Reliability Analysis And Selective Maintenance Of System Considering Multiple States

Machinery,electronics,hydraulics and control are highly integrated in modern engineering system.As a cutting-edge complex system,its failure would result in significant economic losses.With the development of frontier fields such as aerospace,deep sea and deep earth exploration,research on the reliability of complex systems has been included in the National Science and Technology Development Program.Both ensured the complex system with a high reliability and formulated scientifically effective maintenance decision-making have received extensive attention from the engineering and academia.In this paper,several complex systems are taken as the research object.Its reliability analysis and selective maintenance decisionmaking have been deeply studied from the perspectives of complexity,multi-state and uncertainty.Furthermore,the methods for reliability analysis and maintenance decision-making of the complex system are proposed.Meanwhile,the effectiveness and practicability of the proposed method are verified through specific engineering cases.The main research contents and results are described as follows:(1)Considering the premature state,a fuzzy Goal-Oriented(GO)reliability analysis method for uncertain multi-state systems is proposed.Modern engineering systems have a wide variety of subsystems and components with different functions.There is a widespread nonfaulty premature state during operation,such as gyroscopes,accelerometers,motors,etc.Ignoring the premature state will lead to inaccurate system reliability assessment results.In addition,the internal relationships of complex systems are intricate,and the operating environment is complex and changeable,resulting in fuzzy uncertain effects of components and subsystems on the system reliability.Therefore,with the combination of fuzzy mathematics and the GO reliability model,the triangular fuzzy number and the extended principle are introduced,in order to quantify the fuzzy uncertainties of input and output data respectively.As a result,a novel fuzzy GO reliability analysis method of the uncertain multi-state system is proposed.Taking a certain PINS as an example,the reliability analysis is conducted by FTA(Fault Tree Analysis)method,traditional GO methodology and fuzzy GO methodology respectively.The results show that the fuzzy GO method is effective,which is a reasonable promotion of the traditional GO method.(2)A grey fuzzy Bayesian Network(BN)reliability analysis method for uncertain complex multi-state systems with intermediate states is proposed.During the evolution from normal state to failure state,the system and its components exist a variety of failure states.And the boundaries between the states are not obvious,with a certain degree of fuzzy uncertainty.In addition,the limited number of test samples and insufficient data would make it difficult to accurately determine the failure probability or performance level of system components.In this paper,the uncertainty of the complex multi-state systems is fully considered.It consists of two aspects:the uncertainty of failure logic relationship between the system and the component,and the uncertainty of the component failure state.Fuzzy mathematics and gray system theory are introduced into the reliability analysis model of multi-state BN.A complex system reliability analysis method based on uncertain membership functions and interval feature quantities—grey fuzzy BN method is proposed.By the use of the proposed method,the reliability of the satellite propulsion system is analyzed,and the system reliability under the condition of the current failure state and failure rate of each root node is solved.The importance of the component is analyzed by the two-way reasoning mechanism of BN,and the reliability characteristics,such as system reliability and component importance,are calculated and expressed in the form of interval value.The research results can provide an important reference basis for reliability analysis,fault diagnosis and maintenance decision-making of complex uncertain systems.(3)A selective maintenance model for complex systems based on multiple repairpersons and Particle Swarm Optimization(PSO)algorithm is proposed,and a maintenance allocation algorithm for system components based on multiple repairpersons is also proposed.No maintenance,minimal maintenance,replacement maintenance and multiple intermediate maintenance levels are considered in the proposed model.More importantly,factors,such as the pre-maintenance state of the component,the effective age of the component and the maintenance cost,are introduced into the imperfect maintenance model,which is more in line with the engineering reliability.And the problem of how to allocate multiple maintenance tasks to multiple repairpersons so as to minimize the system maintenance time is also solved by the proposed algorithm.The proposed algorithm is introduced into the PSO algorithm to solve the selective maintenance model of the complex system under the condition of multiple repairpersons and imperfect maintenance,which expands the application of the PSO algorithm to maintenance decision-making under multiple repairpersons.Taking the Strap-down Inertial Navigation System(SINS)as a case,the influence of the number of repairpersons on the SINS selective maintenance decision and imperfect maintenance(intermediate maintenance levels)in the imperfect maintenance model on the reliability of SINS is studied.Meanwhile,the influence of maintenance resources on the reliability of SINS is also studied.In addition,the effectiveness and advantages of the proposed model and algorithm are verified.(4)A selective maintenance model for multi-state systems considering the random uncertainty of the system mission period,the random uncertainty of the system mission break,and the requirements of different system performance levels is established.In the established model,random variables that obey a certain distribution(e.g.normal distribution)are utilized to characterize the randomness of the mission period and mission break,and the service age regression model is utilized to describe the effective age of the system components after maintenance.At the same time,the quantitative relationship among the system maintenance cost,maintenance time and the service age regression factor is constructed in the established model.Based on the multi-state system reliability and the general generating function technology,the task completion rate model of the component and system is established,and the solution method for the selective maintenance decision of the multi-state system based on the PSO algorithm is given.Taking the coal transportation system of a thermal power station as a case study,the influence of random uncertainty of mission period on the selective maintenance decision of the system and random uncertainty of mission break and imperfect maintenance on the selective maintenance decision of multi-state system is studied.The sensitivity analysis of the system task completion rate is carried out under a variety of conditions,and the effectiveness of the proposed model and solution algorithm is verified.The research results of this paper not only can provide a useful reference for the design,manufacture,use and maintenance of modern engineering systems,but also provide new methods for the reliability analysis and selective maintenance decision-making of modern engineering systems.