WHAT WE DIDN’T LEARN FROM THE WAVERLY TRAGEDY
On February 24, 1978, a railroad tank car loaded with liquefied petroleum gas (IPG) ruptured following a derailment in the small western Tennessee town of Waverly. In the massive fireball that followed the tank failure, 16 people were killed, including the town’s fire chief, police chief, and a Tennessee Emergency Management Agency hazardous-materials specialist.
What we didn’t learn from this tragedy is almost as sad as the injuries and loss of life that occurred. Public safety agencies have failed to study this hazardous-materials catastrophe, and thus many of the factors contributing to the magnitude of the accident could return to haunt us in any incident with similar circumstances. This article studies what happened, without placing blame, since the “state-ofthe-art” mistakes that happened in Waverly could occur anywhere today.
THE INCIDENT
A 92-car Louisville & Nashville (L&N) railroad freight train was west-bound for Memphis out of Nashville on the night of February 22, when a wheel on a boxcar cracked about six miles east of Waverly. The cracked wheel caused the two-axle truck it was part of to derail and be dragged along by the moving train. This type of failure is not uncommon and hardly ever can be detected by the train crew, since it changes nothing in the performance of the train or locomotives.
A switch for a siding in the middle of Waverly caught the derailed truck and caused the major derailment of the train, dumping 24 cars off the track at about 10:38 p.m. Wednesday night.
The weather in western Tennessee had been below freezing a few days prior to the derailment, and a light snow was falling at the time it occurred. A number of boxcars came to rest on top of two Department of Transportation (DOT) specification 112 tank cars, each loaded with about 28,000 gallons of LPG. The tank cars had been loaded in Nashville, and the temperature of the contents was estimated to be near the average ambient air temperature for the preceding several days, about 40°F.
L&N railroad officials responded immediately to the derailment, including Resident Vice President Phillip Hooper, a retired colonel. Colonel Hooper, a seasoned railroad executive, assumed command of railroad operations at the scene, which were designed to clear the wreckage and get the track reopened.
Although police and fire officials responded to the initial derailment, which was only a few blocks from city hall, there was no major concern until the following day, when the seriously damaged LPG tank cars were uncovered in the wreckage. At 8:06 a m. on Thursday, about 30 residents in a onequarter-mile area were evacuated for about eight hours while the situation was assessed. Rail and public safety officials examined the two LPG tank cars and, detecting no leaks, relaxed their evacuations and continued on with wreckage-clearing operations.
A practice still in use today was used to move the two damaged, loaded LPG cars from their position, which was directly astride the destroyed track. A heavy crane secured a steel cable sling around one end of the car, lifted it several feet off the ground, and swung it to the side, pivoting the car about the opposite end on the ground. In this fashion, both cars were positioned alongside the right-of-way, and the track was rebuilt.
Arrangements were made on Thursday for Liquid Transporters Inc. of Louisville to send several empty LPG highway cargo tank trucks to the scene the following day so that the product could be transferred from the damaged tank cars. These tank cars do not contain pumps or other equipment necessary to off-load themselves.
The wreckage was cleared Thursday afternoon. By Thursday night the track was rebuilt and a freight train passed through town on the new track. No particular hazard was attached to the still-loaded LPG tank cars lying alongside the track.
A SHIFT IN CONDITIONS
Friday morning saw an unseasonable change in Tennessee weather, with clear skies and a significant warming trend that sent temperatures climbing to the mid-50s by that afternoon. In preparation for the off-loading of the LPG, neighboring buildings were partially evacuated; the evacuation process was only marginally successful. Customers were drinking in a basement bar less than 50 feet from the scene, and two employees were unloading a truck of building materials at a lumber yard 100 feet away.
The evacuation order also did not keep a retired firefighter and a teenaged neighborhood resident from walking into the area to watch. Around noon, preparations got underway to off-load the propane, and Fire Chief Wilbur York had one of his engines stationed at the scene. Police Chief Guy Harnett, a longtime friend of York’s, was talking to him when a rapid series of events unfolded.
