(146c) Improving Checklisting with Predictive Failure Analysis

Check listing is a very common way of making sure that we have thought about everything of concern to us, whether it be going to the grocery store, getting ready for a presentation, making sure that all our instrumentation is working in a chemical plant, or ensuring that our evacuation procedures are up to date with current conditions. All of us have these. In the chemical industry, we primarily use HAZOP as our mechanism for doing this. It consists of asking basic questions about our process as designed and operated:

1. What if the (flow, temperature, pressure) is not what it is supposed to be? Higher, lower, the same?

2. What if more or less of a substance is present?

3. What if none of the substance or flow is present?

4. What if a different or varying composition of a substance is present?

And there are many more. The nature of these questions is ?what?? They basically do not question the basic design proposal or operation unless a serious and obvious flaw is identified. Despite these extensive and time consuming reviews, the chemical industry continues to have accidents, some of which cause major property loss as well as injuries and loss of life. After one of these incidents, it seems we are always able to identify what went wrong. On some occasions, it is a failure to follow a known procedure and not that we did not understand what might go wrong. However, in some cases, this isn't the case and post incident investigations turn up errors in design or operation that were overlooked in these checklist processes. We then add this learning to the next generation of plants or change the procedures for the existing plant.

How can we improve our ability to ensure that accidents don't happen? By becoming deliberate saboteurs and thinking about HOW to cause a problem, and not thinking about WHAT can cause the problem. In a traditional HAZOP review, much caffeine is consumed going through every process variable and asking questions like, ?higher, lower, none, combined with?. This is a tedious process with little energy. If however we ask the participants to cause the process to fail, produce hazardous releases, injure people, etc. the mental process changes in a very sadistic, but productive way. When asked to figure out how to make a process release hazardous materials ALL the time and cause a major human and environmental catastrophe, a group will not only find a lot more possibilities, but will have a lot of fun doing it. This presentation will review several major incidents from this perspective and demonstrate how this technique could have been used proactively to improve the process design and minimize the potential for the disasters that happened.