Coupling mechanism and data-driven approaches for high power wet clutch torque modeling and analysis

Abstract

We propose a hybrid model that integrates conventional mechanism methods with state-of-the-art data-driven techniques to study torque characteristics during the high power clutch engagement process. Experimental validations using clutch bench tests demonstrate that our hybrid model has an error of 7.0%, outperforming conventional mechanistic models and reducing the average error in predicting clutch torque by 12%. With this developed hybrid model, we systematically investigated the impact of oil pressure, oil temperature, and disc size on torque characteristics under various working conditions. Our findings reveal that pressure has the most significant impact on clutch torque. While the influence of the other two factors on overall clutch torque is relatively diminished, temperature can significantly change the viscous torque and disc size can markedly change torque generation time.