Research On Modeling And Solving Method Of Maintenance Strategy Optimization For Repairable Structures

The strength,bearing capacity and reliability of the structure gradually decrease with the development of time due to the combination of external load,service environment and internal factors of materials during its service,which has an important impact on its safety and effective service.Reliability is an important index to indicate the condition of the structure,so it is of great importance to evaluate the structural time-dependent reliability during its service and to maintain them properly accordingly.Different maintenance strategies,including maintenance times,intervention time and maintenance options,have different impacts on the performance of the structure as well as expenses.Therefore,determining the optimal maintenance strategy of the structure plays a key rule in the balance between its safety and economy.To solve this problem,the main contents of this dissertation are as follows:(1)Considering maintenance planning is a multi-stage decision-making process,a cooperative PSO-DP method,named co-PSO-DP,is proposed for obtaining the optimal solution for the multi-stage multi-option maintenance problem.Thereafter,a total life-cycle cost-based(LCC-based)reliability-based design optimization(RBDO)model is formulated,in which the initial construction cost together with the subsequent maintenance cost is considered simultaneously and both the structural parameters and the maintenance strategy parameters,including the intervention time and options,are taken as the design variables.Again the proposed co-PSO-DP algorithm is applied to identify the optimal solution of the model.The results show that compared with the traditional particle swarm optimization or genetic algorithm,the proposed method can effectively solve the complicate multi-stage maintenance optimization problem,and has good robustness and great advantages in computational efficiency.(2)Considering that the structure experiences both continuous degradation and random shocks in the process of degradation,the time-dependent reliability model based on stochastic process and its calculation method are studied.Thereafter,a condition-based aperiodic maintenance optimization model is established.The aim of this model is to determine the threshold of preventive maintenance and the number of maintenance times,so as to minimize the cost per unit time over the infinite or a finite time span.Two typical failure modes may occur during the service life of the structure,i.e.(a)soft failures caused by continuous degradation and medium shock loads;(b)hard failures caused by severe shock loads of the same load process.Gamma stochastic process is used to simulate the continuous degradation caused by time effect,and homogeneous Poisson stochastic process is used to describe the stochastic shock loads.The results show that the additional random shock loads have significant effects on the time-dependent reliability and the optimal maintenance strategy of the structure compared with the pure continuous degradation process.In addition,the optimal maintenance strategy highly depends on the structural design life T.(3)A generalized multi-mission selective maintenance optimization model for a system with several multistate components over a finite time horizon is established.The model aims to determine the optimal maintenance time and the corresponding maintenance options of each component under reliability constraints so as to minimize the total cost.The homogeneous continuous time Markov chain is utilized to characterize the degradation process of each component,and a state probability vector is used to describe the condition state of each component at any time during the deterioration-maintenance process.Considering the structural dependence of components in the system,a directed graph is used to represent the precedence relations among these components,and this correlation is incorporated into the corresponding cost model.The results show that ignoring the dependence between the components in the system may result in a conservative solution,which may lead to more resources consumed.In addition,compared with the maintenance strategy with periodic time intervals,performing maintenance actions with aperiodic one has better economic benefits.