Research On On-line Monitoring Method Of Milling Process Stability

High-speed milling processing is one of the main processing methods in key fields such as aviation,aerospace,vehicles,and ships.Its processing stability has a significant impact on improving product quality,production efficiency,cycle time and reducing costs.Cutting chatter is a typical unstable phenomenon in the machining process,which has a destructive effect on the surface quality of the product,the wear of the tool,and even damages the machine tool spindle in severe cases.In order to ensure the stability of the machining process,the current main research method is to establish a stability prediction model based on the dynamic characteristics of the machine tool.However,because the actual processing is affected by the environment,load,tool,machine tool and workpiece structure,the deviation between the theoretical model and the actual model makes it difficult to be widely used in practice.Online monitoring and identification of chatter through sensor signals has become a more important direct and effective method.This article mainly focuses on the research of online monitoring method for the stability of the milling process.The research work is as follows:(1)Analyze the cause and influence of machining instability,analyze the milling process dynamics,and study the optimal selection and arrangement method of sensors.The research shows that the unreasonable design of cutting parameters is the main factor causing the abnormal stability in the actual machining,and the acceleration sensor sensitive to the abnormal vibration is selected to collect the signal.(2)The analysis and processing methods of vibration signal and cutting force signal in the process of chattering are studied.Based on the time domain,frequency domain,and time-frequency domain,respectively,the signal characteristics and change trends of the processing in stable,flutter transition and flutter phases are studied,which lays the foundation for the feature extraction of flutter state.(3)A feature extraction algorithm based on wavelet packet energy ratio is proposed,combined with LSSVM model for online monitoring and prediction of machining chatter.By comparing K-nearest neighbors,BP neural network,decision tree,SVM,LSSVM and other classification models,it is found that the classification algorithm based on LSSVM is better in recognition time and recognition accuracy,and can effectively monitor and predict chatter.(4)The design method of the stability region of the machining process under mass production conditions is studied.In view of the vibration signal energy frequency shift that occurs when abnormal conditions such as machining chatter occur,a stability monitoring algorithm based on S transform-singular value entropy is proposed.Experimental research shows that the singular value entropy is significantly sensitive to the abnormal machining stability of machine tools and has certain predictive characteristics.Through the accumulation of cutting data and learning statistics,based on the singular value entropy,the stability domain of different machining processes is established,which can be effectively applied to the evaluation of the stability of the machining process.