A hybrid physics-data-driven surface roughness prediction model for ultra-precision machining

The surface finish quality is critical to the service performance of a machined part, and single-point diamond ultra-precision machining can achieve excellent surface quality for many engineering materials. This study studied the problem of predicting the surface roughness for titanium alloy workpieces in ultra-precision machining. Process data and surface roughness measurement results were obtained during end-face machining experiments. A deep learning neural network model was built based on the ResNet-50 architecture to predict surface roughness. We propose increasing prediction accuracy by using the energy ratio difference (ERD) as a stability feature that can be extracted using fast iterative variational mode decomposition (FI-VMD). The roughness value obtained with an analytic model was also used as an input feature of the prediction model. The prediction accuracy of the proposed approach was depicted to be improved by 8.7% with the two newly introduced roughness predictors. The influence of the tool parameters on the prediction accuracy was investigated, and the proposed hybrid-driven model exhibited higher robustness to errors of the tool parameters than the analytic roughness model.