A novel prediction approach of polymer gear contact fatigue based on a WGAN-XGBoost model

Polymer gears have long been used on power transmissions with the fundamental durability data, including fatigue S-N curves, yielding important data informing reliable and compact designs. This paper proposed a prediction method for polyformaldehyde (POM) gear fatigue life based on the innovative WGAN-XGBoost algorithm. The findings generated herein revealed that the proposed method performs well in terms of prediction accuracy. The predicted fatigue lives, analyzed under different loading conditions, were within 1.5 times dispersion band compared with experimental results. Furthermore, based upon the enhanced fatigue dataset, a thermo-mechanical coupled prediction formula for POM gear contact fatigue life was proposed. These findings offered a new approach for high-power density design of polymer gears.