Space-Time Kriging Surrogate Model to Consider Uncertainty of Time Interval of Torque Curve for Electric Power Steering Motor

Abstract

In rotating motors, an electromagnetic analysis using finite-element analysis is performed according to the time interval such as the steps of the rotation angle. If the uncertainty of time variable is necessary to be handled, the time interval should become smaller than uncertainty values. It causes an increase of the cost even if the surrogate model is applied. Thus, space-time kriging surrogate model is proposed to consider time variable as well as space variable in the design stage. The proposed model handles time variable and predicts responses between time intervals to obtain outputs for the tiny uncertainty of time variable. The performance of prediction is verified with simulation data and then the reliability estimated by the proposed model is compared with measured data from manufactured prototypes.