Modeling and estimation for stepped automatic transmission with clutch-to-clutch shift technology.

摘要

A major problem in designing controllers for automatic transmissions in automobiles is that many of the relevant variables characterizing the performance of the power train are not measurable because of sensor cost and reliability considerations. This dissertation presents the development of model-based estimation for a stepped automatic transmission with clutch-to-clutch shift technology, to provide real-time information about some necessary but un-measurable variables. Specifically, model-based real-time estimation of clutch pressures for the clutches involved in the gear shifting process is developed and validated.; In this dissertation, mathematical models for all systems affecting the dynamic behavior of the transmission, namely, the torque converter, transmission mechanical components, shift hydraulic system, and vehicle and driveline are developed. The main focus, however, is on the development of mathematical models describing the dynamic behavior of the transmission mechanical components, viz, the dynamics of the gear sets inside the transmission and the shift hydraulic system. This dissertation presents the development of a nonlinear dynamic model for the shift hydraulic system for the transmission of interest. The result of the model development presented here includes a fully detailed model, which is complex and highly nonlinear, as well as a simplified model which is used for real-time implementation. Resulting models are validated against experimental data.; The availability of the shift hydraulic model leads to the development of the model-based clutch pressure observer. Sliding mode observers are used due to their ability to deal with nonlinear systems, robustness against uncertainties, and ease of implementation. In this dissertation, both continuous-time and discrete-time sliding mode observers are developed and implemented. Since the turbine torque is needed by the observer but is not available, an adaptive sliding mode observer is also implemented in this research to improve the on-line estimation of the turbine torque, and hence improve the accuracy of the clutch pressure estimation. The resulting observers are validated via off-line simulation tests, and are also implemented in real-time at different sampling frequencies on a test vehicle, in order to demonstrate observer performance and establish the feasibility of using the designed observers on current transmission control units.