An Application of Inverse Probability to Fault Isolation

Inverse probability is applied to the problem of isolating faults in an electrical network by an analysis of failure symptoms. Prior to performing fault isolation, the performance of a network with respect to possible failures is established. Then, identifying each possible failure as a separate event, the procedure for isolating a fault is presented in terms of extracting a ``signal'' from random amplitude ``noise.'' An example demonstrating these procedures is presented. Finally, an approximate solution involving less computation is summarized.