Control-oriented modeling of wet-clutch friction considering thermal dynamics and iterative learning control of filling phase

Abstract

Although wet-clutches have excellent thermal endurance, they are technically challenging to control due to their high system nonlinearities and model uncertainties. In general, the largest uncertainty in the wet-clutch system occurs in the clutch friction model, and the feed-forward control performance depends on the accuracy of the model. The largest nonlinearity occurs in the filling phase, which is the initial period of operation of the wet-clutch. In this dissertation, a friction model for a wet-clutch with a paper-based material is proposed for compensation of the model uncertainty. The effect of several variables, including slip speed, clutch friction surface temperature, and normal force, on clutch friction was modeled using a test bench. The asperity friction coefficient appeared to be significantly affected by the friction surface temperature, and it is inferred that it has a linear relationship within the experimental range. The contact friction coefficient structure was proposed based on the experimental data, and its model parameters were obtained by multiple linear regression. Secondly, an observer-based iterative learning control with an output feedback control scheme is proposed for a high relative degree nonlinear system with an unmeasurable control state. Multiple differentiation of measurements can make the control result sensitive to measurement noise when controlling such systems. A convergence analysis was established that can use only lower-order differentiation regardless of the highest-order differentiation, based on the robustness condition to measurement noise and initial estimation error. The application to the vehicular wet-clutch system is presented to illustrate the effectiveness of the proposed method and its learning gain selection. These contributions are verified through theoretical analysis and simulation of the wet-clutch system. The simulation results show the effectiveness of the proposed approach for a class of nonlinear systems with a high relative degree. An example of automatic transmission shift control using the proposed control-oriented modeling of wet clutch friction is presented using simulation.