Design, modeling, and control of automotive power transmission systems.

Abstract

This thesis focuses on investigating the design, modeling and control methodologies, which can enable smooth and energy efficient power transmission for conventional, hybrid and future automotive propulsion systems. The fundamental requirements of the modern power transmission system are: (1). It should be able to shift the torque transmission ratio efficiently and smoothly to enable the fuel efficient operation of the power source. (2). It should be able to reject/damp out the power source torque oscillation in an energy saving fashion to avoid rough torque transfer to the driveline. Critical factors determining the successful power transmission include the appropriate control of the power transfer key components (clutches), the optimal power transmission coordination with the automotive driveline system, and the capability to smooth out the power source input oscillation in a fuel efficient fashion. To meet these resolutions, this thesis will investigate the enabling design and control methodologies for power transmission in three levels: the fundamental clutch level, the intermediate driveline level, and the entire propulsion system level. First, the clutch level design is investigated in two categories: open loop control and closed loop control. For the open loop, two key issues are addressed. One is to ensure the consistent initial condition with optimal valve structure design, and the other is the clutch fill process optimization using a customized dynamic programming with reduced computational cost for stiff hydraulic system. For the closed loop control case, the solutions are further divided into two groups. One is to enable feedback with pressure sensor measurement, and the other is to close the control loop without any sensor. Through experiments, both methods are shown to enable precise, fast and robust clutch actuation. Second, the driveline level design considers optimizing the power transmission coordination with the driveline. Optimal conditions to achieve efficient and smooth torque transfer are formulated. The nonlinear optimization is then solved using the Dynamic Programming. Finally, in the propulsion system level, the engine start/stop torque oscillation rejection problem for hybrid vehicle and future advanced combustion system is discussed. Through proper formulation, this problem can be treated as disturbance rejection for a linear parameter varying (LPV) system under the internal model principle. To experimentally implement the state of the art controller design, two problems should be solved. First, the vibration signal is periodic with changing magnitude, whose generating dynamics has not been studied before and needs to be derived. Second, the current linear time varying internal model control is lack of robustness, and the design method of a low order yet robust internal model stabilizer is still unavailable. This thesis proposes promising approaches to address this fundamental bottleneck issue in the time varying internal model control theory, which is one of the key contributions in this thesis. The proposed stabilizer synthesis method is treated in a general form, and can potentially be applied to other applications beyond the automotive field as well. Experimental results are also shown to validate the effectiveness of the proposed algorithm. In summary, the contributions of this thesis span from the control applications to the fundamental control theory. Application wise, this thesis formulates the smooth and efficient power transmission design and control problem in three levels, and proposes design, dynamics analysis and control methodologies to address the critical challenges in each level respectively. For control theory, a robust and low order stabilizer synthesis method is proposed to enable reference tracking/disturbance rejection based on linear time varying internal model principle. This stabilizer design addresses one of the most critical issues in the linear time varying internal model control synthesis, which facilitates experimental investigation of the internal model controller in the LTV setting.