An Adaptive Optimal Process Control of Clutch for AMT with the Use of ADAMS

A dynamic model of a manipulative machine for automobile clutch is setup in ADAMS and further analysis is made to an automatic clutch for AMT. Since the engaging process of clutch is complex and related to riding performance and life span of friction plate, it has to be researched intensively using modern control theory. Adaptive control is effective to avoid engine flameout, which guarantees successful power transmission. Optimal process control can achieve quality engagement given attention to riding performance and life span of friction plate. The function builder and control module of ADAMS provide convenience to model, control and verify the adaptive optimal control strategy for automatic clutch in visual prototyping. Keyword: AMT automatic clutch, optimal control, model reference adaptive control Introduction Automatic Mechanical Transmission (AMT) has been a hotspot in automatic transmission field in recent years. The control for clutch engagement is the key to AMT. A membrane spring clutch is considered in this paper. The difficulty in research on electronically clutch unit (ECU) lies in two, (1) prediction of engine state; (2) optimal engaging law compromising riding performance and life span of friction plate. To verify a control strategy, a lot of experiments using actual vehicle should be performed unless visual prototype experiments is possible and reliable. ADAMS provides a round tool for its function builder and control module. Compared with other kinds of ordinary spring, diaphragm spring (DS) has nonlinear elastic characteristic, which effectively avoids disengage force increasing due to normal wear. And it needs for disengage force almost the same as engage force instead of much larger when coil spring is used. The geometry of a membrane spring is simple but its elastic characteristic is complex, which is described in function builder. The engine output performance is nonlinear with respect to throttle angle and its revolution. The relationship between its stable and dynamic performance has been studied. Thus the engine output performance is described by building 3-D splines in data element. The prediction of the engine state is carried out by model reference adaptive control strategy. The model is set up involving engine, membrane spring, friction plate, manipulative mechanism, clutch parts, shock absorber, gear box, tire and etc.. The visual model is verified by actual experiments, which guarantee correctness and reliability for further control and analysis using ADAMS model. According to minimum theory, an optimal control is performed to achieve compromising optimal performance, that is, less lash and less wear. Synthesizing the above model reference adaptive control for engine and optimal control for riding and life, an accurate control law is worked out.