
that the chlorine gas release that occurred in this accident resulted when the shell of the
9th car on the train was punctured by the coupler of the 11th car.
Photographs show that the punctured tank car came to rest angled slightly to one side
with the puncture opening extending below the midline of the resting tank. With the tank car
in this position, it is likely that the liquefied chlorine above the level of the puncture was
released within minutes and pooled on the ground. Measurements by emergency responders
of the liquid level of the chlorine remaining in the tank on the day of the accident revealed
that the tank car retained only about one-third of its original load of 180,000 pounds (13,830
gallons) of liquefied chlorine. Thus, approximately 120,000 pounds (9,218 gallons) of
liquefied chlorine were released before the responders arrived on scene.
The liquefied chlorine rapidly vaporized and expanded when it spilled from the tank
car. (The vaporization of liquefied chlorine at 32º F at atmospheric pressure can generate a
gaseous cloud with a volume 450 times greater than the volume of the liquid released.59)
As a result, the released chlorine created a large toxic cloud around the derailment site.
59 Richard J. Lewis, Jr., Hawley Condensed Chemical Dictionary-Thirteenth Edition. New York: John
Wiley & Sons, Inc., 1997.
Analysis 50 Railroad Accident Report
Because gaseous chlorine is 2 1/2 times heavier than air, the toxic cloud tended to settle in
low areas around the railroad tracks and remain more concentrated in these areas. Judging
from the locations of the fatalities, the cloud likely expanded to the west and into the plant,
followed the local topography running downhill to the south and west, and was blown to the
north by light winds. All fatalities resulting from the accident were caused by inhalation of
chlorine gas. Given that both train 192 crewmembers survived the collision (the engineer
died later from exposure to the gas), no fatalities or serious injuries would have resulted from
this accident had a tank car of chlorine not been punctured.
Tank Car Performance
Tank Car Crashworthiness
As previously noted, the 9th of 42 cars in the train was struck and punctured by the
coupler of the 11th car transporting steel coils. The combined weight of the striking steel
coil car and the rest of the trailing cars in the train was about 2,618 tons. The estimated
impact speed was determined to be about 42 mph. This combination of mass and velocity
subjected the punctured chlorine tank car to severe impact forces during the derailment,
with the most concentrated forces being applied in the area struck by the coupler.
The punctured tank car was built in 1993, and therefore was required to have both
the tank heads and the tank shell constructed of normalized steel. The normalizing heat
treatment typically increases the fracture toughness and lowers the ductile-to-brittle
transition temperature of steel plate. Thus, for a given composition of steel, normalized
steel is less susceptible to catastrophic brittle fractures and requires more energy to
fracture than non-normalized steel.
Charpy impact testing showed that the normalized steel in the tank shell of the
punctured chlorine car had a fracture toughness that was significantly greater than the
fracture toughness of the non-normalized steels of the catastrophically ruptured tank cars
involved in the derailment of a Canadian Pacific Railway freight train in Minot, North
Dakota, in January 2002.60 The steel in the Minot tank cars exhibited relatively low
fracture toughness, and cracks propagated rapidly around the circumference of each tank.
The higher fracture toughness in the Graniteville tank car contributed to the relatively
quick arrest of the crack even though there was brittle fracture in its outer portions.
Chlorine tank cars such as the punctured ninth car are tested to a pressure of 500 psig
compared to a test pressure of 300 psig for tank cars used to transport anhydrous ammonia and
liquefied petroleum gas. To be rated for the increased operating pressure, the tanks of chlorine
tank cars must have greater tank wall thicknesses than tanks of the lower pressure cars.
Because of the improved properties of normalized steel and the increased wall thickness, the
60 National Transportation Safety Board, Derailment of Canadian Pacific Railway Freight Train 292-
16 and Subsequent Release of Anhydrous Ammonia Near Minot, North Dakota, January 18, 2002, Railroad
Accident Report NTSB/RAR-04/01 (Washington, D.C.: NTSB, 2004).
Analysis 51 Railroad Accident Report
punctured car was among the strongest tank cars currently in service. The Safety Board
therefore concludes that, as shown in the Graniteville accident, even the strongest tank cars in
service can be punctured in accidents involving trains operating at moderate speeds.
