The Application of Generalized Predictive Control in CVT Speed Ratio Control

The continuously variable transmission (CVT) speed ratio system is described with a feeble nonlinear characteristic with pure time-delay by analyzing its structure and basic control principle; this is the basic reason that induces speed ratio tracking fluctuation when the common-used control algorithm is applied in this system. Aiming at this problem, a generalized predictive controller optimized by the genetic algorithm is designed according to the requirements of CVT speed ratio control. The proposed controller utilizes the commonly used controlled autoregressive integrated moving average (ARIMA) predictive model to forecast the future control effect. The optimal control is determined by using a simple genetic algorithm; the introduction of this genetic algorithm improves the optimal control searching efficiency effectively and solves the problem of real-time performance in control process. By analysis of bench test results, this control algorithm improves the system dynamic and steady response performance compared with general PID. At last, for the problem of control quality decline caused by the mismatching of the CARIMA model, an on-line solution scheme to adjust the parameters of the CARIMA model is suggested in this paper.