This short article deals with the research needs of reliability and failure prevention in the field of anti-earthquake design of industrial facilities like nuclear power stations, oil refineries, and petro-chemical industries. The discussion will start from the current design flow and will look for the reliability concept in the flow. We usually put design margins everywhere in the design flow. And this design margin can be directly converted into the reliability failure, if we have appropriate data. The author examines the design parameters and tries to find what type of data is needed for this purpose. Next step is data collection of modes of failure which were observed in past destructive earthquakes. These data lead us to design checking points. Unless we have exact knowledge of how equipment, piping, vessels, and other installation failed, we can not design them nor evaluate failure probability of these items. The failure prevention procedure of such plants under destructive earthquake conditions has not been well established. In the case of earthquake engineering, this subject should involve failure prevention, failure detection during earthquake, and post-earthquake inspection after the event. One of the roles of reliability analysis in this field is evaluation of the overall probability of occurrence of plant disasters, like the Three Mile Island accident, under a certain level of destructive earthquakes. The author quickly reviews what data are needed for this purpose and what data are currently unavailable. The example will be taken fron nuclear power station study.

This content is only available via PDF.
You do not currently have access to this content.