Determining replacement policies for bus engines

In this study, a power law process (PLP) for the failures of an engine, regarded as a complex repairable system, in a minimal repair set‐up (i.e. only a small proportion of the constituent parts of the engine are replaced on repair) with the engine regularly replaced within period T or replaced at the Nth failure after its installation, whichever occurs first, was examined. First of all, the Laplace test was used to check for the existence of a deteriorating trend in the failure data. Second, model parameters of the PLP were estimated using the maximum‐likelihood estimation method. Third, the Cramer‐von Mises test was used to test its goodness of fit. Finally, the optimal replacement policy based on minimising the long‐run expected cost per month for each type of engine was determined. The statistical inference procedure involving the maximum‐likelihood method for the PLP is based on the associated large‐sample theory. This implies the need to have a lot of data before conducting a statistical analysis. Unfortunately, the authors do not have sufficient data to conduct a real statistical analysis and to bring a significant conclusion to the considered application. The paper describes an industrial application of a PLP and a theoretical replacement model.