Study On Preventive Maintenance Strategies Of Mission-Based System

Preventive maintenance including inspection,preventive replacement and preven-tive repair is a common activity to prevent system's failures.The purpose of inspections is to reveal the system's working state,and thereby to carry out necessary maintenance actions if needed.Firstly,we propose a postponed repair policy for a single-unit mission-based system.The system's states can be classi-fied into four states,namely,normal,minor defective,major defective and failed states.The durations of the normal,minor defective and major defective states constitute a three-stage failure process.Failure of the system is self-announcing,whereas the de-fective states both minor defective state and major defective state can only be identified by inspections.Inspections are schemed at periodic times as well as every comple-tion of missions to identify the defective states.Most preventive maintenance models assume the instantaneous execution of repair once the defective state is revealed.In contrast,we relax this assumption and allow repair to be postponed for the purpose of saving cost.We assume if the system is major defective at a random inspection,re-pair will be postponed if the time to the following periodic inspection is less than a predetermined threshold,otherwise repair will be immediate.We also consider that if the system is major defective at a periodic inspection,it is repaired immediately.The renewal probabilities of the system are presented and then the long run expected cost per unit time are derived accordingly using the renewal-reward theorem.Three another policies,i.e.,a pure periodic inspection policy without postponed repair,a pure random inspection policy without postponed repair and a periodic and random inspection policy without postponed repair are proposed for comparison.The numerical example shows that when both the random inspection cost and the postponed repair cost are low,the postponed repair policy will be the most cost-effective policy compared with another three policies.The related papers discussed maintenance problems on a single-unit mission-based system fall into two categories,i.e.,non-repairable categories and repairable categories.For non-repairable categories,the replacement policy is used.One of the replacement policies is an age-dependent policy.Most systems suffer some deterioration with time from wear,fatigue,and damage,and ultimately fail(internal deterioration).In addition to internal deterioration,the systems may expose with particular effects,i.e.,external shocks,such as icing,thunder and lighting,earthquake,high temperature and so on,leading to the failure of the system immediately.This paper,secondly,takes up three age replacement models with random working cycles subject to fatal shocks.A standard age replacement policy(SARP),where corrective replacement(CR)is done immediate-ly after failure caused by both internal deterioration and external shocks,and preventive replacement(PR)is done before failure at a total operating time T,is proposed.More-over,the unit is replaced immediately after failure caused by both internal deterioration and external shocks,and preventive replaced at a total operating time T or at a ran-dom working cycle Y before failure,whichever occurs first and last,which are called replacement first policy(RFP)and replacement last policy(RLP),respectively,are for-mulated.Expected cost rate for each policy is obtained,and the optimal operating time which minimizes it is derived analytically.Comparisons among optimal expected cost rate of SARP,RFP and RLP are made in detail.It is shown that neither RFP nor RLP is better than SARP when PR cost at T is equal to PR cost at Y theoretically and numeri-cally.Furthermore,it is shown numerically how the mission rate affects the comparison results.For repairable categories,there's the common assumption:as a failure occurs,the system suffers one of two types of failure based on a specific random mechanism,type-I(repairable)failure is rectified by a minimal repair,and type-?(non-repairable)failure is removed by a corrective replacement.Most maintenance(replacement or repair)policies are generally assumed that there is an unlimited storage of spares at any time available for immediate replacement.However,this assumption might be unrealistic when storage is costly.In this case,a random lead time for spares delivery upon order should not be neglected.In this paper,finally,three generalized age replacement poli-cies considering the random lead time for an operating system which works at random time and maintains imperfectly are proposed.The mean cost rate of the assumed mod-els are derived explicitly,and the optimal schedules that minimize the mean cost rate are also determined analytically and computed numerically.