ABSTRACT
Use of generic data in risk analysis is always dangerous. There is the ever present problem that in the actual plant, components may perform more poorly than a general industrial average. The result may be an overly optimistic calculation. On the other hand, it is in practice rare to obtain a significantly pessimistic calculation. For combinations of failures, for component such as pump and valves if average figures are used, since it is difficult to operate components so that their failure rates differ significantly from average. Specially susceptible components are found quickly, and replaced.
During plant design, however, generic failure rate data is all that is available. Some calculations will need to be made even before equipment suppliers are chosen. For this reason, a set of typical data is given here. The data have been obtained by taking fairly representative components which have been studied by the author or his colleagues, and recording their failure modes and failure rates. The results have been cross checked with available data base figures. Note that the reader should contact the author for the most recent data, since estimates are under constant revision.
