Study On Model-Based Engagement Control For AMT Automotic Clutch

The engagement control of clutch is the key problem in the research of AMT. There are two evaluation indicators,namely jerk and friction work,and they are conflicting. In order to harmonize the two evaluation indicators each other, this paper is focused on the model-based control in the engagement process of AMT automatic clutch, mainly include:According to the structure of AMT transmission, the dynamic equations of the vehicle were established and simplified. Moreover, the dynamic equations of the experiment bench were established in slipping and synchronous stages. The friction radius and the maximum engagement speed of clutch were calculated. The friction torque was tested on the bench. It was testified the correction and feasibility of dynamic model by the simulation model.The simulation model was constructed to identify the parameters, such as inertia, stiffness and damping coefficient. The signals were tested from AMT experiment bench, including the output torque of engine, the friction torque, the engine speed, the clutch speed and the equivalent load speed. The signals were preprocessed as input and output of the simulation model. Then, the least square method and genetic algorithm were used to identify the parameters for AMT in second gear and verify in first gear.The control strategy was designed after parameter identification. According to the jerk and friction work in different stages, Four methods were designed, namely, fuzzy control, clutch speed model control, engine constant speed control and torque matching control. The control models were designed based on MATLAB/Simulink. According to jerk and friction work, both engine constant speed model and torque matching model were determined to take experiment.According to the above analysis, the electronic control system for clutch engagement was designed based on Freescale MC9S12DT128 MCU and dSPACE. The MC9S12DT128 MCU was to collect the signals from each sensor on the AMT test bench as well as outputing the control signals to the actuator of the clutch. And at the same time, it also sent the signals that collect to the dSPACE by CAN bus. Then, according to the output signals of each sensor, the dSPACE made the rapid prototyping calculate about the clutch engagement and the gear shift. The results sent to the MC9S12DT128 MCU by CAN bus. The engine constant speed model and torque matching model were used to control clutch engagement on the bench. The test results showed that: the model-based control strategy could ensure smooth and fast start of vehicle under various conditions.