Research On Tool Wear Prediction And Feedback Control Of Cnc Machine Tools

The CNC tool is the direct execution equipment of the automated processing system.Due to the contact between the workpiece and the cutting edge during the operation of the CNC machine tool,wear is inevitable.The degree of tool wear directly determines the surface quality of the workpiece and even the performance of the machine tool,and then has an impact on the processing efficiency of the entire manufacturing workshop.However,the development of tool materials is far from meeting actual needs,and the research on the wear resistance of cutting tools has reached a bottleneck.Therefore,in view of the actual problems such as unstable processing quality and high cost caused by untimely tracking of machine tool wear information in the manufacturing enterprise's workshop,researches are carried out on the acquisition of tool wear perception data,the construction of wear prediction models,and the optimization of processing parameters.Study the monitoring of machine tool operating status and tool wear.To avoid the influence of sensor noise,use OPC technology to realize CNC communication,a detection mechanism with dual-lens vertically distributed is used to obtain sensing data.Online monitoring of the life cycle of the tool in the processing stage,predict the wear of the tool in the next machining cycle,use this as a reference to control the machine tool to perform coordinated actions such as length/width compensation,tool replacement,etc.Make full use of tool life while reducing failure rate.Research the prediction model of CNC tool wear.To enhance the generalization ability of the model,using Dropout's improved deep belief network as a predictive model.First reconstruct the optimized weights in the feature extraction stage,then in the feature matching stage,the label quantity data is introduced for feature recognition.Use support vector regression and neural network algorithms for comparative experimental analysis,the results show that the improved deep belief network algorithm is significantly better than the traditional model in terms of prediction accuracy and stability,the average prediction accuracy is 94.1%.Research the multi-objective optimization model of machine tool operating parameters.Taking the improvement of machine tool processing efficiency as the starting point,a nonlinear multi-constraint decision-making problem with processing efficiency,processing quality,and processing cost as the objective function is established.Use analytic hierarchy process to convert multi-objective function into single objective function and then use genetic algorithm to optimize.The non-dominant sorting genetic algorithm modified by adaptive mutation probability is used to optimize the multi-objective function.The optimization results show that the optimization suggestions of the improved non-dominant sorting genetic algorithm can increase the processing efficiency by an average of 54.6%,the tool cost can be saved by an average of 36.4%,The quality of workpieces can be improved by an average of53.7%.Research on the control system for online tool monitoring and scheduling.Analyze the CNC machine tool management process,and connect to the MES interface to complete the data exchange and back-end information processing of the tool in the manufacturing workshop throughout the life cycle.C/S architecture software is developed based on the C#development environment,combined with Python programming and Modbus communication protocol.