Research On Optimization Methods Of Kernel Parameters Of Tool Condition Intelligent Monitoring

The goal of the tool condition intelligent monitoring system is to be able to timely and accurately monitor abnormal conditions such as wear and damage during tool processing and to make prompts and alarms,in order to improve production efficiency and reduce production costs.But the current level of technical automation and identification is not high enough,and the adaptation to the process and working conditions is not strong enough.Therefore,based on the investigation and summary of the influencing factors in various aspects of the system application process,this paper studies the optimization methods of the three core parameters of window length step,tool wear monitoring threshold and tool breakage monitoring feature value.The main research contents are as follows:Firstly,the effects of process conditions and signal acquisition conditions on the length and step length of the pretreatment window were qualitatively analyzed.The influence of the change of window length and step size on the signal characteristics was quantitatively investigated.Based on this,the automatic setting method of two parameters of window length and step length in the signal preprocessing was established.The characteristics of time and frequency domain and application conditions of common wear characteristic values are analyzed and summarized.According to the abnormal phenomena appearing in the calculation process of wear eigenvalues,the causes of the occurrence and the influence of different process conditions on the judgment of wear state are analyzed.Then,treatment plan based on the difference between numerical and quantitative data points in the numerical data and the normal data is proposed.Because the wear threshold value is used to determine the wear state of the tool,the accuracy of the tool wear state and the adaptability is poor.Combined with the statistical characteristics of the tool wear characteristic value,the dynamic setting method of the tool wear threshold is proposed,effectively reducing the impact of factors such as changes in the blank.The principle of the tool damage,the influencing factors and the characteristics under different process conditions are analyzed and summarized.The time and frequency domain characteristics of the tool damage signal are explored.It is found that the dynamic envelope method has insufficient accuracy for the weak phenomenon of the tool breakage.A tool breakage monitoring method based on signal waveform characteristics is proposed.This method uses the wave difference between the broken signal and the normal signal as the monitoring feature value,which can significantly improve the accuracy of damage monitoring.Finally,using the tool state monitoring platform built by the industrial site to collect data,the preprocessing window length step,the tool wear monitoring dynamic threshold and the tool waveform monitoring feature value based on the signal waveform characteristics are optimized.The use of these methods effectively improve the accuracy and adaptability of tool status monitoring,and significantly enhance the intelligence of the technology.