Abstract: When worktable is traveling along the X, Y, or Z axis of machine tool back and forth, due to frictional interaction between moving parts, a great amount of heat is generated in ball–screw drive systems, and part of it is transferred to the worktable, by which two sides of the worktable parallel to the guide way will be warped. So the thermal deformations of various positions on the worktable are different. In order to improve the accuracy of thermal error compensation of large size worktable, a multi-location forecasting model with piecewise fitting compensation method is proposed. The corresponding thermal error model of the multi-point on the worktable is established, and piecewise fitting of multi-point prediction model is used to implement a thermal error compensation for whole worktable. Furthermore, to improve prediction accuracy and robustness of the thermal error model, particle swarm optimization algorithm is utilized to identify the model parameters based on the real-time feedback data, and thus the thermal error model could adapt to the latest working condition of the machine tools. The thermal error compensation experiment of the worktable in a three coordinate milling machine is carried out under the condition of X-axis’s fast movement. The experimental results indicate that this method has a good robustness and a high precision, which can compensate thermal error of large size worktable at any position, and this method is of universal applicability.

Key words: piecewise fitting, real-time identification, thermal error, worktable, CNC machine tools