Sun Dongye

Fuzzy Neural Network Control of AMT Clutch in Starting Phase

A fuzzy controller of automated mechanical transmission (AMT) clutch in starting phase is developed based on driverpsilas experiences. Focused on the problem that ordinary fuzzy controller design has a lot of subjective factors and the fuzzy controller is difficult to optimize, the fuzzy control model of the AMT clutch was optimized based on adaptive-network-based fuzzy inference system (ANFIS) algorithm and the training, test and simulation analysis were performed for the fuzzy neural network model. The results prove that the membership functions of fuzzy model obtain the satisfactory effect. Applying the optimization model in the simulation, the new control system can avoid the start jerk and better respond to the starting intention of driver.

Model Based Fault Diagnosis of the Clamp Force Control System of Continuously Variable Transmission

On-line diagnosis is one of the most important modules in the control software of Continuously Variable Transmission (CVT). With the development of a prototype CVT, fault diagnosis method of the clamp force control system was developed. Configuration and working mechanism of the clamp force control system was analyzed in detail, based on which mathematic model of the system was established using bond graph theory. Precision of the model was then tested by experiment, and the test results indicated that the model was precise enough for diagnosis use. Fault diagnosis algorithm of the clamp force control system was developed based on the model. Experiment indicated that the proposed diagnosis method of the clamp force control system was effective.