Abstract: In order to eliminate the influence of thermal error on machining precision，a novel method combining the data processing merit of grey model (GM) with that of artificial neural network (ANN) is proposed for thermal error modeling in machine tools.bining gtion Relevant temperature and thermal error data of a turning machine are measured respectively through digital temperature sensors and non-contact displacement sensors. A series of experimental data are used to establish the thermal error compensation model by means of adjusting the length of data sequence in GM and the weight and threshold value in ANN. The results show that the new model performs better than the traditional GM and ANN model. It is less-requiring for the original data, convenient in calculation, good in fitting and precise in prediction under a variety of working conditions. So the new model is more suitable for the complex industrial applications.

Key words: Artificial neural network, Error compensation, Grey model, Machine tool, Thermal error