Operating and maintenance cost reduction using probabilistic risk assessment (PRA)

"January 1992."Includes bibliographical references (pages 129-132)Final report, "Operating and maintenance cost reduction using probabilistic risk assessment (PRA)"This study quantifies the change in one measure of plant risk, the frequency of loss of long-term decay heat removal, due to changes in maintenance at the James A. Fitzpatrick (JAF) plant. Quantification is accomplished in two steps. First, the effects of maintenance are quantified in terms of changes in: a) the frequency of common cause failure of residual heat removal (RHR) pumps and b) the frequency with which operators fail to correctly restore the RHR system following maintenance. These parameters are selected as the result of an importance analysis for the plant. Second, the changes in these two parameters are propagated through a simple plant model to obtain the associated change in plant risk. Based on this study's assessment of the current maintenance program at JAF, it appears that the potential for significant risk reduction due to improved maintenance is not extremely large; an optimal program might lead to an 80% reduction. The optimal program would place a stronger emphasis on predictive maintenance, and would employ improved procedures for RHR pump maintenance. There is potential for significant risk increase (around a factor of 70) if the maintenance program is significantly degraded (e.g., if post-maintenance is deemphasized). This study shows how, at a simple level, maintenance program changes can be quantified without explicit modeling of the details of a plant's management and organizational structure. However, such modeling may be required: a) to more strongly justify the quantitative factors used in the analysis and b) to quantify the effect of other program changes not yet treated (e.g., the strengthening of program elements ensuring feedback of information to organization). In addition, failure data specific to the JAF plant are also needed to increase the confidence in the quantitative results of this study.Sponsored by New York Power Authority, White Plains, NY under contract no. S-90-0019