The Safety Board addressed the improvement of crashworthiness of railroad tank
cars in its Minot, North Dakota, accident report. The Board stated in its report:
Improvements in the crashworthiness of pressure tank cars can be realized
through the evaluation of alternative steels and tank car performance standards.
The ultimate goal of this effort should be the construction of railroad tank cars that
have sufficient impact resistance and that eliminate the risk of catastrophic brittle
failures under all operating conditions and in all environments. Achieving such a
goal does not necessarily require the construction of a tank car that is puncture�
proof; it may only require construction of a car that will remain intact and slowly
leak its contents if it is punctured.
To address these concerns, the Board recommended that the FRA:
R-04-6
Validate the predictive model the Federal Railroad Administration is
developing to quantify the maximum dynamic forces acting on railroad
tank cars under accident conditions.
R-04-7
Develop and implement tank car design-specific fracture toughness
standards, such as minimum average Charpy value, for steels and other
materials of construction for pressure tank cars used for transportation of
U.S. Department of Transportation class 2 [gases] hazardous materials,
including those in low-temperature service. The performance criteria must
apply to the material orientation with the minimum impact resistance and
take into account the entire range of operating temperatures of the tank car.
On August 9, 2004, the FRA responded and described the actions being taken to
address each recommendation. In response to Safety Recommendation R-04-6, the FRA
stated that it has identified ongoing programs at the Volpe National Transportation
Systems Center and the University of Illinois at Chicago to evaluate in-train forces
associated with train derailments. The FRA said it anticipates that the modeling program
will be completed in early 2006. Regarding Safety Recommendation R-04-7, the FRA
stated that further research in this area is required and that it may require a �three-year
effort� to develop adequate tank car design-specific fracture toughness standards.
On June 22, 2005, in addressing the FRA�s response to Safety Recommendation R-
04-6, the Safety Board noted that programs to analyze in-train forces have already been
identified and that it expects validation of the models to be a standard part of any model
development. Based on FRA�s response, Safety Recommendation R-04-6 was classified
�Open⎯Acceptable Response.� In addressing the FRA�s response to Safety
Recommendation R-04-7, the Safety Board stated that implementation of tank car design-
Analysis 52 Railroad Accident Report
specific fracture toughness standards, such as Charpy impact value, can be achieved for
standard manufacturing processes without waiting for the results of the modeling effort
associated with Safety Recommendation R-04-6. The Safety Board added that evaluation
and analysis of the dynamics of the Minot accident can provide data about the levels of
fracture toughness that may be necessary for pressure tank cars and that data from
subsequent accidents in Macdona, Texas, on June 28, 2004, and Graniteville will provide
additional information. Based on the FRA�s response, Safety Recommendation R-04-7
was classified �Open⎯Unacceptable Response.�
Congress also recognized the significance of the Safety Board�s safety
recommendations to the FRA by incorporating them into the �Safe, Accountable, Flexible,
Efficient Transportation Equity Act,� which was signed into law in August 2005. Section
20155 of the act stipulated that the FRA was to validate the predictive model within 1 year
of enactment and initiate a rulemaking to implement appropriate design standards for tank
cars within 18 months of enactment.
The Macdona and Graniteville accidents, both of which have occurred since the
Minot report was issued, resulted in the puncturing of two chlorine tank cars and the death
of 12 people from chlorine inhalation. When a liquefied gas such as chlorine, which is
poisonous by inhalation, is released, large clouds at lethal concentrations can be generated
within minutes. There is often little or no time to alert citizens and to take effective action.
Based on AAR data on tank car shipments in the United States for 2002, chlorine and
anhydrous ammonia ranked as, respectively, the fourth and seventh most commonly
shipped hazardous materials by tank car. Furthermore, these products routinely travel
through communities of varying size, including large metropolitan areas.
Reduction of Tank Car Vulnerability
It is the belief of the Safety Board that modeling of accident forces and the
application of fracture toughness standards as recommended in the Minot report will
provide the most effective improvements in the crashworthiness of tank cars. However, at
best, it will be several years before a significant percentage of pressure tank cars in service
will have been so designed and constructed. Therefore, the most expedient and effective
means to reduce the public risk from the release of highly poisonous gases in train
accidents is for railroads to implement operational measures that will minimize the
vulnerability of tank cars transporting these products.
