New approach to the modelling of motor failure data with application to the engine overhaul decision process

Abstract

This paper presents a method to transform data on motor failures, that is extensively used in the literature for the benchmarking of reliability models, into the form of time-to-failure (TTF) data. A normal distribution is appropriate for modelling the TTF data. The estimated mean lives are subsequently fitted to a power-law failure point process model, and the life standard deviation after a failure can be obtained using a recursive relation. These are useful for inferring the life distribution after a future failure. The resulting model is applied to optimize the lifespan and predict the number of overhauls in a given operational interval. The approach developed in this paper is not limited to the specific case for which it is derived; it can be applied to similar problems or situations.