关键词: 热误差预测, 运动精度, 混沌演化, 精度劣化, 深度学习

Abstract: Thermal error is one of the main factors leading to the decline of NC machine tool's accuracy. It is proposed to use the chaotic characteristics of the thermal error to reveal the internal law of the temperature rise process hidden in the disordered and complex representation, so as to predict the early deterioration of the motion accuracy of the machine tool. The chaotic phase space was reconstructed from the historical data of machine tool temperature measuring points. The Lyapunov exponent was used to prove that the temperature rise process of NC machine tool was actually a complex nonlinear system evolution motion with chaotic characteristics. The system was identified from multi-dimensional space and perspective, and the law of thermal error contained in the system was excavated. A thermal error prediction model based on chaotic phase space evolution and long short term memory neural (CPSE-LSTM)was proposed. The reconstructed temperature series was used as the input of the prediction model. The temporal and spatial characteristics of dynamic chaotic time series were extracted to improve the accuracy and generalization ability of machine tool thermal error prediction model under different conditions. The circular motion repeated positioning error in the temperature rise process was defined. According to the mapping relationship between the thermal error of the machine tool spindle and the circular motion accuracy, the motion error of the machine tool was evaluated, and the decline of the motion accuracy of the NC machine tool was predicted in early stage. The experimental results show that CPSE-LSTM model has high prediction accuracy and generalization ability under different conditions, and the evaluated value of machine tool motion accuracy is highly consistent with the measured value.

Key words: thermal error prediction, motion accuracy, chaotic evolution, accuracy deterioration, deep learning