Reliability Analysis And Maintenance Strategy For Complex Systems Based On Generalized Stochastic Petri Net

In recent years,safety critical systems have been widely applied in many industries,such as automobile,aerospace and energy.These systems have high demands upon reliability.The extensive application of redundant technology ensures the high reliability of the system.However,redundant technology increases the complexity of the system structure.Reliability analysis of complex systems faces some challenges such as epistemic uncertainty,dynamic failure behaviors and common cause failure.The reliability analysis and maintenance strategy of these systems are meaningful and significant researches.Firstly,a dynamic fault tree is utilized to capture the dynamic failure behaviors of complex systems.The framework based on expert evaluating method is proposed to deal with the problem of epistemic uncertainty.This framework invites a group of heterogeneous experts to evaluate the failure rate of basic events.The intuitionistic fuzzy weighted averaging operator is used to aggregate the opinions of experts.The failure rate of basic events is obtained by defuzzification procedure.The fuzzy analytical hierarchy process method and decision alternative ratio evaluation method are utilized to decrease the impact of expert subjectivity on the results.Secondly,traditional analysis methods for dynamic fault tree have the problem of state space explosion.Therefore,the dynamic fault tree is converted into corresponding generalized stochastic Petri net to calculate some reliability parameters.The logic gates,bottom events,intermediate events and top event of the dynamic fault tree are mapped into the corresponding generalized stochastic Petri net.By using the Oris simulation software,the system reliability parameters can be acquired under different task time.Simulation result verifies the effectiveness of the proposed method.Thirdly,common cause failure is considered in reliability analysis.The ? factor model and efficient decomposition and aggregation approach are utilized to deal with the problem of common caused failure under different situation.The advantages of generalized stochastic Petri net simplify the process of the efficient decomposition and aggregation approach,which is simpler and more effective.Finally,aiming at the multi-attribute decision making problems of multi-source heterogeneous data,a multi-attribute decision making algorithm is proposed based on the improved grey relational analysis.Maximizing deviation method is incorporated in the algorithm to determine the weights of attributes,which increases the diagnosis accuracy.The closeness coefficient is obtained to formulate the maintenance strategy of complex systems.Simulation results prove the effectiveness of the proposed method.