Research On Tool Fault Diagnosis Of NC Lathe Based On Improved SVM

With the continuous progress of science and technology,the intelligentization of CNC machine tools has become one of the important signs of the national industrial level.And CNC machine tool cutting covers almost all applications.Whether it malfunctions or not becomes the most important factor affecting product production efficiency and processing quality.At present,most of the CNC machine tools used by small and medium-sized enterprises in China do not have the function of tool condition diagnosis,and only rely on experience to make judgments and analysis,which leads to great differences in the accuracy of diagnosis.Therefore,in order to avoid the above problems,an effective method is urgently needed to diagnose and judge the state of the CNC machine tool.In this paper,the wear status of turning tools used in CNC lathe machining is studied and monitored online.The following work is done:1.Analyzes the tool's damage mechanism and wear process,and the tool's bluntness standard,establishes a CNC lathe tool fault diagnosis experimental platform,determines the experimental scheme,and collects the tool vibration signal data under different processing states during turning.2.Aiming at the phenomenon of modal aliasing in the processing of nonlinear vibration signals by the empirical mode decomposition method,an improved complementary set empirical mode decomposition(CEEMD)method is selected.This method is used to process the tool wear vibration data,which solves the problems of modal aliasing and endpoint effects;and experiments have shown that the first five high energy percentage values of each IMF component after CEEMD decomposition are selected as the characteristics to characterize the tool wear state The values are more differentiated and repeatable.3.The principle of support vector machine algorithm is introduced in detail,and the problem of parameter selection of support vector machine model is analyzed.Aiming at the problem that the support vector machine cannot adaptively select thekernel function parameters and penalty factors,a model diagnosis method based on the bacterial foraging algorithm to optimize the support vector machine is proposed.By using the bacterial foraging algorithm,the accuracy and computational efficiency of the selection of the parameters of the support vector machine decision function are improved,and experiments have shown that the optimized support vector machine model has higher efficiency and accuracy in identifying the tool wear status.4.Use the Win Form function of C# to build a real-time online monitoring and diagnosis system.The main functions include three parts.The first is to use hardware equipment to complete the collection and storage of lathe vibration data;the second is to use the above algorithm to perform data analysis and fault diagnosis on the data to identify the running status of the tool;finally,the Win Form communication function is used to store the diagnosis results in a database and Shown at the interface layer.The establishment of this system can allow users to better understand the current wear status of the tools,reduce unnecessary downtime,reduce maintenance costs,and improve production efficiency.