November 10, 2001

From the Conrail Technical Society:

Memorandum on the CSXT Ohio Runaway Train (2001)
Memorandum on CSXT Runaway:

In accordance with your instructions, the following report summarizes
the work of the investigation committee which was assigned to review and
analyze the events of May 15, 2001, in which a locomotive with cars
departed from Stanley Yard on the CSXT near Toledo, Ohio and traveled
south to Kenton, Ohio, with no crew member on board. I was assisted in
this review by MPE Inspector Mike Lusher and OP Inspector Ed Scalzitti,
both of whom responded initially to the incident, and by Chief Inspector
Harold Rugh, who provided essential technical information.

On May 15, 2001, at approximately 12:35 p.m., DST, an unmanned CSX yard
train consisting of one model SD-40-2 locomotive, 22 loaded, and 25
empty cars, 2898 gross trailing tons, departed Stanley Yard, which is
located in Walbridge, Ohio. The uncontrolled movement proceeded south
for a distance of 66 miles before CSX personnel were able to bring the
movement under control. At the time of the incident, the weather was
cloudy with light rain. The ambient temperature was 55 degrees
Fahrenheit. There was no derailment of equipment or collision. There
were no reportable injuries as a result of the incident.

Circumstances Prior to Incident: 
Yard crew Y11615, consisting of one engineer, one conductor and one
brakeman, reported for duty at Stanley Yard, Walbridge, Ohio, at 6:30
am, DST, May 15, 2001. After the normal job briefing with the
trainmaster, crew Y11615 performed routine switching assignments until
approximately 11:30 am, at which time the crew received new instructions
and a second job briefing. A few minutes before 12:30 p.m., the crew
entered the north end of track K12, located in the classification yard,
with the intent to pull 47 cars out of K12 and then place these cars on
departure track D10. Locomotive CSX 8888 was positioned with the short
hood headed north. The engineer was seated at the
controls on the east side of the locomotive.

The locomotive coupled to the 47 cars on track K12, as instructed and
planned. The air hoses between the locomotive and the cars were not
connected, as is normal during this kind of switching operation. The air
brakes on the cars were therefore inoperative. The brakeman notified the
engineer by radio to pull north from K12. After the rear or 47th car
passed the brakeman's location, he walked west to position the switch
for the reverse movement to proceed into the assigned track D10.

The movement continued north out of K12 passing the conductor, who was
positioned on the ground at the "Camera" switch. The conductor advised
the engineer by radio of the number of cars that had passed him and
received an acknowledgement from the engineer by radio.

The Incident:
With eight cars remaining to pass over the "Camera" switch, the
conductor notified the engineer by radio to prepare to stop. The
engineer did not respond to his communication. The conductor again
notified the engineer when four cars remained to clear the switch, but
again there was no response from the engineer. The conductor then
ordered the engineer to stop movement, but again there was no response
from the engineer and the movement continued.

In his interview, the engineer stated that as he pulled north out of
K12, he was notified by radio by the conductor that the trailing point
switch for track PB9 off the lead was reversed. The engineer understood
that it would be necessary for the movement to be stopped short of the
PB9 switch in order to line the switch for movement further along the
lead. Neither the conductor nor the brakeman were near the PB9 switch,
and the engineer intended to stop his train, dismount from the
locomotive, and align the switch to its normal position, if necessary.
The speed of the movement up the lead had now reached 11 mph. The
engineer observed the reversed switch, but due to the wet rail
conditions and the number of cars coupled to his locomotive, he foresaw
that he could not bring the equipment to a stop prior to passing through
the misaligned switch. The engineer responded by applying the
locomotive's independent brake to full application. The independent
brake applies the brakes on the locomotive but does not apply the brakes
on the individual freight cars. In addition, he reduced brake pipe
pressure with a 20 psi service application of the automatic brake valve.
The automatic brake is pneumatic braking system designed to control the
brakes on the entire train. Still certain he would not stop short of the
switch the engineer attempted to place the locomotive in dynamic brake
mode. The dynamic brake utilizes the locomotive propulsion system to
brake the train. Dynamic braking is analogous to down shifting a truck
or automobile. Unfortunately, the engineer inadvertently failed to
complete the selection process to set up the dynamic brake. Under the
mistaken belief that he had properly selected dynamic brake, the
engineer moved the throttle into the number 8 position for maximum
dynamic braking. The engineer believed that the dynamic brake had been
selected and that additional braking would occur.

However, since dynamic brake set up had not been established, the
placing of the throttle into the number 8 position restored full
locomotive power, instead of retarding forward movement of the train.

