Optimal preventive maintenance for a device submitted to random and deteriorating failures

A Markov model is used to characterize the performance of a continuously operating device that is subject to random failure as well as failure generated by deterioration. The optimum preventive maintenance of this device is determined by maximizing its availability with respect to the mean time to minimal preventive maintenance. An example is provided to illustrate the method. The computational work is performed through Matlab^® software. The results indicate that, as deterioration rate decreases, availability increases and total cost of maintenance decreases, but the respective increase/decrease occurs in a non-linear way. On the other hand, as random failure rate decreases the respective increase in availability and decrease in total cost of maintenance occurs in almost a linear way. Thus it is concluded that preventive maintenance is more sensitive to the range of deterioration failure rates than that of random failure rates.