Prediction of ultimate bearing capacity for EV differential housings using elastoplastic finite element analysis

Abstract

Accurately predicting the ultimate bearing capacity of differential housings is crucial for the reliable and lightweight design of electric vehicle (EV) transmission systems. This study proposes a novel engineering approach integrating elastoplastic finite element analysis (FEA) with a torque-angle curve extrapolation technique to predict the ultimate bearing capacity of EV differential housings. The method involves simulating static torsional loading at 12 circumferential positions of the driven gear, generating torque-angle curves for each. A zero-slope criterion, identified as the transition point from strain hardening to structural failure, is applied to determine the limit torque for each curve. A quadratic polynomial extrapolation based on the least squares method is employed when the zero-slope point is not directly reached in the simulation. The minimum extrapolated torque among all positions is defined as the predicted ultimate bearing capacity. The predicted ultimate torque is 1436 Nm, which is in good agreement with the experimental bench test result of 1400 Nm, with a relative deviation of approximately 2.6%. Moreover, the predicted capacity exceeds 2.5 times the maximum design torque required by the automotive industry standard (QC/T 1022-2015), confirming that the housing satisfies static strength criteria. In contrast, conventional linear elastic FEA combined with FKM guidelines yields conservative safety factors below 1.0, incorrectly suggesting insufficient strength. The proposed elastoplastic method reduces design redundancy and provides a more accurate basis for lightweight optimization. This work offers a reliable predictive framework for the ultimate strength assessment of differential housings and similar load-bearing components in automotive transmission systems.