According to the National Transportation Safety Board (NISH) report on the incident, one of the LPG tank cars had sustained a small crack during the derailment, probably caused by the entire car sliding over the top of a derailed wheel-and-axle assembly. This contact caused a significant “wheel burn,” which dented the pressure vessel and caused “cold-working” of the steel, creating the crack. (Railroaders are more familiar with the form of wheel burn caused by the car dropping straight down and creating a deep semicircular-shaped dent and gouge. This crease was long and relatively shallow.) The crack apparently was propagated further by the movement involving booming it about one end by the crane during the derailment-clearing operation.
The unseasonably warm weather heated the contents of the tank car, causing a pressure increase. The higher internal pressure caused the crack to grow and catastrophically propagate itself, first along the long axis of the car for a number of feet. As the crack velocity increased, it began to run perpendicular to the long axis of the car, causing the tank to come apart, dumping its LPG contents.
This massive release of the LPG formed a rapidly growing cloud of flammable gas. The gas reached an ignition source at the derailment scene and immediately was transformed into a huge fireball hundreds of feet in diameter. Five people on the scene died instantly.
Fire Chief York and Police Chief Harnett were horribly burned and succumbed to their injuries in a Nashville hospital several days later. One firefighter suffered extensive burns and subsequently would undergo dozens of operations over a period of several years.
Tom Doss, director of the Humphries County Ambulance Service, had visited the scene just minutes prior to the tank failure. He tried to talk the police and fire chiefs into completely evacuating the area. When they failed to share his concern, he quietly returned to his ambulance headquarters two miles away, called in every offduty ambulance driver and attendant, and began to stock his ambulances for the tragedy he felt positive was coming.
When the tank failed. Doss got on his emergency radio and called for every available ambulance in western Tennessee to respond, since he knew all too well the magnitude of burn injuries that would result. Due to the efforts of his crews and a smoothly functioning disaster plan that immediately was implemented at Waverly’s Nautilus Hospital, the evacuation and treatment of the injured were models of efficiency involving some 49 ambulances and seven medical evacuation helicopters, some from as far away as Fort Campbell, Kentucky.
Dozens of pieces of fire apparatus responded from towns as far away as Nashville, and by that evening the numbers of fires burning for blocks around the scene had been brought under control. Tennessee Governor Ray Blanton, as well as General Carl Wallace, head of the Tennessee National Guard, which included the Tennessee Emergency Management Agency, quickly responded to the scene.
Relief operations were extensive and effective. Amid quotes from public officials talking about new increased safety rules being necessary and calls for an investigation, things slowly returned to normal.
ANALYSIS
Waverly sometimes is referred to as a “cold BLEVE,” but that designation is misleading, since the typical connotation of a boiling-liquid, expandingvapor explosion involves tank failure through heating of the external shell from fire, which precipitates a weakening of the steel due to flame impingement. It is incorrect, therefore, to describe the Waverly failure with the term traditionally used to describe an entirely different mechanism for pressure vessel failure.
The NTSB’s exhaustive investigation utilizing metallurgical analysis of the failed tank is fairly conclusive. The Waverly car failed due to a crack caused by “cold-working” of the steel during the derailment. There was no flame impingement on the shell, and the LPG inside the tank certainly was not boiling at the time of failure. The huge fireball, similar in appearance to a classic BLEVE, is simply characteristic of the ignition of a massive flammable vapor cloud.
In the years following the tragedy, stories occasionally surfaced regarding alleged mishandling of the incident, which supposedly contributed to its causes. One of these erroneous stories stated that the fire department had employed master stream appliances and placed large amounts of water on the damaged cars in the hours prior to the tank failure. Interview’s with Assistant Chief F.X. “Dutch” Geisenhoffer and several surviving firefighters who were at the scene the entire time revealed this did not occur.
Master stream appliances are evident in photographs of the aftermath of the disaster and were used extensively to play water on the remaining loaded LPG tank car the next day, when it was feared that it too would fail. Judging from photographs taken prior to the tank failure, the tank that survived actually looked like it was much more heavily damaged than the car that failed. The intact car suffered extensive damage and serious denting on one end, which was not evidenced at all on the car that failed.
