Research on time-varying dynamic characteristics of heavy-duty vehicle powertrain system based on digital twin

Abstract

Under harsh operating conditions, the powertrain system of heavy-duty vehicles is subjected to severe impact loads. Restricted by structural space, internal dynamic loads cannot be measured directly, making accurate health status assessment difficult to realize. Therefore, this paper constructs a digital twin (DT) system, which integrates a dynamic model, a shift frequency model, and a Dynamic Graph-based Gated Recurrent Unit (DG-GRU) prediction model with dynamic gating regulation. Through simulation mapping, the system realizes real-time collection and visual monitoring of indirectly measurable dynamic signals. Bench test verification demonstrates that the DT simulation signals of output shaft speed and torque are in high agreement with experimental data, with steady-state errors below 5%. The research reveals the degradation law of coefficient of friction (COF) of different clutches with the accumulation of shift cycles, among which the C1 and C2 clutches suffer the most severe wear. The evolution law of drive shaft torque performance with clutch degradation is investigated, and an adaptive oil pressure compensation mechanism is proposed to offset the performance loss of components caused by friction degradation. The research results indicate that the proposed mechanism can effectively reduce the adverse effects of clutch wear on drive shaft performance. This DT system can accurately predict the dynamic characteristics throughout the full life cycle and provides a theoretical basis for predictive maintenance of heavy-duty vehicles.