STOCHASTIC ANALYSIS OF A TWO-UNIT PARALLEL SYSTEM WITH PREVENTIVE MAINTENANCE AND SINGLE DEGRADED FAILURE

This paper deals with analysis of a parallel system having two identical components. If any one of the two new units is operational the system is considered in upstate and system is considered in down state as and when single degraded unit is available for use and such a failure is known as single degraded failure. There are two types of servers, ordinary and expert, and visit the system as soon as it is needed. The preventive maintenance is provided to the new operative unit after its maximum operation time. The repair and preventive maintenance of new units are performed by ordinary server. While the inspection and repair of the degraded units is performed by expert server. Upon failure, the degraded unit is inspected and if an inspection shows that repairing of a degraded unit at the time of failure will be detrimental to the system, it is replaced with a new one. All the random variables related to failure, repair, inspection and preventive maintenance follow arbitrary distribution. The regeneration point technique and the semi-Markov process are used to calculate system metrics such as mean time to system failure (MTSF), mean sojourn time, availability, busy period of the servers and expected number of visits by server. The numerical findings for MTSF, availability, and profit of the system model are also analysed for a particular case.