RECOGNIZING THE PROBLEM
Perhaps the greatest tragic legacy of Waverly involves our failure to study the incident and understand it so that we can learn from the accident and improve our capabilities. This lesson is best underscored by a quote from a fire official, who stated a number of years later what should be done with loaded rail LPG tank cars in similar circumstances: “Put all the water you can get on ’em. Put your deluge guns in position…”
The firefighting water supply for the town of Waverly is drawn from the Tennessee River, which flows about 12 miles to the west. At the time of the derailment and for the entire next day, river temperatures were estimated to be a number of degrees warmer than the prevailing ambient air temperature. If this water had been played on the car that cracked following its repositioning, some who have analyzed the incident believe it simply would have caused the car to fail 24 hours sooner than it did —and for the exact same reason, increased internal pressure due to heating of the contents caused by an external source, in this case, water, instead of the unseasonably warm Friday temperature.
The exact contribution made to the tank failure by the initial derailmentsustained crack, the propagation due to the crane movement, and the final failure from internal pressure never will be determined. What is clear, however, is that emergency responders must recognize when similar circumstances could lead to the same tragic sequence of events. Therefore, one primary rule that can be stated from the Waverly experience is based on both what happened and an NTSB conclusion from an entirely separate study we will address: The most knowledgeable expertise should be brought to the scene to analyze any significant damage to LPG pressure vessels in derailments.
The Association of American Railroads now teaches at its Pueblo Transportation Test Center that when a pressure car has received certain sizes of dents or the type of cold-working from a “wheel burn” as had happened in Waverly, the possibility of imminent failure should be anticipated. Unfortunately, the number of people who have attended courses at this facility or who have ever actually seen a critical “wheel burn” nowhere near approaches the number of times this type of expertise is critically needed at derailment scenes.
The other NTSB study, Special Investigation Report: Tank Car Structural Integrity After Derailment (NTSBSIR-80-1, Oct. 16. 1980), noted above involved some tests with a derailment-damaged tank car that was filled with water and pressurized. Experts were asked to determine if it would fail in the test, and all said it would not. It, in fact, failed at an extremely low pressure. Among the major conclusions of this enlightening study are the following:
- No accurate method for estimating the residual strength of damaged railroad tank cars at derailment sites currently exists.
- The method now used for determining the structural integrity of damaged tank cars during wreck-clearing operations has been empirically derived and can lead to underestimation of the dangers present.
- The existing body of knowledge on damaged tank cars does not produce accurate predictions on the behavior of such cars in a postderailment environment.
- Small surface cracks in the heataffected zone of welds can develop into ruptures if the damaged tank shell contour is changed by increases in internal pressure.
These disquieting test results logically lead to several thoughts. It is hoped that future studies in the hazardous-materials transportation accident field will include further consideration of the safest methods for moving and handling damaged loaded pressure vessels. Obviously, we also need a better understanding of the severity of varying types of damage to pressure vessels. Except for the 1980 NTSB study noted above, there has been no significant research involving the assessment of the structural integrity of damaged pressure vessels.
LESSONS LEARNED
The several obvious lessons learned from the Waverly tragedy include the following:
- It is better to be overly cautious in ordering and maintaining evacuations than to hope that “things will work out.”
- Only a bare minimum number of workers and public safety officials should remain inside the hazardous perimeter.
- Seriously damaged loaded pressure vessels should be pumped off prior to movement.
Two more “lessons learned” that can be logically derived from the Waverly experience are certain to generate controversy in the railroad industry due to the commonality of the still-utilized practice of booming loaded cars about one end. However, I responded to a derailment several years ago involving another seriously damaged LPG tank car. Senior rail officials and a nationally recognized authority were not able to guarantee that this practice would not lead to results as tragic as those at Waverly, and the car was not moved but pumped off where it lay.
- If damaged loaded pressure vessels must be moved, use multiple cranes or lifting points at both ends to equalize the stresses involved.
- If water must be applied to damaged pressure vessels, consider its temperature and any effect it may have on internal pressures or thermal shock to the car shell itself.
It has been said that those who do not learn from history are doomed to repeat it. By studying the tragic events that led to 16 deaths and more than 200 injuries in 1978 in Waverly. Tennessee. we hope to avoid repeating past mistakes.
