Coupling mechanism and data-driven approaches for wet clutch frictional heat modeling and analysis

This study proposed a hybrid model that integrates mechanism-based and data-driven approaches to investigate temperature characteristics during the sliding friction process in wet clutches. The model incorporates friction coefficient variations induced by operating conditions and interface temperature, as well as complex heat exchange and heat partition processes at the contact interface, to solve for the actual dynamic variation surface of the heat partition coefficient. UMT (Universal Micro-Tribotester) experiments and sliding friction temperature-measurement experiments are conducted to validate the model, achieving a 3.59 % friction coefficient prediction error and a temperature prediction error not exceeding 12.18 %. Three distinct temporal stages of the heat partition coefficient are identified: inertial, time-varying, and steady-state stages. In the steady-state stage, the partition coefficient is influenced by radial heat dissipation and the friction pair thickness ratio. Specifically, the steady-state value increases with the friction disc-to-separator disc thickness ratio, reflecting its dependency on relative thermal capacity. This work provides a novel framework for modeling dynamic thermal behavior in wet clutches, enabling more accurate temperature field calculation and guiding thermal management design.