Performance Evaluation of Japanese Nuclear Power Plant Based on Open Data and Information

Abstract

The operation term of Japanese nuclear power plants is less than 13 months. Moreover, the refuel outage period is longer than the other countries. It is likely that excessive preventive maintenances result in early “infant mortality” failures (early failures). However, the statistical evaluation report about failure type of Japanese nuclear power plants was not found. Therefore, Evaluation of plant engineering performance was tried on open data/information with a statistical method. The Weibull distribution/analysis, which needs not only failure data but also censored data like the preventive maintenance, was applied to plant performance evaluation. As open data/information, an annual report; “Operational Status of Nuclear Facilities in Japan” is very popular for the operation database of Japanese nuclear power plants. However, maintenance date/information of failed equipment which caused plant shutdowns were not reported in this annual report. Therefore, every equipment was assumed to be maintained during every shutdown, because this assumption generally makes conservative results of failure rate and Mean Time Before Failure (MTBF) for discussion about early failures. Data till March 2007 were collected from these annual reports. As a result of plant performance evaluation, failure type was early “infant mortality” failure type. Excessive maintenances probably resulted in early failure type. Moreover, influence of early failures and effects of plant design modifications were evaluated. Almost 30% of failures occurred within one month after restart, and failure type excluded these failures was chance (random) type. Early failures within one month may be closely evaluated to improve maintenances. Design modification effect of Japanese nuclear power plants was confirmed.