The Research Of Monitoring Technology For Tool Wear Based On HMM

In recent years, with the rapid development of automated manufacturing equipmentand integrated manufacturing model, so that the tool wearing become the weak link ofmanufacturing systems. How to improve the tool life, and to obtain excellentprocessing quality, is the urgent problem need to solve. It can not only save cost, butalso improve the production efficiency. Therefore, the research and development of thetechnology of tool condition monitoring in machining process, has an importantsignificance in improving the level of manufacturing.Modern manufacturing industry is developing toward higher productivity andhigher manufacturing precision. In parts machining if we can grasp the tool cuttingparameters and maximum life, we can replace the tool or blade before the tool enter thefront stage of rapidly wear and the tool reach the largest quantities of workpieces, In thisway we can not only guarantee the machining quality, and can make the tools to achievethe maximum using rate, reduce the frequency of tool change and improve productionefficiency.In this paper, we take the milling process as the research object, and establish theexperimental system of cutter wear monitoring, which taking milling force and thevalue of vibration acceleration as a monitoring signal. We carried out a systematictheoretical and experimental research, which aim at feature extraction of milling forcesignal and vibration acceleration value, as well as the technology of milling cutter wearmonitoring which based on HMM.The state of cutter wearing has a serious impact on the surface quality ofworkpieces and productivity. So the national standards have been specified allowablecutter wear range. However, due to the complexity of the milling process, it is difficultto judge by the milling time to wear of the cutter. Therefore, to achieve real-timemonitoring of cutter wearing on line has good economic and scientific value. In themilling cutter wear monitoring, the milling force signal is considered to be the mostsuitable, because for monitoring of tool wear and breakage, the milling force signal canprovide a clear feature. Through the analysis in time domain and frequency domain for vibrationacceleration measurements and milling force signal in experiments monitoring, from theperspective of time domain and frequency domain, we study the relationship betweenthe milling force signal and their characteristic value in the milling process, found thatthere is the poor correlation between the time-domain statistical characteristics and thecutter wear. And in the frequency domain, there showed a strong correlation betweenmilling force signal power spectral density and the cutter wear, frequency-domainstatistical characteristics show consistent with the trends of cutter wear.