Research Of Monitoring System About Tool Wear Based On Multisensory Information Fusion

The tool plays a very important role in machining, whose machining precision depends largely on rate of tool wear, and is the important factor of restricting automatization and unmanned machining process. During the machining process, it’s inevitable that tool will wear, and it’s necessary to carry out on-line monitoring for the degree of tool wear.The article finishes the research of on-line monitoring system about tool wear multisensory information fusion, and achieves the on-ling monitoring of tool wearing intensitivity. Firstly, by examining sensor and signal input devices, AE signal and vibration signal’s acquisition system has been put up; conducting the experiment for sensor’s installation site, and gains the best installation site for AE sensor and vibrating sensor. Secondly, based on the experiment principle of simpleness and effectiveness, Orthogonal Experiment is employed and analyze all factors that will influence signals, and finalize4main factors as following: wear rate, spindle speed, feed speed and depth of cut. And then Orthogonal Experiment table of four factors and four levels has been designed, and based on it, the experiment of “the influence of factors on date” and “tool process for different wear loss” have been carried on and obtain the required experiment date. Thirdly, Analyze the data in time domain and frequency domain and time-frequency domain, through wavelet packet decomposition and Orthogonal Experiment methods, the best signal feature has been found: the feature of vibration signal is in frequency range of （1.95KHz~3.9KHz）and （5.85KHz~7.8KHz）, and AE in frequency range of （125KHz~156.25KHz） and spindle speed; Lastly, By means of BP neural network, information fusion system based on feature fusion have been established, and achieved one-line defect for tool-wear loss.The article positively research and explore sensor technology, information acquisition techniques, wavelet packet decomposition and information fusion technology based on neural network, which makes a better understanding, especially the choosing of signal feature. It’s the first time that choose feature by means of Orthogonal Experiment, and obtain the abrasion’s influence on signal and analyze other factors’ influence on signal, finally find the feature that properly reflect tool’s wear degree that is not easy to be influenced by other factors, then improve the accuracy of fusion.