Romeoville, Illinois: July 23, 1984

Shortly before 6 p.m. on July 23, 1984, a fiery explosion rocked a 157,000 barrel-per-day refinery near Chicago. Seventeen people died in the disaster. Eleven of the dead belonged to the plant fire brigade, the largest number of American industrial firefighters to die in a single incident.

History has a terrible way of repeating itself, particularly when it comes to industrial fire fighting. On this, the 24th anniversary of one of the worst refinery disasters on record, the time seems ripe to reexamine the events that led to this terrible event.

At about 5:30 p.m., operators at the Fluid Catalytic Cracking Complex were notified by telephone that a vapor release from their unit had been observed. The operators identified a vapor cloud about 15 to 20 feet long and about 12 feet off the ground near 12D701, a 34-ton vessel measuring 8½-feet in diameter and rising more than 53 feet above the ground. The vapor cloud originated from a six- to nine-inch long horizontal crack in 12D701. Because the leak was behind a fixed ladder on the side of the vessel, operators were unable to reach the valve to isolate the unit.

Known as the amine absorber, 12D701 received liquefied propane and butane gas rich in hydrogen sulfide (H2S) at a port near the bottom of the vessel. Liquefied propane and butane gas would rise in the vessel through a series of perforated trays, exiting at the top relatively lean of H2S. An amine solution consisting of monoethanolamine in water would enter the vessel at the top and, flowing downward counter to the flow of the liquefied petroleum product through the perforated trays, would strip the H2S from the liquefied propane and butane gases.

The amine solution would exit the vessel at the bottom, having in the process become rich in H2S. The liquefied propane and butane gases would flow to the next pressure vessel in the process for additional treatment involving the removal of other sulfur compounds.

After sounding an alarm, the operators further inspected the vessel but found no other leaks. Orders were given to connect steam hoses and direct steam at the leak. A second ladder arrived but proved too short to allow operators to activate elevated valves on the side of 12D701 that would have redirected the flow of product. The only other option was to shut down the entire unit, a procedure that would take 30 to 45 minutes.

According to Occupational Safety and Health Administration documents, a refinery fire truck soon arrived. Within 20 minutes of the warning alarm, the first of several explosions occurred. For several seconds, a wall of fire engulfed the leaking vessel. The upper 85 percent of 12D701 was hurled 2,300 feet from the point of origin, spewing its flaming contents as it went. The shift fire crew was caught in the fireball.

Electrical power was down throughout the refinery, limiting in-plant radio communications to hand-held radios. Radio traffic was confused at best. Somewhere a worker trapped by flames was pleading for help. No one could reach him, and he became one of the fatalities.

A second explosion shook the refinery. Fire had spread through an unsaturated gas unit and an alkalization unit with multiple acid settlers. The reactor of the alkalization unit BLEVEed (boiling liquid expanding vapor explosion), spreading further destruction. Responders now had to contend with a compressor fire in a hydrotreating unit, plus fire in a warehouse, alkalization unit and an on-site gas unit. Part of a sulfide stress cracking unit was also on fire. Split lines off two slop oil tanks were burning, as was a crude unit and a coker unit.

With the aid of 35 other departments, the fire was officially extinguished within 17 hours of ignition. Then the investigation into what triggered the disaster began.

Corrosion Defects

Fourteen years before the fire and explosion, a new pressure vessel designated 12D701 was placed in service at the refinery. Fabricated in accordance with the American Society of Mechanical Engineers Code, the vessel was built using 1-inch-thick A516-70 carbon steel. It consisted of a top and bottom head and seven cylindrical sections referred to as "courses" all joined by circumferential welds.

Vessel 12D701 was designed to operate at a maximum pressure of 230 psi at a maximum temperature of 140 degrees Fahrenheit (60 degrees Celsius). That pressure was measured at the top of the tank, meaning that the lower courses exerted greater pressure on the walls due to the weight of the above product.

Starting in 1972, a biannual turnaround inspection of the vessel was required. According to OSHA court findings, that first inspection revealed corrosion defects in the walls of 12D701 resulting from the action of hydrogen on the carbon steel. Hydrogen in its atomic form passes through the steel and dissipates harmlessly into the atmosphere. However, if hydrogen atoms encounter flaws in the manufacturing of the steel such as the inclusion of foreign material, it accumulates and converts to molecular hydrogen.

In the process, tremendous heat and pressure are exerted causing the inclusion to become a lamination within the steel. Where these laminations occur near the surface, they appear as blisters.

"Should laminations, which generally run parallel to the surface of the steel, be joined by cracks which tend to run at right angle to the surface, hydrogen step cracking is said to have occurred," OSHA documents stated. "Also, pitting can occur on the interior surface of the vessel as a result of hydrogen attack." Such pitting and blistering were observed to have occurred on the interior of 12D701.

Management reacted to the corrosion deterioration by ordering that the vessel's second course be replaced during the 1974 turnaround. Using the same grade steel, the replacement course consisted of three pieces requiring three vertical and two circumferential welds. Dissatisfied with the first contractor hired to do the welding, the refinery brought in a second contractor to redo the work. In court, OSHA would later express concern that hard welds producing martensite had the potential to crack if exposed to atomic hydrogen.

More deterioration was observed in the remaining courses of the vessel during the 1974 inspection.

"Visual inspection of the balance of 12D701 during the 1974 turnaround disclosed continuing corrosion and attack by hydrogen in the form of numerous small hydrogen blisters on the interior of the first course of the vessel," OSHA court documents stated.

