Applicability of Operational Research Techniques in CANDU Nuclear Plant Maintenance

As previously reported at ICONE 6 in New Orleans, 1996, and ICONE 9 in Niece, 2001, the use of various maintenance optimization techniques at Bruce has lead to cost effective preventive maintenance applications for complex systems. Innovative practices included greatly reducing Reliability Centered Maintenance (RCM) costs while maintaining the accuracy of the analysis. The optimization strategy has undergone further evolution and at the present an Integrated Maintenance Program (IMP) is being put in place. Further cost refinement of the station preventive maintenance strategy whereby decisions are based on statistical analysis of historical failure data is being evaluated. A wide range of Operational Research (OR) literature was reviewed for implementation issues and several encouraging areas were found that will assist in the current effort of evaluating maintenance optimization techniques for nuclear power production. The road ahead is expected to consist first of resolving 25 years of data issues and preserving the data via appropriate knowledge system techniques while post war demographics permit experts to input into the system. Subsequent analytical techniques will emphasize total simplicity to obtain the requisite buy in from Corporate Executives who possibly aren’t trained in Operational Research. Case studies of containment airlock seal failures are used to illustrate the direct applicability of stochastic processes. Airlocks and transfer chambers were chosen as they have long been known as high maintenance items. Also, the very significant financial consequences of this type of failure will help to focus the attention of Senior Management on the effort. Despite substantial investment in research, improvement in the design of the seal material or configuration has not been achieved beyond the designs completed in the 1980’s. Overall, the study showed excellent agreement of the relatively quick stochastic methods with the maintenance programs produced at great cost over years of trial and error. The pivotal role of expert opinion via experienced users/problem owners/maintenance engineers in all phases of the method and its application was noted and will be explored in subsequent efforts. The results are displayed via economic alternatives to more easily attract the attention of Maintenance Managers. Graphical overviews of the data demonstrated that substantial insight can be gained by simply organizing the data into statistically meaningful arrays such as histograms. The conclusions highlight several very positive avenues to evaluate at this particular juncture in time.Copyright © 2002 by ASME