All it took was three minutes. The Williams Olefins chemical plant in Geismar, Louisiana, which was routinely executing everyday operations, was instantly transformed into a scene of destruction, panic and mayhem. A catastrophic equipment rupture led to an explosion and fire killing two workers and injuring 167. In retrospect, it was a tragedy that could have been prevented.
The explosion, which occurred on June 13, 2013, took place when heat levels began to rise in an offline “reboiler”. This increase in heat created an intense amount of pressure due to liquid thermal expansion. The reboiler shell catastrophically ruptured, triggering a boiling liquid expanding vapor explosion (BLEVE) and fire.
As is the case involving most safety-related disasters across the globe, it wasn’t just one isolated error that led to the explosion. Rather, a series of lapses that accumulated over the years, combined to trigger a series of disastrous chain reactions that could no longer be controlled and contained. A post mortem investigation revealed that factors such as lack of overpressure protection, poor change management, inadequate hazard analysis procedures, and a weak process safety culture all contributed to the blast.
Post-event safety analysis
In order to peel back the layers of the safety breach onion, we need to analyze the history of the events that transpired.
During a morning meeting with operations and maintenance personnel the day of the disaster, the plant manager noted that the quench water flow through the operating propylene fractionator reboiler (Reboiler A) had dropped over the last 24 hours. The group then analyzed plant data and observed that the water circulation rate appeared deficient. An operations supervisor investigated the operating reboiler (Reboiler A) and speculated that fouling in the reboiler could be the cause of the problem. He recommended that the propylene fractionator function be rerouted through Reboiler B to correct the quench water flow issue.
The operations supervisor attempted to meet with the operations manager to discuss switching the reboilers so that they could initiate the corrective action. The operations manager was not available, so the operations supervisor returned to the field and continued evaluating the quench water system.
Both Reboiler A and Reboiler B were designed to supply heat to the propylene fractionator — a distillation column that separates propylene and propane. The process fluid on the shell-side of these reboilers is heated by hot “quench water,” flowing through the tubes. Reboiler B had been offline for 16 months while Reboiler A was in operation, but was clean and available for use when Reboiler A needed to shut down because of fouling.
At 8:33 a.m., the operations supervisor opened the quench water valves on Reboiler B. Three minutes later, Reboiler B exploded. Propane and propylene process fluid erupted from the ruptured reboiler and from the propylene fractionator due to failed piping. The process vapor ignited, creating a massive fireball. The force of the explosion launched a portion of the propylene fractionator reboiler piping into a rack located 30 feet overhead.
The fire lasted 3.5 hours, and Williams Olefins reported releasing over 30,000 pounds of flammable hydrocarbons in the atmosphere during the incident. The plant remained shut down for the next 18 months.
In retrospect, the countdown to the disaster started 12 years earlier in 2001, when Williams Olefins installed valves on the shell-side and tube-side reboiler piping to allow for continuous operation with only one reboiler operating at a time. The other reboiler would be offline but ready for operation, isolated from the process by the new valves. This configuration allowed for cleaning of a fouled reboiler while the propylene fractionator continued to operate.
However, these valves inadvertently introduced a new process hazard. If the new valves were not in the proper position (open or closed) during each phase of operation, the reboiler could be isolated from its protective pressure relief valve located on top of the propylene fractionator. If that happened, control over the pressure levels would be lost.
As it turned out, Reboiler B’s tube-side hot quench water valves were open, but the shell-side process valves were closed, which isolated the shell of Reboiler B from its protective pressure relief valve on the top of the propylene fractionator.
When the Reboiler B hot quench water valves were opened, the liquid propane within the standby Reboiler B shell began to heat up. This caused the liquid propane to increase in volume due to liquid thermal expansion, filling any remaining space within the shell. When the liquid could no longer expand due to confinement within the blocked-in Reboiler B shell, the pressure rapidly increased until the internal pressure exceeded the shell’s mechanical pressure limit and the reboiler shell failed.
As a result of lessons learned and in order to avoid a repeat of the disaster, Williams Olefins redesigned the propylene fractionator reboilers to include a pressure relief valve on the shell side of each reboiler.
In order to find out how some of the various mistakes that occurred over time should have been addressed and prevented, click here.