Prediction of Machine Tool’s Accuracy Degradation Based on Chaotic Phase Space Reconstruction and Depth GRU

Abstract

Aiming at the problem that it is difficult to analyze the precision deterioration law of CNC machine tools by mathematical modeling, a prediction and evolution method of CNC machine tool precision degradation based on chaos evolution and deep Gated Recurrent Unit (GRU) network was proposed. The method of phase space reconstruction was proposed to reconstruct the one-dimensional historical data sequence of motion accuracy of CNC machine tools. The evolution process of motion accuracy of CNC machine tools was described by the trajectory of phase points in multi-dimensional phase space. Furthermore, the prediction model based on gating cycle network was established. By learning the temporal and spatial characteristics of the phase point trajectory, the motion accuracy degradation process was evolved and predicted by using the phase point trajectory prediction curve. The experimental results show that the prediction model based on chaotic evolution and deep GRU network can well track the degradation trend and law of motion accuracy of CNC machine tools, and has high prediction and evolution accuracy.