Research On Tool Wear Detection Method For Turning Large Pitch Thread

As a non-standard part,the large-pitch threaded part has a long cutting stroke,a long cutting edge and a large overhang of the cutting tool,which leads to a large amplitude and obvious fluctuation of the vibration and cutting force of the cutting tool and rapid wear of the cutting tool,thus affecting the processing of long-stroke and high-quality large-pitch threaded parts.Therefore,it is important to monitor tool wear to improve the productivity and quality of large pitch threaded parts.In this paper,we propose a tool wear monitoring method for turning large pitch threaded parts and design a tool wear monitoring system that can be applied to turning large pitch threaded parts.The main research contents of this paper are as follows.Firstly,the tool wear process of turning large pitch threaded parts is studied.And the tool wear in this test mainly showed hard point wear and bonded wear;the tool wear curve was drawn and the tool wear stages were divided according to the curvature of the curve;the tool wear criteria were clarified according to the roughness of the machined surface;and the basis for the subsequent construction of the tool wear model was provided.Secondly,the monitoring signals and tool wear during the cutting process were analyzed and processed.The signals were analyzed and relevant eigenvalues were extracted using time domain and frequency domain methods,and the eigenvalues were extracted after noise reduction of the signals using the integrated empirical modal method of time-frequency analysis;then the correlation between the extracted eigenvalues and tool wear was screened using the Relief-F feature selection method to determine the input quantity of the model to be built later.Thirdly,the tool wear model was established based on the deep learning method.A BP neural network was used to build a tool wear prediction model;a multi-grain scanning cascade forest method was used to build a tool wear stage identification model respectively.The accuracy of the tool wear prediction model and tool wear stage identification model were 90.93% and 92.06%,respectively,using the empirical formula method or the repetitive mapping method.Finally,the tool wear monitoring system for turning large pitch threads was developed and validated.The signal analysis processing and tool wear model were written into the system to make it a whole;a tool wear experiment for turning large pitch internal threads was designed and the tool wear monitoring analysis method for turning external threads was applied to this experiment.In the tool wear monitoring for internal threads,the accuracy of the tool wear prediction model was91.27% and the accuracy of the tool wear stage identification model was 93.56%,which verified the the effectiveness of this monitoring method and the reliability of the system.