Digital Twin-Based Torque Degradation Prediction and Compensation for Hydro-Viscous Drive

Abstract As one of the high power density transmission devices, the Hydro-Viscous Drive exhibits torque transmission and degradation characteristics under impact and fatigue conditions that directly affect the operational stability of equipment. Therefore, accurate prediction of torque transmission behavior and effective suppression of performance degradation are of significant engineering importance for improving system reliability. To address the large prediction deviations of the conventional mechanism model at critical nodes and the inability to characterize torque degradation, the digital twin-based framework for torque prediction and compensation is proposed. The framework employs a mechanism-based torque model to describe the underlying torque transmission mechanisms and integrates a data-driven model to correct prediction deviations at key nodes, thereby forming the hybrid mechanism-data-driven prediction model. In addition, the oil pressure feedback regulation mechanism is introduced to dynamically control the physical hydraulic system, enabling effective mitigation of torque degradation. The results show that, compared with the mechanism model, the hybrid prediction model reduces prediction errors by 33.29% and 40.12% under impact and fatigue conditions, respectively. Furthermore, the effects of cycle count, relative rotational speed, and oil temperature on the torque degradation behavior of the Hydro-Viscous Drive are systematically analyzed. The effectiveness of the oil pressure-based torque compensation strategy in alleviating torque degradation under different influencing factors is also verified, providing a theoretical basis and technical support for torque performance optimization and health management.