While the train was still moving at a speed of approximately 8 mph, the
engineer dismounted the locomotive and ran ahead to reposition the
switch before the train could run through and cause damage to the
switch. The engineer was successful in operating the switch just seconds
before the train reached it. The engineer than ran along side the
locomotive and attempted to reboard. However, the speed of the train had
not decreased as the engineer had expected but had increased to
approximately 12 mph. Due to poor footing and wet grab handles on the
locomotive, the engineer was unable to pull himself up on the
locomotives ladder. He dragged along for approximately 80 feet until he
released his grip on the hand rails and fell to the ground. Unable to
reboard and stop the movement of his train, the engineer ran to contact
a railroad employee, not a member of his crew but in possession of a
radio, located at the north end of the yard. This employee immediately
notified the yardmaster of the runaway train. The yardmaster promptly
notified the Stanley tower block operator and the trainmaster. The
Toledo Branch train dispatcher located in Indianapolis was also
notified. The movement was now
proceeding southward on the Toledo Branch (Great Lakes Division)
governed by Traffic Control System (TCS) Rules. The time was
approximately 12:35 p.m.

The brakeman observed the train depart the yard but did not initially
see the engineer on the ground. The brakeman and another employee used a
personal vehicle to pursue the train to the next grade crossing to
attempt to board the train. Their immediate concern was for the safety
of the engineer, who they feared may have suffered a heart attack while
at the controls of the locomotive. At the grade crossing, the two
employees were unable to board the train as the speed had increased to
approximately 18 mph as it passed the milepost 4.

Local authorities and the Ohio State Police were notified of the runaway
train at
approximately 12:38 p.m.

Attempts to Stop the Runaway:
At a siding called Galatea, near milepost 34, at approximately 1:35
p.m., the train dispatcher remotely operated the switch for the train to
enter the siding. Previously a portable derail had been placed on the
track in an attempt to derail the locomotive and thereby stop the
movement. The portable derail was, however, dislodged and thrown from
the track by the force of the train passing over it, and the movement of
the train was not impeded.

Northbound train Q63615 was directed by the dispatcher into the siding
at Dunkirk, Ohio. The crew was instructed to uncouple their single
locomotive unit from their train and wait until the runaway passed their
location. at approximately 2:05 p.m., the runaway train passed Dunkirk,
and the siding was lined for the crew of Q63615 to enter the main track
and to pursue the runaway train.

At Kenton, Ohio, near milepost 67, the crew of Q63615 successfully
caught the runaway equipment and succeeded in coupling to the rear car,
at a speed of 51 mph. The engineer gradually applied the dynamic brake
of his locomotive, taking care not to break the train apart. By the time
the train passed over Route 31 south of Kenton, the engineer had slowed
the speed of the train to approximately 11 mph. Positioned at the
crossing was CSX Trainmaster Jon Hosfeld, who was able to run along side
the unmanned locomotive and climb aboard. The trainmaster immediately
shut down the throttle, and the train quickly came to a stop. The time
was 2:30 p.m. and the runaway train had covered 66 miles in just under 2

An examination of the controls confirmed that the locomotive independent
brake had been fully applied, automatic brake valve was in the service
zone, and the dynamic brake selector switch was not in the braking mode.
All brake shoes had been completely worn to the brake beams.

The railroad was prepared to place an additional fully manned locomotive
ahead of the runaway south of Kenton, if necessary, to further slow the
train. This rather hazardous option was fortunately not required.

Post Incident Investigation:
The engineer of Y11615 was slightly injured, but he declined medical
treatment. He was released from service with his crew at 5:30 p.m. The
CSX did not require Drug or Alcohol testing of the crew, nor was federal
testing required.

The engineer first hired on the Pennsylvania Railroad in 1966, and he
was promoted to engine service in 1974. He received his most recent
check ride with a supervisor in January 2001. The engineer's discipline
record is clean.

A Federal Railroad Administration Motive Power and Equipment Inspector
arrived at the location where CSX 8888 was stopped and performed a full
mechanical inspection. He found all systems to function normally,
including sanders, headlight, auxiliary lights, bell, horn and the
alerter. The brake cylinder piston travel could not be determined,
because all brake shoes were completely burned off. 

Air Brake System: 
Locomotive CSX 8888 is a model SD40-2 manufactured by General Motors
Corporation (EMD). This unit is equipped with 26L type air brake system.
The Alerter system is connected directly to the air brake system which
functions to provide an automatic full service penalty application of
the air brake system, and a power knock out (PC) caused by failure to
acknowledge the time out feature usually about 40 seconds. When the
Alerter time out has expired, the engineer must acknowledge by tripping
the acknowledging switch which will reset the time out feature. The
Alerter system is nullified when locomotive brake cylinder pressure of
20 psi is developed in the Independent Application and Release pipe.
This also prevents the P2A Application Valve from triggering a service
brake application and PC action.

When the engineer of Y11615 placed the locomotive independent brake
valve into full application, a design pressure was developed in the
brake cylinders, depending on the type of relay valve, which nullifies
the Alerter System. Had the engineer not placed the independent brake
into the full applied position and caused a build up of brake cylinder
pressure, when the time out feature had expired, the system would have
functioned causing not only an application of the locomotive brakes but
a PC trip which would have resulted in a Power Knock Out bringing the
movement to a stop.

