| Bayesian network is one of the powerful methods for system reliability modeling.However,the traditional Bayesian network reliability analysis method does not consider the correlation failure modes among components,which makes the accuracy of reliability modeling based on the assumption of component failure independence is poor.Therefore,a multi-dimensional dynamic Bayesian network model of correlation failure based on Copula function is proposed,which comprehensively considers dynamic,correlation failure and multi-dimensional influencing factors,making the reliability modeling closer to the actual system.Firstly,based on the continuous time Bayesian network model,the concept of multidimensional is introduced,and the reliability analysis method of multi-dimensional dynamic Bayesian network is expounded.Copula function is introduced to describe the correlation failure relationship between components.Through theoretical derivation and calculation,the expression of correlation failure logical relationship between components is given.Copula function is selected by nonparametric kernel density estimation method,and a multi-dimensional dynamic Bayesian network model of correlation failure based on Copula function is proposed.Taking a hydraulic system as an example,the Copula function is optimally selected.At the same time,the hydraulic system of the cutting part of the roadheader is taken as the research object,and the reliability analysis and importance and sensitivity analysis are carried out by using the related failure Bayesian network model,which provides theoretical guidance for system troubleshooting and maintenance.Then,through the reverse reasoning mechanism of Bayesian network,starting from considering the posterior probability under the correlation failure,the reliability assignment method based on Bayesian network is studied;the network posterior probability replaces its prior probability,so as to achieve the purpose of selecting the best maintenance plan.Finally,taking the hydraulic system of a horizontal waste paper baler as an example,by analyzing its working principle and failure mechanism,a correlation failure Bayesian network model is constructed to solve the multi-dimensional fault probability distribution,importance and sensitivity distribution,and the reliability distribution and maintenance decision of the system components are carried out,which provides reference for system design and maintenance. |
