Fundamental study on cost benefit brought by insulation diagnosis

This paper deals with the statistical estimation of the maintenance cost reduction when insulation diagnosis for power apparatus is performed. A failure loss is expected if a proper alert has not been made because of "too loose diagnosis". On the other hand, prematurity loss is expected if the apparatus has been replaced because of a too early alert made by "too severe diagnosis". The expected total loss is estimated as a function of tolerable risk ratio of the diagnosis. The loss depends on accuracy of diagnosis and age of apparatus as well. Condition for the minimum cost is realized by selecting appropriate tolerable risk ratio. The cost reduction effect by diagnosis was also compared with the maintenance costs without diagnosis. If the loss is lower than both "expected prematurity loss with immediate replace" and "expected failure loss with continuous operation", a cost benefit is brought by the diagnosis. The better diagnosing accuracy brings the lower loss. A case study was performed assuming some realistic parameters.