According to the interview of the engineer, he made a 20 psi brake pipe
reduction with the automatic brake valve before dismounting the
locomotive. This by no means would have provided any braking power in as
much as the brake pipe was not connected to the cars and the system was
not charged.

Dynamic Brake:
General Motors Model SD-40-2 dynamic brake system is established by
placing the selector lever into Dynamic Brake Mode. This will convert
the traction motors to generators to produce voltage and amperage which
is dissipated in the form of heat through the braking grids. Excitation
for the fields of the traction motors is developed through the main
generator and is regulated by increasing main generator outputby
increasing diesel engine rpm. This increase in engine rpm is
accomplished by increasing throttle position 1 (setup) through 8
positions. The effectiveness of the dynamic brake system is generally
maximized at speeds above 40 mph. At very speeds, below 10 mph, the
dynamic brake system is not effective. When the engineer failed to
properly move the selector lever into the dynamic brake mode, the
traction motors remained in the motoring mode. By placing the throttle
handle into number 8 position in this set up, maximum locomotive
power was thus developed and diesel engine rpm would increase in a like
manner to dynamic braking. Without first observing the load meter, and
at low locomotive speeds, it may be difficult to determine immediately
if the locomotive was in braking or power mode. 

In the days following the incident, state and local officials expressed
concern about the potential for this type of event to occur in the
future. While FRA does not dismiss any potential safety concern, the
exact circumstances that combined to cause this incident are highly
unlikely to recur. It is not uncommon, today, for our inspectors to
observe an engineer bring his locomotive to a full stop, dismount his
locomotive and operate a switch. This is done safely and in accordance
with railroad safety rules and does not pose any special hazard to
employees or the general public. Railroad operating rules generally
prohibit any employee from dismounting, or mounting, moving equipment.
Engineers are required by railroad operating rules to apply a hand brake
and take other steps at the control stand to immobilize the locomotive
before dismounting.

FRA does have an initiative which will minimize the possibility of
runaway equipment resulting from unsecured equipment left unattended in
a yard. This issue falls under the umbrella of the Switching Operations
Fatality Analysis (SOFA) program, in my view, and we have taken steps to
emphasize the securement of equipment in our ongoing SOFA activities.
For example, since the incident, State of Ohio and Federal inspectors
have visited CSX and NS terminals in Ohio, and elsewhere, to review with
railroad managers policies and procedures relating to switching safety,
including securement. We have not
been able to identify any systemic problems or shortcomings in training,
supervision, or operating practices that are cause for alarm. Inspectors
do observe various local, non-systemic safety issues, and they are
addressed promptly with local managers, in accordance with our standard
policy and practice. 

I might comment further on why this incident is unlikely to recur. The
cause of the incident was multiple gross errors in judgement by the
locomotive engineer. For the incident to have occurred, each error
needed to be committed in sequence. First, the engineer was not properly
controlling the speed of his train on the lead, if he is unable to stop
for a switch improperly lined. This is covered by the railroad's
operating rules. Second, if the engineer cannot stop for a switch
improperly lined, the correct action to take is simply run through the
switch and then stop without backing up, to avoid derailing the train.
Third, an engineer should never dismount his locomotive while it is
moving, except in extremely rare emergency circumstances, such as an
imminent collision. This is also covered by the railroad's operating
rules. Fourth, the engineer should not have relied on dynamic braking at
low speed, since dynamic brakes are ineffective at speeds of less than
ten mph, except on an AC locomotive. This is well known among railroad
engineers. Fifth, the engineer seemed to believe, in error, that an
automatic brake application would improve braking power on single
locomotive with the independent brake fully applied. Sixth, the engineer
misapplied the selector handle for "power" or "dynamic brake," an error
that can only be understood if we assume the engineer acted with extreme
haste and negligence. That all of these actions were taken by an
apparently well-qualified, fully rested employee with a good service
record is simply incredible. 

It should be remarked that this incident could only have occurred during
freight car switching and not during passenger car switching. Most
freight switching is done "without air," that is without air brakes
functional in the train. This is the
industry standard, and it has been so for many decades. Passenger
switching, on the other hand, is usually performed "with air," that is
with air brakes functional in the train. In addition, it should be noted
that passenger coaches are rarely switched with passengers aboard,
because of concerns with passenger safety. In those rare circumstances
where passengers are on board during switching, the air brakes would
always be fully functional.

Finally, this incident could not have occurred to either a passenger or
freight train involved in over-the-road operations. Before any passenger
or freight train embarks on a run outside of the yard where the train is
assembled, federal regulations require that the train receive a thorough
inspection of its braking system and that the brakes be 100% operational
before the train is permitted to begin its journey.
SOURCE:  Conrail Technical Society  |  News Page  |  Send News Tip