A Fault Isolation Approach for Data-Driven Device Replacement Decision Making

Despite the ubiquity and the prominence of device replacement decisions, the process is not always as simple as it seems. Especially as data-driven approaches are becoming more vital and intelligent systems have recently outperformed the classical techniques. The objective of this paper is to propose a solution to provide the set of decisions for replacing a whole asset. To accomplish this, first, we translate the deterioration level into the monetary scale via a data-driven approach. This is done by modeling the aging effects via fault occurrences; then utilizing fault isolation techniques for the determination. Finally, the set of decisions can be made by the optimization procedure.In the end, to implement the proposed solution, an actual fault isolation data set is reproduced to form a pertinent data set for a fictitious firm. Then an ANN is exploited for the fault isolation, and a game-theoretic approach comprising GA is utilized for the decision-making part.