Wear Monitoring Method Research Of Milling Tool Based On Current And Vibration Signal

The fault of CNC machine tool is mainly caused by the wear-out failure of the cutting tool in the machining process. Inaccurate and delayed fault condition monitoring would decrease the machining efficiency and shorten the tools working life, then the processing cost will be increased. Therefore, the on-line fault condition monitoring of CNC machine tool is beneficial for prolonging trouble-free worktime and reducing the probability of machine workpiece scrapped.The research of this paper is based on milling cutter of CNC machine tools. After analyzing the current domestic and foreign research situation, especially making research on three-dimensional dynamometer. Since it requires change the machine tool structure in installation and its measure precision is influence by machine tool vibration greatly, it is difficult to be applied n in the industrial production. The fault condition monitor methods of cutter wear based on current signal and vibration signal is proposed in this paper. The main works of this method are as following :(1).After analyzing the wear-out reason of tool and making research on the wear-out process of tool wear, the tool flank wear erosion rate is utilized as a standard of measuring milling cutter wear.(2).The experiment platform of milling cutter wear based on LabVIEW is set up, then, the location of sensors are confirmed after analyzing the milling experiment condition. The communication between the acquisition device and the sensors is realized by setting correct parameter of the NI-DAQ and the data acquisition card.(3).The dynamic model for the main shaft transmission system is established. The characteristic frequency of milling force is extracted after the correlation analysis of two-phase spindle current signal power spectral density. The analysis result confirm that current signal can be utilized as the replacement of milling force in on-time monitoring The milling parameter is confirmed by the method of single factor analysis.(4) The current signal and vibration signal of CNC milling machine are collected, then they are utilized to extract the signal characteristic of time domain, frequency domain and time-frequency domain. Then, in order to gain the best milling cutter wear characteristics, kernel principal component analysis is utilized to reduce the original characteristics.(5).After making research on fault condition monitor methods of cutter wear based on BP neural network and support vector machine(SVM), the training of them are realized. The result illustrates that fault recognition rate of BP neural network is 83.3% and fault recognition rate of support vector machine is 91.7%. The fault recognition rate of both are extremely high. Thus, the fault condition monitor methods of cutter wear based on BP neural network and support vector machine(SVM) are useful.The research in this paper show that the precision rate of tool wear’s monitoring method based on current and vibration is extremely high. The method is simple. It is easy to realized and it can be utilized for tool wear condition monitoring effectively.