Improved unscented Kalman filter algorithm-based rapid identification of thermal errors of machine tool spindle for shortening thermal equilibrium time

Abstract

Thermal equilibrium of machine tools is needed before machining to ensure the accuracy. To shorten the warm-up time and improve the machining efficiency, a rapid identification of thermal errors is proposed. It can predict the thermal errors in the shortest identification time so that the thermal error compensation can be started in advance to shorten the thermal equilibrium time. A Singular Value Decomposition-Adaptive Unscented Kalman Filter algorithm (SVD-AUKF) is proposed with singular value decomposition and adaptive rule for the fast identification. A general exponential thermal error model is established and applied into the SVD-AUKF. RMSE of the thermal error data within each time is calculated for the evaluation to obtain the shortest identification time. K hysteresis rule, Δt judgment rule and maximum time rule are proposed for the prediction accuracy. Different experiments are proposed with two machine tools to testify the effectiveness of the shortest identification time.