Robust Thermal Error Modeling and Compensation for CNC Machine Tools.

Thermal errors are one of the most significant factors affecting machine tool accuracy. Error compensation has been widely used to reduce the thermal errors, the robustness of the thermal error models, however, still needs to be improvement. Currently, five-axis machine tools are becoming more important and extensively utilized in industry. In this regard, the geometric errors of rotary axis must be properly measured and corrected to assure the accuracy of five-axis machining. Thermal error model, relating temperature variations to thermal errors, is the core of an effective thermal error compensation strategy. Thermal modal analysis, unveiling the essence of thermo-elastic process, is explored for the determination of temperature sensor placement based on the finite element analysis and eigen analysis. Thermal error models are thus derived based on the temperature variations collected from the specified temperature sensors. The robustness of the derived models is investigated in terms of linear extrapolation and frequency sensitivity. Numerical simulation and experiments are conducted to illustrate the existence of thermal modes and validate the robustness of the thermal error models. Thermal loop analysis is developed for the thermal error compensation of an entire machine tool. A machine tool is first decomposed into several thermal links along an identified thermal loop. For each thermal link, the thermal modal analysis is carried out for the derivation of thermal error model. These thermal links are finally reassembled for the thermal error prediction of the entire machine tool. The thermal loop analysis mitigates the inaccurate modeling of machine joints, and extensively facilitates the utilization of the finite element method in the thermal error modeling and compensation. Calibration of rotary axis of five-axis machine tools is usually time-consuming and laborious by using laser interferometer or autocollimator systems. The Telescopic Magnetic Ball Bar is explored to estimate error components induced by the rotational motion of a rotary axis. The calibration algorithm is developed based on the rigorous mathematical derivation. The setup errors, including parameter variation and eccentricity, have been accounted for through the numerical simulation, enabling the practical utilization of this method. This approach shows the advantages of easy setup and quick assessment.Ph.D.Mechanical EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/60857/1/zhujie_1.pd