A turnaround inspection in 1976 revealed numerous blisters ranging in size from ¼-inch to 1-inch in diameter. The lamination of the blisters varied from .16-inch to .38-inch. Courses 3, 4, 5 and 6 were now observed to have "dirty metal or lamination." A monel metal liner was welded to the bottom head of the vessel and the interior surface of the first course. The vessel was again returned to service.

Beside visual inspections, the thickness of metal throughout the vessel was monitored frequently using straight beam ultrasonic examination. However, OSHA authorities maintained that this technique was prone to error when encountering lamination.

Finally, a turnaround inspection in 1984, only three months before the explosion, found that the vessel shell was "above nominal thickness." A visual inspection found six hydrogen blisters, each a half inch in diameter, in the feed tray area, all located within one foot of the circumferential weld joining course one and two at the bottom of the vessel.

Again, vessel 12D701 was returned to service. The court would later find that the refinery had failed to adequately inspect for cracks near field welds or in the walls of pressure vessels exposed to a hydrogen environment. The refinery failed to use methods such as magnetic particle testing, liquid dye penetrant testing or angle beam ultrasonic testing which might have revealed cracks invisible to the naked eye.

Training Issues

OSHA documents cite that in 1979 the refinery fire chief wrote a letter to management calling attention to difficulties in getting shift fire crew personnel to attend fire training sessions. He pointed out that the problem had existed for many years and stated "… the lack of attendance at trainings and response to off-shift emergencies cannot be ignored any longer.

"The potential of the loss of life and property without an efficient off-shift fire and rescue squad is too great," he wrote.

Prior to the 1984 explosion, the refinery's written emergency procedures required all operating personnel to extinguish or contain minor fires located in their work areas as part of their regular duties. Fires beyond the capabilities of the operating personnel were to be handled by the "day fire crew" – volunteers trained to be proficient in fire fighting. Available between 7:30 a.m. and 4 p.m. on weekdays, the day fire crew trained on a monthly basis.

Members of the "shift fire crew" – refinery employees trained to perform fire fighting duties as part of their mandatory job description – were to respond when the day fire crew was not present. This crew trained on a bi-monthly basis. The shift fire crew was to call in the day fire crew from their homes should a fire develop beyond their capabilities. As of 1979 all ten shift fire crew chiefs had attended off-site fire training at either Texas A&M University or the Western Oil and Gas Association training facility in Reno, NV.

The varying degrees of training for fire-fighters expected to deal with the same level of risk would later become an issue in court.

In a December 1981 letter, the fire chief expressed his disappoint-ment in the attendance of shift fire crew personnel at training sessions throughout that year. The shift fire crew chiefs had only a 66 percent attendance, he wrote.

"We have determined that if each classification in the Shift Fire Crew achieves an 80 percent attendance in each of the six annual classes, they should be able to maintain the basic skills required for the job … We believe that attendance below this minimum greatly increases the risk of property damage and personal injury to the members themselves."

The chief continued to let his displeasure with the training situation be known. In March 1983 he wrote some shift fire chiefs had a zero percent attendance at training classes that month.

"The problem is now to the point that not only I seriously question the Shift Fire Chief's ability, but many members of the Shift Crew … have also begun to ask the same question … Such a lack of confidence would prove disastrous to any progress we have made with the Shift Fire Crew over the past few years.

"Continuing neglect of training by the shift fire crew chiefs was also noted by the chief in June 1983. In January 1984, he reported in another letter that while attendance of the shift fire chiefs at training sessions had risen, it was still less than 50 percent. Shift crew member attendance had improved and was observed to be "marginally acceptable."

OSHA court documents noted that the chief had no input in the selection of shift crew members or chiefs and on at least two occasions suggested that shift fire crew members should be replaced in their job classifications if they failed to have satisfactory attendance at training sessions or failed a written competency test.

That the refinery gave production priority over safety fire training is demonstrated in the closure of the refinery's fire training field in 1983, the OSHA documents stated. Without providing an alternative fire training site, refinery management opted to use the existing site for a new production facility.

"From May of 1983 until around the first of July 1984, the respondent's Fire Department was without a live training field where it could provide hands-on training in putting out actual fires to its fire fighting personnel," OSHA stated. The chief "considered live fire fighting drills to be a necessary part of the training."

Within a month of having a training facility restored to him, the refinery fire chief faced the worst live-fire training scenario possible. But it was not a test.

Conclusion

OSHA issued four citations to the refinery carrying a total fine of $31,000. The refinery was cited for failure to have an effective maintenance and inspection program for pressure tanks, failure to provide protective equipment for workers, failure to adequately train a fire brigade and lack of effective emergency procedures.

IFW has published this article in the hope that industrial management worldwide will appreciate the consequences of not taking proactive actions when signals and signs are apparent. Unfortunately, while the deaths of industrial firefighters are rare, we continue to have incidents that frequently kill workers.

I routinely perform training in plants and evaluate emergency response organizations. My greatest concern is that in many cases we are training the wrong people. We are training the better fire brigades while the poorly trained and led brigades are not attending class.

Recently, I conducted training in a large city for a combined group of municipal and industrial firefighters. I was amazed to learn that the city thought the plant firefighters would handle any major fires while the plant thought the city would. What kind of pre-planning is this?

A team of industrial firefighting experts, myself included, testified in federal court regarding this Illinois refinery disaster. The plant management barely escaped prosecution for negligent homicide. Today, given the political atmosphere and the growing emphasis on emergency response training, it might be very difficult to keep someone operating an industrial facility in this manner out of jail following an event of this magnitude.

Editor's note: This after-action review appeared in "Disasters Man-Made" by David White and Anton Riecher that was published in 2011.