Spindle thermal error modeling method considering the operating condition based on Long Short-Term Memory

Thermal error has a significant effect on machine tools' precision. Traditional methods usually establish the thermal model by temperature. As temperature only describes instant state of tool, it is inadequate to clarify generation process error. To improve description completeness process, operating condition data such as spindle speed, rotational number and power are employed besides Furthermore, Long Short-Term Memory (LSTM) based was proposed which combined with This not considers dynamic characteristics error, but also selects remembers previous information. accuracy for whole speed range, K-Means clustering analysis adopted obtain optimal categorization, then different LSTM sub-models were established separately. Experiments demonstrated that predicting segmental method (K-S-LSTM) over 85.0% under conditions. Moreover, comparison experiments indicated performance K-S-LSTM better than Elman IRNN models. Finally, compensation experiment showed can decrease from 49 μm 12 after in actual numerical control system.