Supplemental operational measures are already imposed for the transportation of
certain high-risk materials. For example, the DOT�s Pipeline and Hazardous Materials
Safety Administration requires that pipeline operators have an integrity management plan
for high-consequence areas, which are identified on the basis of population densities and
environmentally sensitive areas. The regulations are designed to identify high-risk areas and
to develop a process for evaluating the risks within areas identified as high-consequence.61
61 Under the regulations, a high-consequence area may include an urban area with a population greater
than 50,000, or a population density of 1,000 people per square mile, or other area (an unincorporated town
or village, for example) that contains a concentrated population.
Analysis 53 Railroad Accident Report
The integrity management plan for pipeline operators describes both preventive
and mitigative measures that a pipeline operator must take to protect a high-consequence
area. Such measures include implementing enhanced damage prevention practices, better
monitoring, shorter inspection intervals, improved system monitoring and detection, and
additional personnel training/drills with emergency responders.
For rail transportation of hazardous materials, the AAR since 1990 has published
Circular No. OT-55, Recommended Railroad Operating Practices for Transportation of
Hazardous Materials. This circular contains recommended operating practices for
member railroads that include speed restrictions for �key trains� and enhanced track
inspection standards for �key routes.�62 Because the train involved in the Graniteville
accident was not a key train and the main line track was not on a key route, neither the
operational restrictions nor additional inspections applied. Further, even if train 192 had
met the definition of a key train, Circular No. OT-55 would not have restricted its speed
below that at which it was already operating.
Two research studies have also been conducted that address operational measures
to reduce the vulnerability of tank cars transporting hazardous materials. The 1992 FRA
report, Hazardous Materials Car Placement in a Train Consist,63 concluded that the rear
one-quarter of a train is the most desirable location for cars containing hazardous
materials and that reducing the speed and size of trains can reduce the number of cars
derailed in an accident. The second study, �Minimizing Derailments of Railcars Carrying
Dangerous Commodities Through Effective Marshaling Strategies,�64 prepared for the
Transportation Research Board, reached similar conclusions and provided some additional
statistical information to validate those conclusions.
Both these reports address operational measures that might have made a difference
in the Graniteville accident. Placement of the three tank cars transporting chlorine near the
front of the train and ahead of most of the trailing tonnage increased the probability that
the cars would be damaged and would release chlorine in an accident. Had the chlorine
cars been placed behind the other loaded cars in the train, the reduction in the trailing
tonnage would have reduced the impact forces on the tank cars. A reduction in train speed
would also have significantly reduced the derailment forces on the tank cars. These
operational measures, taken individually or collectively, may have been sufficient to
prevent the puncture of the tank car and the release of the chlorine.
While the FRA report notes that car placement might be detrimental to train
handling and dynamics and that switching cars to change their order in the train might
62 Under the recommended practices, a key train includes any train with five tank car loads of poison
inhalation hazard (PIH) cargo; or 20 carloads of a combination of PIH, flammable gas, explosives, and
environmentally sensitive chemicals; or one or more carloads of high-level radioactive waste. A key train
cannot exceed 50 mph.
63 R. E. Thompson, E. R. Zamejc, and D. R. Ahlbeck, Hazardous Materials Car Placement in a Train
Consist, Vol. 1 (Review and Analysis). Report DOT/FRA/ORD/18.I. Federal Railroad Administration, U.S.
Department of Transportation (Washington, D.C., 1992).
64 F. F. Saccomanno and S. El-Hage, �Minimizing Derailments of Railcars Carrying Dangerous
Commodities Through Effective Marshaling Strategies,� Transportation Research Record 1245.
Transportation Research Board, National Research Council (Washington, D.C., 1989).
Analysis 54 Railroad Accident Report
result in exposing these cars to additional dangers, railroads should be able to take these
factors into account and still reduce the vulnerability of tank cars transporting chlorine,
anhydrous ammonia, and other liquefied gases that are poisonous by inhalation. Given the
risks involved in the transportation of such liquefied gases, the Safety Board believes that
the FRA should require that railroads implement operating measures, such as positioning
tank cars toward the rear of trains and reducing speeds through populated areas, to
minimize impact forces from accidents and reduce the vulnerability of tank cars
transporting chlorine, anhydrous ammonia, and other liquefied gases designated as
poisonous by inhalation.
