High-fidelity modeling of dog clutch and its control and design studies based on statistical analysis

Abstract

Compared to friction clutches, which were mainly used in vehicle transmission, dog clutches can provide better transfer efficiency and constitute a more compact transmission. Besides, speed synchronization of dog clutches is a simple problem for electrified vehicles because motors can derive positive or negative torque freely. Therefore, dog clutches are a good substitute for friction clutches in electrified vehicles. The only problem of a dog clutch is the “torque hole” phenomenon during the shift, which is a disconnection of power during the synchronization of speed and engagement, and it critically harms the comfort of the driver. However, there is no sufficient precedent research about fundamental dynamics and control strategy of dog clutch. In this research, high-fidelity modeling of dog clutch is introduced, and every collision during the engagement is considered in modeling. A spring-damper model with deflection depth models contact force, and this depth is determined by the current state of engagement, which is determined by axial displacement and the relative angle between dogs at that time. With high-fidelity modeling, statistical analysis for proper engagement conditions based on various simulations is studied to guarantee better shift performance. Then, these proper engagement conditions are used as criteria of the proposed control algorithm. The proposed modeling and control algorithm is expected to be the theoretical foundation of dog clutch which is an essential component in electrified vehicles.