Research And Application Of Wear Prediction Method Of NC Milling Cutter Based On Data-driven

In the application of IIo T,one of the most important application directions is fault prediction and health management of industrial equipment.PHM uses the monitoring information collected by sensors in industrial production to monitor and predict the health status of industrial equipment with the help of information technology and artificial intelligence methods.It ensures that corresponding measures are taken before the system fails,and gives a series of supporting decisions under the existing resource conditions.As the basic tool of industrial equipment,the wear of CNC machine tools will directly affect the quality of product processing,and even cause workpiece scrap and machine tool damage,thus affecting the production efficiency of the whole project.Therefore,how to use the condition monitoring information in the working process of the tool to accurately predict the wear degree and residual life of the tool,which has high application value and practical significance in machine tool processing and production.Under the application background of industrial big data and industrial equipment fault prediction and health management,this topic uses the algorithm model of artificial intelligence to predict the wear of milling cutter of NC machine tool,and then accurately predicts the remaining life of cutter.The main contents in the research and analysis process of this paper are as follows:(1)firstly,data preprocessing is carried out for the original monitoring signal of milling cutter,including truncation of invalid data and elimination of abnormal data.(2)Then,with the help of statistics and signal theory knowledge,multi-dimensional feature extraction is carried out for cutting force signal,vibration signal and acoustic emission signal in time domain,frequency domain and time-frequency domain,and then feature screening is carried out for the extracted feature set.In this process,a weighted feature screening evaluation index function is proposed.(3)In the process of designing tool wear prediction algorithm,this paper proposes a random forest regression algorithm based on weighted feature screening evaluation index function,which can accurately predict the relative remaining life after inputting the monitoring signal of machine tool milling cutter.(4)In order to realize the modularization and visualization of model training function and milling cutter life prediction function,finally,the milling cutter life prediction system of NC machine tool is designed and implemented.Various comparative experiments in this paper show that the prediction algorithm proposed in this paper has higher and more stable prediction accuracy than the algorithm model method in other articles.At the same time,the design of the milling cutter life prediction system of NC machine tool conforms to the normative principle of system design,which is simpler and more intelligent than other tool life monitoring systems.