Thermal error modeling by integrating GWO and ANFIS algorithms for the gear hobbing machine

Abstract

Improving the machining accuracy of the gear hobbing machine is effective to enhance the processing efficiency of gear production. Thermal error is the main source of the machining error of the gear hobbing machine, and establishing accurate thermal error prediction model is necessary to reduce the thermal error of gear hobbing machine. In this paper, an accurate thermal error model was proposed based on the improved gray wolf optimizer (IGWO) and adaptive neuro-fuzzy inference system (ANFIS). What’s more, the K-means clustering and gray model GM (0, N) were used to conduct the selection of temperature-sensitivity points to guarantee the robustness and accuracy of the prediction model. ANFIS was used to predict the thermal error on different working condition based on the data collected in the thermal characteristic experiments. IGWO was applied to optimize the parameters of ANFIS. And the combined IGWO-ANFIS was applied to predict the thermal expansion of the hobbing shaft of the gear hobbing machine. By comparing the prediction performance of IGWO-ANFIS and other existing algorithms, it can be concluded that it has better generalization performance than other algorithms.