Case Study: BP Texas City Explosion 23 March 2005
provides a full description of the complex’s system.) The Board Operator – far removed from the physical location of the process unit undergoing startup81 – depended on the system to provide him with crucial process information, which requires a well-designed control board. On the day of the incident, however, the computerized control system display provided neither flow data in and out of the raffinate unit on the same display screen, nor a material balance calculation, hindering the Board Operator’s ability to recognize the need to send liquid raffinate to storage.
Ineffective supervisory oversight and technical assistance during unit startup.
The ISOM/AU2/NDU complex lacked effective supervisory oversight during the startup of the raffinate unit. When Day Supervisor A left the refinery for a family medical emergency, 84 no technically trained personnel with ISOM unit experience were assigned to assist and supervise the Board Operator. Two significant staffing issues arose: it was unclear who was responsible for ISOM unit supervision once Day Supervisor A left, and the one individual available to provide such supervision lacked technical knowledge of the unit. Had the second Day Supervisor on shift (Supervisor B) left his work at the ARU to assist in the raffinate startup, his presence in the control room would likely not have been helpful, as he had little technical expertise on the unit.85 The two Process Technicians (PTs) who had ISOM knowledge and experience were not assigned to assist with the startup.86
Amoco’s process safety guideline, “Supervisory Personnel – Startups and Shutdowns,” adopted by BP, states that times of unit startup and shutdown are often unpredictable and much more likely than normal operations to go awry. For this reason, the BP guideline states: “Experienced operating personnel should be assigned to each process unit as it is being started up and as it is being shut down.”87 Examples of such personnel who could provide supplementary assistance include supervisors and operating specialists. On March 23, 2005, Texas City management did not ensure that such personnel were assigned to the startup.
Insufficient staffing to handle board operator workload during the high-risk time of unit startup.
On March 23, 2005, the ISOM unit was understaffed for the task of unit startup. One board operator was in charge of monitoring and controlling the NDU, AU2, and ISOM units, which under normal conditions, would take about 10.5 hours of a 12-hour shift to run, if all units would be running at a steady state (normal). On the morning of the incident, however, the Board Operator was also responsible for managing the startup of the ISOM raffinate section. A startup is an abnormal unit condition that requires significantly more manual control of a process, as well as critical thinking and decision-making that goes beyond normal unit operation.
Lack of a human fatigue-prevention policy.
Several of the supervisors had worked twelve-hour days, seven days per week, for nearly 30 days prior to the accident. According to the Baker