Although the processes to be employed for disposal of chemical munitions at the Pueblo and Blue Grass Chemical Agent Destruction Pilot Plants (PCAPP and BGCAPP) are fundamentally different from those used at earlier chemical agent disposal facilities and vary somewhat among themselves, many similarities exist within and among the processes employed at all of the disposal facility sites. Prominent examples of similarities shared by the incineration sites and PCAPP and BGCAPP include the use of rocket shear machines (at sites where M55 rockets have been stored), thermal decontamination of metal parts, and methods used for waste treatment. The committee believed that evaluating documentation from those other sites concerning process incidents, any process safety metrics used, and process hazards could prove useful and would offer guidance on what process safety metrics might be useful for PCAPP and BGCAPP. Although none of the other chemical agent disposal facilities employed formal process safety metrics, the committee believed that an analysis of chemical events that occurred at those sites could provide insights into those process steps that are most subject to failure and might identify opportunities where the use of leading and lagging metrics could help to prevent failures.¹ Thus, there are aspects of the operations of even the baseline incineration facilities that could be useful for deriving process safety metrics for use at PCAPP and BGCAPP.

The National Research Council Committee on Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities (the chemical events committee) reviewed all of the chemical events that had occurred since the commencement of destruction operations through the end of 2001, totaling 81 process and non-process-related incidents at that time. The chemical events committee issued its report, Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities, in 2002. This committee reviewed that report to identify which of the events could be classified as process incidents.² Interestingly, the chemical events committee found that the causal factors underlying those events were not process related but were independent of any specific process. The chemical events committee’s analysis of causal factors is summarized in Table 3-1.

The following definitions extracted from the report on chemical events apply to the terms used in the tables in this chapter:

• Standard operating procedure (SOP) deficiencies refer to nonexistent SOPs, inadequate SOPs, or SOPs being circumvented or ignored as a routine operating practice.

• Equipment malfunction refers to the failure of equipment to function as designed but does not include design deficiencies. Such failures range from the simple tearing of waste bags to break­downs of critical instrumentation such as flow-meters and sensors.

• Design deficiency applies to equipment or facilities found to perform their operating functions inadequately as a result of poor design.

• Unexplained human error refers to human actions that were wrong for no reason recorded in the investigation reports or for which there is no apparent explanation. One example is when an operator assembled a piece of equipment incorrectly.

• Mindset refers to the mental attitude people have about the process of disposal and the state of the system during processing. One example is when a person assumes an agent alarm is false because of a historical pattern of frequent false alarms.

• Improper technique refers to a manner of performing tasks that causes either a hazard or a malfunction. An example is using equipment for purposes other than those dictated by design.

• Failures of communication refer to failure to communicate essential information, failure to heed communicated information, and inadequate communications systems.

This committee requested an update on process-related chemical events from all operating sites and the two sites that completed destruction since the end of 2001. In all, 147 events were reported to the committee,³ of which 26 were reviewed by the chemical events committee in 2002. This committee evaluated the remaining 121 incidents for frequency of event type, process activity involved (e.g., maintenance, waste handling, weapons transfer, and agent transfer), consequence, and causal factors. The frequencies of incident types, activities, and causal factors mirror those that were noted in the 2002 Chemical Events Report. A summary of the causal factors for these 121 events is presented in Table 3-2.

“SOP deficiencies” was the most frequent causal factor, followed by “equipment malfunction” and “human error.” Significantly, almost all of the events were noted to have had multiple causal factors, as is evidenced by the fact that 215 causal factors were identified for 147 events (also see notes to tables).

The activities that had the most incidents and events were maintenance and waste handling. Thirty-one events happened during waste handling, including hydrolysate transfers and spills. Twenty-two events

happened during maintenance activities. Only two incidents happened during munitions transfer and two during agent transfer; however, not all agent transfer incidents were tabulated. Twenty-five incidents occurred in the rocket shear machine, but the causes of 21 of these were not, or could not be, assigned.

In summary, it appears that the frequency, types, and causal factors of process safety events in chemical agent disposal facilities could not be correlated with the type of facility (neutralization or incineration), type of chemical weapon (blister agent or nerve agent), or how

the agent is stored (in assembled munitions or bulk storage). Consequently, PCAPP and BGCAPP can reasonably expect to experience the same types of events with similar causal factors. Because of the unique nature of the processes to be employed at PCAPP and BGCAPP, however, and the extensive use of first-of-a-kind equipment, it may be reasonable to expect more events in the early part of operations and, based on the data in Tables 3-1 and 3-2, a shift in the relative proportion of causal factors. For example, design deficiencies might be more prevalent in new facilities than in older or second-generation facilities.

Finding 3-1. The causal factors involved in past events at chemical agent disposal facilities are not process specific. Consequently, the Pueblo Chemical Agent Destruction Pilot Plant and the Blue Grass Chemical Agent Destruction Pilot Plant can reasonably be expected to experience the same types of events having causal factors similar to those experienced at the Chemical Materials Agency sites. Also, there may be an increase in the frequency of events and a shift in the relative frequency of causal factors.

Causal factors for process safety events are discussed further in Chapter 4.

NRC (National Research Council). 2002. Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities. Washington, D.C.: The National Academies Press.