Locating the faulty component in a malfunctioning repetitive operation

When the output of a repetitive operation is nonconforming, it is desirable to use a method that can locate the particular component that caused the nonconforming output. To minimize the probability of a wrong decision, a Bayes decision approach is used to develop a criterion for locating the source. The criterion is the ratio of the posterior and prior probability values of the parameters associated with each component when parameters assume values within their tolerance ranges. This criterion provides an ordering of inspection of the components of the process and locates the most likely malfunctioning component. It can be applied to an existing faulty model whose parameters can have been determined in an arbitrary way (physical principles, brainstorming, etc.). The criterion can be used in conjunction with expert opinion to improve decision accuracy. It is particularly suitable when there is little information about the performance of the components of an operating process. Supporting examples are provided.<
>