Reliability Evaluation And Prediction Of High Performance CNC Machine Tools In Service

With the continuous integration of new generation information technology,artificial intelligence technology and manufacturing technology,the transformation of manufacturing industry to intelligence has become an inevitable trend.The high performance CNC machine tools and its flexible manufacturing system are the important foundation of intelligent manufacturing.The high performance CNC machine tools with complex structure,variable operating conditions and complex operating environment will lead to irreversible degradation of the performance of the CNC machine tools system and key functional components.In the service stage,performance degradation and frequent faults will seriously affect the machining accuracy and production efficiency.Therefore,how to ensure the service performance of CNC machine tools has become the focus of designers,manufacturers and users.For high performance CNC machine tools,it has the characteristics of multiple component degradation characteristics and less reliability data samples.The traditional single performance evaluation method based on failure data has some limitations.Based on "state monitoring data","standard S-shaped specimen" and"multi-source data fusion",this paper evaluates and forecasts the service performance of high performance CNC machine tools in terms of life prediction and reliability evaluation of the whole machine operation.The main work of this paper includes the following aspects:(1)Based on the hybrid prediction method,the Remaining Useful Life(RUL)prediction model of key components is proposed.The degradation degree and failure time of key functional components of CNC machine tools with degradation failure are discrete due to the influence of different operation conditions,use environment,maintenance degree and other factors.A RUL prediction model based on RVM and improved power function is constructed by using the integrated prediction method of data-driven and artificial intelligence.The model can adapt to the uncertainty of the degenerated process,quickly and conveniently evaluate the running state without affecting the actual cutting process.(2)The evaluation method of the reliability of high performance CNC machine based on S-shaped specimen is studied.For the high performance CNC machine tools in service,the degradation of its performance and accuracy retention in time dimension is different under the complex and multi factor dynamic action.At present,there is no unified standard for reliability evaluation based on machining accuracy in operation condition.This paper discusses the error transfer model between the performance degradation of joint surface and the mapping of machining accuracy,and puts forward the evaluation model based on the operation reliability of S-shaped specimen.The model applies constant exciting force to the whole machine through standardizing the processing conditions of the S-shaped specimen,and regularly monitors the characteristic signals of the fixed cutting conditions.Three dimensional evaluation indexes(entropy dimension,visualization dimension of three-dimensional Hilbert amplitude spectrum,and numerical quantification dimension of center of gravity frequency of marginal spectrum)are constructed to comprehensively evaluate the deterioration degree of machine tool joint surface,so as to quantify and evaluate the operation performance and machining quality of CNC machine tool.(3)The comprehensive reliability evaluation model of high performance CNC machine tools based on multi-source information fusion is constructed.The operation reliability of high performance CNC machine tools is not only related to the inherent reliability in the design and manufacturing stage,but also to the use and maintenance in the service stage.This paper systematically studies the reliability data fusion modeling method of three dimensions:fault maintenance record data,real-time monitoring sensor information of running state and machining accuracy,and constructs the comprehensive evaluation model of high performance CNC machine tools based on Fuzzy AHP.The evaluation index system of operation reliability and quality reliability is established.The reliability evaluation method can not only take into account the "congenital factor" reflected by machine tools failure time,but also take into account the "acquired factor" reflected by operation status and processing accuracy,so as to evaluate the comprehensive reliability of CNC machine tools in multi-dimensional and accurate way.(4)The reliability evaluation model of CNC machine tools in production line based on Fuzzy Bayesian network is constructed.Deeply integrate subsystem reliability test data,field operation and maintenance data,and maintenance data of similar systems.The fuzzy theory and Bayesian network are combined to solve the problem that it is difficult to accurately obtain the state probability of each root node in the multi-state system,and improve the ability to deal with uncertainty problems.(5)A joint maintenance decision model based on life prediction is proposed.In order to ensure the high reliability and low cost operation of high performance CNC machine tools,a joint decision-making model based on condition based maintenance and planned maintenance is proposed to solve the problem of over repair or under repair caused by planned maintenance.The model comprehensively utilizes the life distribution function of the key functional components and the life prediction results of individual components.The minimum average maintenance cost is taken as the optimization objective,and the mainterrance interval and residual life maintenance threshold are used as the optimization variables.Through Monte Carlo simulation,the collaborative analysis of maintenance cost,maintenance interval and maintenance threshold is carried out,which provides technical support for the decision-making of maintenance plan and the budget of maintenance cost.