Research on high-precision gear form-grinding technology with consideration of thermal error real-time compensation

Aiming to the problem that thermal error seriously weakens grinding accuracy, according to the involute property and involute helical gear form-grinding principle, this work proposes a novel thermal error predicting model with the combination of steady-state temperature field working condition and non-steady-state temperature field working condition to optimize machining accuracy, eliminate warm-up time, and improve processing efficiency. In view of the above, the independently developed new type of high-precision horizontal forming CNC gear grinder L300G is taken as research object, and the combination of fuzzy clustering algorithm and mutual information theory is taken to optimize temperature variables, which is also verified with the comprehensive gray correlation; on this basis, L300G thermal error predicting model is established based on the combination of autoregressive time series and BP neural network. Then, the model is given practice tests through temperature field finite element analysis, temperature and thermal error measuring, temperature variables optimizing, thermal error predicting, and involute helical gear form-grinding. Results show that the model is fairly effective in thermal error predicting; what is more, after thermal error real-time compensation, it can eliminate warm-up time of gear grinder about 6000 s, and the Gleason measuring results show that the involute helical gear form-grinding tooth profile accuracy is improved from level 6 to 4, which verifies the correctness. So that it is expected to be widely used in high-precision form-grinding of horizontal gear grinder; besides, it also has significance for high-precision form-grinding research of vertical gear grinder.