抄録
Wheel bearings are essential components that ensure smooth rotation during both straight-line driving and cornering. A catastrophic failure of wheel bearings can lead to serious vehicle damage or even personal injury. Therefore, wheel bearings must be designed with a high level of reliability and robustness. Due to the increasing demand for quietness with the shift to Electric Vehicles, abnormal noise defects caused by denting have become an important problem in the market. Due to the above reasons, it is important to be able to predict the raceway indentations with precision during the design phase. In this research, developing a surrogate model to predict the actual indentations was completed. In order to predict actual indentations, four types of calculations were performed. First, FEM analysis was performed under specific loading conditions corresponding to actual tests in order to determine raceway loading. Second, FEM analysis was performed to get the numerical indentations values using raceway loading acquired from FEM analysis based on actual tests. Third, a surrogate model was established to predict the numerical indentations acquired from the FEM analysis using the Response Surface Method. Lastly, a surrogate model was established to predict the actual indentations using the Ridge regression. The surrogate model to predict the actual indentations provides the designer with the predicted indentations values with a high degree of precision.
