Research On Tool Wear Status Monitoring Technology Based On Deep Learning

The tool is one of the most important processing elements in mechanical production.Its wear not only affects the dimensional accuracy and the surface quality of the workpiece,but also causes the vibration shock of the machine tool to affect the accuracy of the machine tool,and also indirectly affects the machining efficiency and production cost.Therefore,the monitoring of tool wear status has become more and more important,and has become one of the hot issues researched by domestic and foreign scholars.The wear problem of the milling tool was studied in this paper.The machined workpiece surface texture was obtained via by a visual sensor.According to the characteristics of the tool wear image,to obtain the wear level of the tool,the convolutional neural network was used to directly extract the characteristic information of the sample.The experimental results show that the convolutional neural network can effectively reduce the amount of pre-processing work and has a high network recognition rate.The main research contents of the thesis are as follows:Firstly,experiments on the machining of hard aluminum alloy with white steel end mills were performed to analyze the tool wear mechanism from the theoretical views in this paper.As the tool wear increased,the workpiece surface texture would become worse and worse.Based on this point,an image processing method based on texture analysis was put forward.Secondly,an improved convolutional neural network algorithm was proposed to overcome the shortcomings of the traditional pattern recognition.According to the characteristics of the tool wear,the parameters of the convolutional neural network and the training algorithm were optimized.The superiority of the proposed algorithm was verified by the comparison with other algorithmsThirdly,aiming at the problems of the existing tool wear detection instrument such as large volume,high cost and off-machine detection,and combining with the image acquisition method and the characteristics of the proposed algorithm,an embedded tool wear detection system was designed based on FPGA and DSP,which has the advantages of integration and intelligence.Finally,to verify the feasibility of the designed tool wear detection system and the effectiveness of the proposed algorithm,some experiments were carried out.The results showed that the wear detection of the tool could be accomplished by using the designed system and the wear degree of the tool with high accuracy could be reported by using the proposed improved algorithm.