Research On Tool Condition Monitoring In Milling Process Based On Multi-sensor Information Fusion

Cutting tools are the direct participants in material removal in machining,and wear or breakage of tools will occur as the machining continues.Excessive wear or breakage of tools directly affect the quality of the machined parts and increase the manufacturing cost.Therefore,the on-line monitoring of tool status is very important for improving the quality of products,saving manufacturing costs and improving manufacturing efficiency.Meanwhile,It is very important for manufacturing industry to become intelligent and dehumanized.The research done in this paper is as follows:Firstly,the machining signals under different machining conditions and tool wear states are obtained through full factor experiments.The feature extraction method and feature selection method for wear and fracture monitoring signals of milling cutters are studied.Time domain analysis,power spectrum analysis and wavelet packet decomposition method are used to analyze the time domain analysis,frequency domain analysis and time-frequency domain analysis of monitoring signals.Then the characteristics of the signal in time domain and time frequency domain are selected to obtain the characteristics of the wear and fracture state of the milling cutter.Secondly,Multi-sensor information fusion technology based on BP neural network is applied to wear condition monitoring of milling cutter.lt studied wear state of the tool in the machining of fixed and variable parameters.During monitoring process of tool wear,it gives the fixed parameter monitoring method based on threshold and variable parameter monitoring method based on wavelet packet decomposition and BP neural network multi-sensor information fusion.The characteristic values of signals collected by various sensors were extracted by wavelet packet decomposition and correlation coefficient method,and the tool wear VB values are measured by tool microscope.The training samples are used to train the BP neural network identification system,and the test samples are used to verify the network.Then,The eigenvalues are used as inputs of BP neural network to get tool wear values.At the same time,a variable parameter integrated network based on BP neural network with the same cutting cross section area is proposed.The combination of different machining parameters with the same cutting area is classified into the same sub network for training and monitoring,which reduces the number of subnets and improves the fault tolerance of the system.Finally,this paper uses LAB VIEW programming language and MATLAB data processing software to develop system for monitoring the milling tool state,and obtains the software system including waveform display,data transmission,data storage and cutter alarm,real-time monitoring of milling cutter wear value,etc..The experimental results show that the monitoring system technology can effectively identify the tool wear state,and get the corresponding wear value,and give the tool fracture reminder.The cutting experiment design,signal acquisition,sensor signal analysis,feature extraction and tool state pattern recognition are explored in this paper.The tool condition monitoring technology is developed,and the practical model of tool condition monitoring is put forward,which improves the precision and stability of tool online detection system.