Dynamic features and sensitivity analysis of an electric vehicle powertrain in the presence of stochastic temperature uncertainties

Abstract

Understanding the complex interplay between random temperature fluctuations and the vibration dynamics of electric vehicle (EV) powertrains is crucial for enhancing performance, durability, and noise control in modern automotive systems. This study investigates the influence of random temperature variations on the dynamic behavior of EV powertrains. A typical EV configuration is employed, integrating a two-stage helical gear system with a differential mechanism. The mesh stiffness of helical and straight bevel gears is modeled using the potential energy method, slice theory, and superposition principles, with offset concepts incorporated. Flash and bulk temperatures, along with the mechanism’s geometric and material properties, are treated as random variables. The dynamic and sensitive analyses of the mechanism in the presence of various random temperatures are conducted. To address the resulting random differential equations, the sparse grid method is combined with the polynomial chaos expansion technique. Four sparse grid configurations are evaluated to optimize computational efficiency while maintaining accuracy. The results have been validated against Monte Carlo and Latin hypercube sampling methods. Key findings reveal that temperature variations affect the system’s dynamic characteristics by altering its natural frequencies and increasing the vibration in the mechanism. Additionally, the sparse grid approach, particularly when applied within the polynomial chaos framework, proves effective in handling EV powertrains with multiple random parameters. While the Smolyak grid delivers the highest accuracy, the hyperbolic cross set offers a viable alternative with reduced computational demands, particularly when assessing whether vibration magnitudes fall within tolerable limits.