Research On Prediction Method Of Milling Cutter Wear State Based On Cutting Force And Vibration Signal

As one of the important terminals of mechanical manufacturing industry,milling cutter wear has always been the focus of research at home and abroad.Relevant data show that the time waste caused by wear failure of milling cutter accounts for more than half of the downtime caused by machine tool failure.It is conservatively estimated that the productivity will be reduced by 10 ～ 60%,which will have an adverse impact on machine tool processing.However,at present,for a series of problems such as whether the milling cutter needs to be changed and under what circumstances,it still depends too much on manual experience.Therefore,this process is greatly disturbed by human factors,which often leads to the waste or over life use of the milling cutter due to the wrong judgment of the operator,which directly affects the machining accuracy and reduces the quality of the final product.In order to reduce the influence of human factors and judge the wear state of milling cutter more accurately,a prediction method of milling cutter wear state based on cutting force and vibration signal is proposed in this paper,which provides some useful ideas for milling cutter intellectualization.The main research contents of this paper are as follows:1.The construction of milling experimental platform and the acquisition of monitoring signals.After comprehensively considering the wear causes,process and wear standards of milling cutter,the milling experimental platform is designed and built to obtain the cutting force and vibration signals reflecting the wear state of milling cutter,and record the wear value VB of the flank of milling cutter.2.Process the collected monitoring signals.The noise contained in the monitoring signal is suppressed by down sampling and wavelet denoising preprocessing;After statistical feature analysis,spectrum feature analysis and wavelet packet feature analysis,the representative frequency band signal rich in feature information is finally obtained.3.The wear state of milling cutter is predicted.Through the optimization of BP neural network by genetic algorithm,a prediction model of milling cutter wear state with high prediction accuracy and small error is established.The traditional BP neural network model and the BP neural network optimized by genetic algorithm are used to predict the wear state of milling cutter at the same time.The comparative test results show that the BP neural network prediction method optimized by genetic algorithm is better than the traditional BP neural network prediction method in accuracy and error,and can effectively improve the prediction accuracy of milling cutter wear state.