Multi-dimensional Dynamic Bayesian Network And Its Application On Pusher Hydraulic System Reliability Analysis

Bayesian network analysis method is an important method to deal with the difficulty of data acquisition and the uncertainty of logical relationship.It is widely used in the fields of reliability analysis,fault assessment and condition monitoring.The discrete-time Bayesian network analysis method has calculation errors in quantitative analysis,and cannot describe the real-time change trend of the system failure probability.The continuous-time Bayesian network analysis method improves the calculation accuracy and can also describe the change of the system failure probability over time.Trends,but still limited to the analysis of single-factor effects,and cannot describe the correlation of failures between components.For this reason,a multi-dimensional dynamic Bayesian network analysis method considering related failures is proposed.Firstly,by introducing the unit step function to describe the failure timing of the root node,and the impulse function to describe the failure time of the leaf node,the modeling and analysis of the continuous-time Bayesian network analysis method are completed,and on this basis,the multi-factors are described.Influencing,the multi-dimensional dynamic Bayesian network analysis method is proposed.Further,in order to quantitatively characterize the influence of components on the failure probability of the system under the influence of multiple factors,a multi-dimensional dynamic Bayesian network importance analysis method is proposed,and the calculation formula is defined,and the reliability analysis content is expanded and extended.The method is compared to verify feasibility.Furthermore,considering the influence of the correlation of failures between components on the system reliability,the multidimensional dynamic Bayesian network analysis method considering the correlation failure is proposed by combining the correlation failure with the multidimensional dynamic Bayesian network,and the reliability of the hydraulic jacking mechanism is analyzed and verified.Finally,through the reliability analysis of the hydraulic system of the double-discharge hydraulic pusher,the dynamic fault tree and the multi-dimensional dynamic Bayesian network model were constructed according to the failure mechanism.Through the proposed analysis method,the hydraulic system of the pusher was obtained.The distribution curve of system failure probability and component importance under the influence of factors,as well as the failure probability distribution curve of the steel machine system under the influence of related failures,provide quantitative information for hydraulic system maintenance and improvement,as well as for system identification weak links and optimization design the basis.