Research on Thermal Error Compensation Technology of Grinding Machine Based on Neural Networks

摘要

In order to fulfill the thermal error modeling of precision abrasive machining, a neural networks model is presented through analyzing thermal error sources of the grinding machine, the structure and algorithm of the neural networks is expatiated then; And because of reasonable sample and systematical training, the accuracy of thermal error models is improved. The hardware system for thermal error compensation is proposed finally, and an experiment is accomplished on the grinding machine. The result shows thermal errors is reduced from 6μm to 1 μm.