Remaining Life Kernel Density Prediction Method Of Multi-Component System Based On Copula Theory

With the highly intelligent equipment,the security and reliability requirements of the equipment are also improved.However,a sudden failure of the system during normal operation can interrupt the entire production process and may lead to serious consequences.Therefore,it is very important to evaluate the remaining life of the system in time and establish a reasonable and effective maintenance plan.In addition,many large systems are expensive to build,and reliability models are difficult to obtain through numerous disruptive life tests.At the same time,a large number of operation data representing system degradation will be generated during the operation of the system.Therefore,the remaining life of the equipment can be predicted by analyzing and processing the data obtained during operation.Based on the data-driven method,the Remaining Useful Life of the system(Remaining Useful Life;RUL),and then use the monitored data of the gear box to verify the research method in this paper.The main work and research results are as follows:(1)A method for predicting residual life kernel density of multi-component system based on a single Copula is proposed.Firstly,the marginal distribution of the remaining life of each component is estimated by using the nonparametric kernel density,and the related parameters in the Copula function are solved by the maximum likelihood estimation.Then,Akagi Information Criterion(AIC)was used for Copula optimization.Finally,a multi-component RUL prediction model was established by combining the Copula function and the kernel density estimator function under the condition of random correlation.Finally,the accuracy of the model is verified by the data of the gearbox.(2)A hybrid Copula based residual life kernel density prediction method for multi-component systems is proposed.Considering that the random correlation of different component degradation processes is very complex,only one of the Copula functions is likely to be biased when fitting the data.Therefore,a hybrid Copula model was proposed to model the residual life of the multi-component system,and the EM algorithm was used to solve the correlation parameters and weight coefficients in the hybrid Copula model.Then,AIC criterion was used to optimize the Copula.Finally,a multi-component residual life prediction model was established by the combination of Copula function and kernel density estimator function under the condition of random correlation.Finally,the accuracy of the model is verified by the data of the gearbox.