On-line prediction of ultrasonic elliptical vibration cutting surface roughness of tungsten heavy alloy based on deep learning

The surface quality of tungsten heavy alloy parts has an important influence on its service performance. The accurate on-line prediction of surface roughness in ultra-precision cutting of tungsten heavy alloy has always been the difficulty of research. In this paper, the ultrasonic elliptical vibration cutting technology is used for ultra-precision machining of tungsten heavy alloy. Based on the idea of deep learning, the surface roughness is discretized, and the fitting problem in surface roughness is transformed into a classification problem. The generalization ability of the prediction model is improved by introducing batch standardization and Dropout. The relationship between the vibration signal and the surface roughness is established. Experimental results show that the model can achieve on-line prediction of cutting surface roughness. The prediction accuracy rate can be improved by more than 10% compared with the direct fitting method.