Reliability Assessment Of Key Functional Components Of CNC Machine Tools Based On Multi-source Information Fusion

CNC machine tools are the important equipment of the manufacturing industry,whose reliability has become a key technical index affecting the development of the industry.It is an important way to ensure the reliability of CNC machine tools that ensuring the reliability of key functional components of CNC machine tools.The reliability assessment method of the key functional components of CNC machine tools can not only be used to assess the reliability,but also provide an important basis for reliability design,reliability improvement and preventative maintenance with assessment results.Therefore,it is of great theoretical value and practical significance to study the reliability assessment method for key functional components of CNC machine tools to obtain accurate assessment results.Failure data is an important basis for reliability assessment,but along with the continuous improvement of the reliability of the key functional components of CNC machine tools,the required reliability test period is continuously extended.The failure data obtained within a limited time is seriously insufficient.The accuracy of the assessment results is very low according to limited failure data.In the whole life cycle of key functional components of CNC machine tools,there is multi-dimensional,multisource reliability information,and it would make up for the inadequacy of failure data by making full use of this information.However,some of the information is missing or difficult to obtain,and how to use the imperfect information to mine out useful reliability information to improve the reliability assessment accuracy is an urgent problem to be solved.At the same time,the current research on reliability assessment methods based on multi-source information fusion usually integrates the information within the same dimension,and the information source is relatively single,which tends to lead to a large deviation of the assessment results.Under the premise of the minimum fusion error,the method of how to integrate the information of different dimensions and sources still needs to be developed.Therefore,supported by the National Science and Technology Major Project of "High-grade CNC Machine Tools and Basic Manufacturing Equipment",a reliability assessment method for key functional components of CNC machine tools based on multi-source information fusion is proposed.According to the characteristics that machine tools have various fault modes,mixture Weibull distribution is used as the distribution model.This method mines and fuses interval censored previous product reliability information,mixed uncertain reliability simulation information,fuzzy judgments information of experts,non-linear performance degradation information and right censored small sample failure information to evaluate the reliability of key functional components of CNC machine tools.It takes the servo turret and spindle as examples to prove that this method can effectively improve the assessment accuracy.The main research contents of the thesis are listed as follows:(1)The reliability information of previous products can be used as an important reference for the reliability assessment of existing products.Aiming at the problems of incomplete records and inaccurate failure time of the previous product of key functional components of CNC machine tools,a reliability assessment method considering interval censored data is proposed.The likelihood function containing interval censored data is established.The EM algorithm and the Delta method are then used for point estimation and interval estimation respectively.The proposed method is verified by Monte Carlo simulation.Finally,the method is applied to the product reliability evaluation of servo turret and spindle,and the previous product reliability evaluation results are obtained.(2)In order to obtain the reliability simulation information as the reliability assessment information source for the key functional components of CNC machine tools,a reliability simulation method based on mixed uncertainty is proposed.Considering the function principle of the system,the fault tree is established by using the method of "function-motion-action".The cognitive uncertainty is quantified by the extended probability box theory and the coefficient of variation.The key functional components of CNC machine tools are considered as multi-state maintainable systems,and the reliability simulation with mixed uncertainty is implemented by using the twolayer Monte Carlo method.Finally,the proposed method is applied to the reliability simulation of servo turret,and the reliability simulation results are obtained.The Bimbaum importance degree is used to evaluate the importance of each subsystem.(3)In order to integrate the previous product reliability information,expert information,and reliability simulation information as the prior information of product reliability assessment,an information fusion framework based on serial structure is established.An expert system is established by using the "best and worst" method and fuzzy theory.Based on the previous product reliability evaluation results,the expert judgment results of the existing product reliability are obtained.Then,the results are fused with reliability simulation information by Bayesian fusion method,and the final multi-source information fusion results are obtained by using the improved sampling importance resampling algorithm.(4)In order to solve the problem of low accuracy of reliability evaluation based on Weibull mixture distribution under small samples,a reliability assessment method of small samples considering right censored information based on Bayesian theory is proposed.The time before failure is predicted by nonlinear degradation process and then fused with the failure data of small samples.The fusion results above are taken as the prior information,and classified by the conditional probability density function.The parameters of each sub-distribution are solved by the Monte Carlo Markov chain method.Finally,Monte Carlo simulation and example analysis are used to prove the effectiveness of the proposed method.Through the research in this thesis,a reliability assessment method for key functional components of CNC machine tools based on multi-source information fusion is proposed.The reliability assessment results are obtained by taking servo turret and spindle as the objects,and it is proved that the proposed method can effectively improve the accuracy of reliability assessment results.It can provide an important theoretical basis for reliability design,reliability improvement,and preventive maintenance in future.