Train Crew Protection From Inhalation Hazards
The Safety Board has found that freight train crews may survive collisions and
derailments only to be injured or killed by hazardous materials released in the accident.
Although the crew of NS train 192 survived the collision and exited the locomotive
unassisted, they could not escape exposure to the chlorine gas. The conductor said that
after getting out of the locomotive, he and the engineer were able to walk some distance
from the collision site. The two were transported to hospitals. The conductor was treated
and released; the engineer died several hours later from inhalation of the toxic gas.
The consequences of this accident are remarkably similar to those of the June 28,
2004, collision of two freight trains near Macdona, Texas. A tank car on the striking train
was punctured and released chlorine gas. Once again, the crew of the striking train
survived the collision and exited the locomotive unassisted into a chlorine-laden
atmosphere. The conductor and engineer had walked about 1,400 feet away from the
collision site when the conductor collapsed and died from exposure to chlorine gas. The
engineer was hospitalized with severe injuries due to his exposure.
Emergency breathing apparatus is commercially available that would give
crewmembers in these circumstances an opportunity to escape a hazardous atmosphere.
According to the manufacturers, many of these devices are approved for use to escape
certain chemical atmospheres, including chlorine and ammonia, as well as fire and smoke.
Emergency escape breathing devices are typically effective for a period of time (5 to 50
minutes) that allow the user to escape and reach a safe location. The devices are used in a
variety of industrial settings. They must also be carried on merchant and passenger vessels
under the Safety of Life at Sea protocols. The Safety Board concludes that had the
engineer of train 192 been wearing appropriate, fully functioning emergency escape
breathing apparatus when he walked away from the collision site, he may not have
succumbed to the effects of chlorine gas inhalation.
The Safety Board therefore believes that the FRA should determine the most
effective methods of providing emergency escape breathing apparatus for all
crewmembers on freight trains carrying hazardous materials that would pose an inhalation
hazard in the event of unintentional release, and then require railroads to provide these
breathing apparatus to their crewmembers along with appropriate training.
55 Railroad Accident Report
Conclusions
Findings
1. Neither train equipment defects nor track condition were causal or contributory to this
accident.
2. In regard to the crew of train 192, fatigue, crew training and qualifications, and drugs
and alcohol were not factors in this accident.
3. In regard to the crew of train P22, neither crew qualifications and training nor fatigue
were causal or contributory to this accident, and no evidence was found to suggest
drug or alcohol use.
4. The execution of the emergency response to this accident was timely, appropriate, and
effective.
5. The crew of train P22 failed to reline a main line switch after using it, leading to the
subsequent and unexpected diversion of train 192 into an industry track where it
struck train P22 and derailed.
6. The crew of train P22 failed to reline the main line switch for one or more of the
following reasons: (1) the task of relining the switch was functionally isolated from
other tasks the crew was performing, (2) the crewmembers were rushing to complete
their work and secure their train before reaching their hours-of-service limits, (3) the
crew had achieved their main objective of switching cars and were focused on the
next task of securing their equipment and going off duty, and (4) the switch was not
visible to the crew as they worked, leaving them without a visual reminder to reline
the switch.
7. Had the conductor of train P22 held a comprehensive job briefing at the Avondale
Mills industry track, as required by Norfolk Southern operating rules, the crew may
have attended to the main line switch, and the accident may not have occurred.
8. At the speed train 192 was traveling as it entered Graniteville, the distance required
for the train crew to perceive the banner of the misaligned switch, react to it, and
bring the train to a safe stop was greater than the distance available.
9. The chlorine gas release that occurred in this accident resulted when the shell of the
9th car on the train was punctured by the coupler of the 11th car.
10. As shown in the Graniteville accident, even the strongest tank cars in service can be
punctured in accidents involving trains operating at moderate speeds.
Conclusions 56 Railroad Accident Report
11. Had the engineer of train 192 been wearing appropriate, fully functioning emergency
escape breathing apparatus when he walked away from the collision site, he may not
have succumbed to the effects of chlorine gas inhalation.
Probable Cause
The National Transportation Safety Board determines that the probable cause of
the January 6, 2005, collision and derailment of Norfolk Southern train 192 in
Graniteville, South Carolina, was the failure of the crew of Norfolk Southern train P22 to
return a main line switch to the normal position after the crew completed work at an
industry track. Contributing to the failure was the absence of any feature or mechanism
that would have reminded crewmembers of the switch position and thus would have
prompted them to complete this final critical task before departing the work site.
