The Neural-fuzzy Thermal Error Compensation Controller on CNC Machining Center.

Abstract

The geometric errors and structural thermal deformation are factors that influence the machining accuracy of computer numerical control (CNC) machining center. Therefore, researchers pay attention to thermal error compensation technologies on CNC machine tools. Some real-time error compensation techniques have been successfully demonstrated in both laboratories and industrial sites. The compensation results still need to be enhanced. In this research, the neural-fuzzy theory was conducted to derive a thermal prediction model. An IC-type thermometer was used to detect the heat sources temperature variation. The thermal drifts were online measured by a touch-triggered probe with a standard bar. A thermal prediction model was then derived by neural-fuzzy theory based on the temperature variation and thermal drifts. A graphic user interface system is also built to conduct the user friendly operation interface with Insprise C++ Builder. The experimental results show that the thermal prediction model developed by neural-fuzzy theory methodology can improve machining accuracy from 80 mu m to 3 mu m. Compared with the multivariable linear regression analysis the compensation accuracy is increased from +/-10 mu m to +/-3 mu m.