Simulation modeling for condition based maintenance optimization of multi-component systems with dependencies

In view of the deficiencies of present condition based maintenance modeling methods of multi-component systems and their difficulty of practical application, this paper presents a simulation model and an optimization method for these systems with economic, structural and stochastic dependencies. First, a Gamma process and parameters estimation method is used to describe the degradation of components. Then, economic dependency strength matrix, structural dependency reachable matrix and stochastic dependence dependency probability matrix are constructed respectively to model these three dependencies based on maintenance workflow, composition relationships and fault information. Finally, considering decision variables of the system and unit level at the same time, a simulation model to obtain the expected cycle costs of the system is presented, and a genetic algorithm (GA) solving process improved by Nelder Mead algorithm (NMA) is given according to the characteristics of the model. Numerical simulation results of a wire flight control system pitching channel subsystem demonstrate that the influence of dependencies on maintenance decision cannot be neglected. Cost saving and decision optimization results are achieved when dependencies and multi-component group maintenance are considered, which verifies the effectiveness and practicability of the models and methods presented above.