Contributing to the severity of the accident was the puncture of the ninth car in the train, a
tank car containing chlorine, which resulted in the release of poisonous chlorine gas.
57 Railroad Accident Report
Recommendations
As a result of its investigation of the January 6, 2005, collision and derailment of
Norfolk Southern train 192 in Graniteville, South Carolina, the National Transportation
Safety Board makes the following safety recommendations:
To the Federal Railroad Administration:
Require that, along main lines in non-signaled territory, railroads install an
automatically activated device, independent of the switch banner, that will,
visually or electronically, compellingly capture the attention of employees
involved with switch operations and clearly convey the status of the switch
both in daylight and in darkness. (R-05-14)
Require railroads, in non-signaled territory and in the absence of switch
position indicator lights or other automated systems that provide train
crews with advance notice of switch positions, to operate those trains at
speeds that will allow them to be safely stopped in advance of misaligned
switches. (R-05-15)
Require railroads to implement operating measures, such as positioning
tank cars toward the rear of trains and reducing speeds through populated
areas, to minimize impact forces from accidents and reduce the
vulnerability of tank cars transporting chlorine, anhydrous ammonia, and
other liquefied gases designated as poisonous by inhalation. (R-05-16)
Determine the most effective methods of providing emergency escape
breathing apparatus for all crewmembers on freight trains carrying
hazardous materials that would pose an inhalation hazard in the event of
unintentional release, and then require railroads to provide these breathing
apparatus to their crewmembers along with appropriate training. (R-05-17)
BY THE NATIONAL TRANSPORTATION SAFETY BOARD
MARK V. ROSENKER ELLEN ENGLEMAN CONNERS
Acting Chairman Member
DEBORAH A. P. HERSMAN
Member
Adopted: November 29, 2005
59 Railroad Accident Report
Appendix A
Investigation
The National Response Center notified the National Transportation Safety Board
of the accident about 4:15 a.m. on January 6, 2005. The investigator-in-charge and other
members of the Safety Board investigative team were launched from the Washington,
D.C., headquarters office and from the Atlanta, Georgia, field office. Investigative groups
were established to study operations, track, signals, mechanical, survival factors, human
performance, and hazardous materials issues. No hearings or depositions were held in
conjunction with this accident. Member Deborah A. P. Hersman was the Board Member
on scene.
Parties to the investigation included the Federal Railroad Administration, Norfolk
Southern Railway Company, the Brotherhood of Locomotive Engineers and Trainmen, the
United Transportation Union, General American Transportation, Union Tank Car
Company, Trinity Industries, Olin Corporation, the Graniteville, Vaucluse, and
Warrenville Volunteer Fire Department, and the Aiken County Sheriff�s Office.
- Cover
- Title Page
Abstract
Contents
Executive Summary
Factual Information
Accident Synopsis
Site Description
Events Preceding the Accident
The Accident
Emergency Response
Fire Department Response
Sheriff’s Office Response
Hazardous Materials Response
Injuries
Damages
Hazardous Materials Cars
Other Railroad Equipment
Emergency Response Equipment
Wreckage
Personnel Information
Train P22
Train 192
Toxicological Testing
Meteorological Information
Operations and Track Information
Postaccident Inspections
Switch and Track
Railcars
Tank Car and Hazardous Materials Information
Tank Cars
Hazardous Materials
Tests and Research
Sight Distance Tests
Event Recorder Data
Laboratory Examination and Testing
Other Information
Accidents Involving Improperly Lined Switches
Postaccident FRA Safety Advisory
FRA Human Factors Working Group
FRA Emergency Order
Switch Position Indicators
Postaccident Action by NS
Hours of Service Act
Analysis
The Accident
Exclusions
Emergency Response
Performance of Train P22 Crew
Train Speeds in Non-Signaled Territory
Hazardous Materials Release
Tank Car Performance
Tank Car Crashworthiness
Reduction of Tank Car Vulnerability
Train Crew Protection From Inhalation Hazards
Conclusions
Findings
Probable Cause
Recommendations
Appendix A
Investigation
- null: this page intentionally left blank
