Machining Center Fault Diagnosis Based On Maximum-probability Path Search Model

Improving the reliability level of domestic machining center is the primary task of complying with the "Made in China 2025" innovation drive and quality first as the basic policy.Frequent faults in domestic machining center can cause problems such as poor user experience,low equipment life,and affecting the production capacity of enterprises.Therefore,timely detection and diagnosis of faults in domestic machining centers have important research value and practical significance.Aiming at the deviation of the traditional single fault diagnosis strategy of domestic machining centers,this paper proposes a comprehensive fault diagnosis technology based on integrated data mining and intelligent search,and uses a series of domestic horizontal machining centers as research objects to verify the application.The main research contents are as follows:Firstly,based on the collected historical fault information,the fault correlation analysis is carried out.Based on the graph theory,the fault transfer relationship between the machining center subsystems is transformed into a directed graph.By using the adjacency matrix to matrix the directed graph model,the RankClus algorithm is introduced,and the machining center subsystem is divided into three modules by modular analysis of the clustering and sorting of the subsystems in the directed graph.According to the "four degrees" theory,the DEMATEL analysis method is used to divide the subsystems in the module into three unidirectional propagation levels to construct a modular fault propagation level model of the machining center.Secondly,based on the fault interval time of the machining center subsystem,the subsystem fault data is analyzed,and the optimal distribution type of the subsystem is determined by the Minitab software.Based on the idea of comprehensive graph estimation and maximum likelihood estimation,the Newton-Raphson iterative method is used to obtain the parameter estimation value,and the parameter estimation value is modified by combining the small sample theory.The K-S test was used to test the goodness of fit,and the subsystem fault distribution model was established accordingly.Thirdly,the Copula theory is introduced,and the subsystem reliability is used as the edge distribution.The binary Gumbel Copula is used as the connection function to establish the joint distribution function between components.The copulafit function in MATLAB software is used to calculate the correlation coefficient.Based on the limit theory of integral sum,it is deduced that when the machining center maintenance time is much smaller than the normal working time,the unreliability function is approximately equal to the failure rate function.Then combined with the joint distribution function and the subsystem reliability function,the conditional probability definition is used to derive the probabilistic model of the associated fault propagation in the machining center.Finally,the modular fault propagation level model of the integrated machining center and the associated fault propagation probability model of the machining center are applied to the fault diagnosis of the machining center subsystem using the maximum probability search model.The simulation test is carried out through the Eclipse platform to verify the correctness of the method.