Condition monitoring and testing for operability of check valves and pumps

A detailed analysis of historical failure data available through the Institute of Nuclear Power Operations` Nuclear Plant Reliability Data System (NPRDS) has been conducted for both check valves and pumps. This analysis, which originated as a part of the Nuclear Regulatory Commission`s (NRC) effort to evaluate the effects of age and wear on nuclear systems components, involved the manual review and characterization of several thousand component failure records according to parameters inherent in the NPRDS database and supplemented by those defined by the analyst for each component type. For example, failure information relative to component size, age, system of service, and NSSS vendor was readily available from the NPRDS database and could be compared relatively easily. Determination of parameters such as extent of degradation, affected area, and detection method, however, had to be determined based on manual review and characterization of individual failure narratives. This paper discusses some of the results of the analyses of historical check valve failure data from 1984 through 1992 and pump failure data from 1990 through 1993. A comparison of the findings of the analyses is made, and emphasis is placed on evaluation of the effectiveness of certain failure detection methods for each component type. Generally speaking, while it was observed that check valve degradation or failure was likely to be detected by code or regulatory required testing, it was discovered that pump degradation or failure was most likely to be discovered by voluntarily implemented plant programs. Failure rates were found to be strongly influenced by the valve or pump application. The type of plant (BWR or PWR) was the overall failure rate and the method of failure detection.