Maintenance Strategy Research For Complex Systems Based On Reliability Constraints

With the growing improvement of productivity and technological innovation ability in recent years,complex systems have been applied more and more in aviation,aerospace,energy,and other high-tech fields.Redundancy technology is increasingly adopted to improve the reliability and safety performance of complex systems,but it also increases the complexity of the system structure and adds more difficulty to the maintenance of complex systems.Traditional maintenance methods of complex systems are that engineering maintenance personnel take regular maintenance or corrective maintenance to the system according to the previous maintenance data,which is prone to excessive maintenance or insufficient maintenance.Therefore,traditional maintenance methods cannot meet the requirements of high reliability and security of complex systems.Aiming at the disadvantages existing among traditional maintenance methods,in this paper,reliability analysis of complex systems is combined with maintenance strategy,and the following aspects are studied:Firstly,a reliability analysis method based on the dynamic evidence network is proposed for the distinctive fault characteristics of complex systems.In this method,a dynamic fault tree is used to construct a fault model of complex systems to simulate the timing and dynamics of faults among components,and an interval number is applied to represent the failure rate of components to address the epistemic uncertainty.Meanwhile,aiming at the problem that traditional solutions for dynamic fault tree fails to deal with interval failure rate of components and common cause failure between components,a dynamic fault tree is transformed into a dynamic evidence network for quantitative analysis,and reliability parameters such as diagnostic importance factor,Birnbaum importance,and risk advancement worth can be obtained.Considering that the system reliability inevitably decreases with the course of running time,the maintenance factor μ is integrated into the fault model of complex systems,and the availability analysis of the engine fuel supply system is carried out.Secondly,due to the redundant nature of complex systems,even if a certain component fails,it will not affect the normal operation of the system and cause the problem of under-maintenance or untimely maintenance.This paper proposes a preventive maintenance strategy of complex systems based on reliability analysis.Firstly,the fault characteristics of complex systems are analyzed.If the reliability of complex systems at a certain time T is lower than the established system reliability threshold,that is,the preventive maintenance threshold of the system,preventive maintenance will be started.A multi-attribute decision making algorithm based on improved CODAS is proposed to solve the multi-attribute decision making problem of multi-source heterogeneous data,such as diagnostic importance,Birnbaum importance and risk achievement worth.The algorithm uses Euclidean distance as the primary metric for evaluation,and hamming distance is applied to compare multiple alternatives if their Euclidean distances are close.By combining Euclidean distance and Hamming distance,the best maintenance components can be found.The effectiveness of the proposed preventive maintenance strategy is verified by taking the hydraulic system of double row steel pusher as the research object.Finally,aiming at the problem of excessive maintenance of the system blindly based on historical maintenance data in traditional post-maintenance resulting in high maintenance cost,a post-maintenance strategy based on reliability analysis is proposed.A post-maintenance framework for complex systems based on reliability analysis is constructed in this paper.In this framework,a sensor placement optimization method based on component diagnostic importance is proposed to find the optimal sensor placement scheme of complex systems using component diagnostic importance factor as a criterion.Then the sensor information is incorporated into the dynamic evidence network of complex systems,and the diagnostic importance of system components is dynamically updated and reordered.The maintenance component of the system is determined by considering the diagnostic importance and maintenance cost of components.The effectiveness of the proposed maintenance strategy is verified by taking the vehicle-ground wireless communication system as the research object.