Wear Condition Monitoring Method Of Milling Cutter Based On Information Fusion

With the improvement of precision and reliability of modern manufacturing equipment,machine tools and fixtures have little effect on the machining quality of workpieces,and tool wear state is crucial to the machining quality and production efficiency.Titanium alloy is one of the most widely used metal materials in aviation and aerospace,but the low thermal conductivity and high chemical reaction of titanium alloy lead to serious tool wear,which directly affects the machining quality of parts.In order to quickly and accurately identify the tool wear state,improve the processing quality of parts and ensure the safety of related equipment is of great significance.First of all,through the analysis of tool wear process and bl unt standard,the initial wear,normal wear and severe wear of three kinds of tool wear state are determined.By analyzing the influence of signal processing on tool wear state,cutting force signal,vibration signal and acoustic emission signal are select ed as the characteristic sources of tool wear state monitoring,and a multi-sensor signal acquisition test platform is built.Secondly,in view of the problem that the original data of the sensor can not directly reflect the change of the tool wear state in the machining process,the signal processing technology is used to preprocess the original signal.The time domain and time-frequency domain analysis methods are used to extract the multisignal feature and select the feature index which is closely relat ed to the tool wear.Then,in view of the limitations of the traditional machine learning recognition model in the monitoring of tool wear state,a stacked sparse self-coding network(Stacked Sparse Auto Encoder,SSAE)based on multi-information fusion is proposed for the recognition of milling cutter wear state,and the recognition results of milling cutter wear state are compared with the single-signal feature SSAE model and the feature fusion SSAE model of two types of signals.At the same time,under the premise of ensuring multi-signal feature fusion,different machine learning models are established to compare and analyze the recognition results of milling cutter wear state,which proves the superiority of the proposed method in recognition accuracy.Finally,an online monitoring system for milling cutter wear state was developed based on Lab VIEW.The monitoring system can realize real-time signal acquisition,storage,historical data reading,signal analysis and online recognition of milling cutter wear state.The experimental verification shows that although the online monitoring system of milling cutter wear state does not identify the tool wear state correctly sometimes,the overall mean square error of online identification is within the range of 6.12?23.33,and the expected goal of online identification of tool wear state is preliminarily realized.