Research On Tool Wear Monitoring Method For Face Finishing Milling

Development of tool wear monitoring technology not only can improve the reliability and security of the machining process,reducing the risk of machining defects,improving the production efficiency,but also make full use of the tool life,improving the economic benefits.Finish machining as the end of the machining process,machining quality directly depends whether parts qualified.To realize the higher precision,higher efficiency,higher quality processing effect,the importance of realizing the tool wear monitoring of finish machining is self-evident.Based on the disc cutter using for finish machining as the research object,the tool wear monitoring methods of the constant cutting depth finishing milling and the variable cutting depth finishing milling are researched.Firstly,finishing milling process characteristics and the tool wear characteristics is analyzed.The wear quantitative indicator is the sum of width of the wear of the two main cutting areas.And the finish milling simulation experiment and monitoring methods are designed.The current signal and vibration signal of the whole finishing milling tool life are analyzed from the time domain and frequency domain.The analysis result provides the signal selection basis for the tool wear features extraction.Based on the Bayesian classification algorithm,the cutter wear state identification method of the constant cutting depth milling is studied.The current signal and the vibration signal of Y direction are both used to extract the wear features from the time domain and the frequency domain.The distance evaluation technique is applied to select features.The wear characteristic vector consisting of the selected features is used to train the model of tool wear state identification based on Bayesian classification algorithm.And the model is verified.Finally,the tool wear prediction method of variable cutting depth finish milling is built.According to the practical application conditions,the vibration signal is determined as monitoring signal.Combined with the natural frequency,the characteristic band is determined,then the tool wear feature is computed.The cutter wear prediction model is obtained by linear fitting the relationship between the value of tool wear and the tool wear feature.And the predict model is verified,the results show that tool wear prediction average error of the last a set of deep cutting is 5.44%.