Improving system maintenance decisions: A value of information framework

Decisions concerning maintenance strategies have become increasingly important as managers face more complex systems in the manufacturing, construction and process industries. The choice among maintenance methods, reactive, preventive or predictive, is often complicated by uncertainty about component failure, reliability of the predictive sensor readings, and the potential for significant financial exposure. This paper describes a decision-science methodology for evaluating alternative maintenance strategies. We use a value of information framework to analyze the maintenance decision in the context of the choice alternatives maintenance managers face, the key sources of uncertainty associated with component failure and sensor readings, and the associated maintenance and repair costs. This work presents an application of the decision model to a gantry crane system used at a hazardous waste retrieval site. This methodology can guide managers toward a more formal evaluation of maintenance methods, give insight into the analysis of competing choice alternatives, and significantly improve the quality of maintenance decisions. The technique described and the application discussed give the engineering economist a more robust approach to analyzing the complex issues associated with this class of decision problem.