智能预测控制在无级变速器速比跟踪中的应用

Application of Intelligent Predictive Control in CVT Speed Ratio Control

无级变速器速比控制系统存在着明显的时滞性和非线性,其中时滞性是引起常规控制算法速比跟踪过程中实际速比波动的根本原因。为解决以上问题,设计了无级变速器速比智能预测控制器,该控制器以支持向量机非线性模型作为控制器的预测模型,解决线性预测模型失配问题;以混沌优化算法作为控制器的滚动优化策略,解决最佳控制量实时计算的问题。台架试验结果表明,与常规PID控制器相比,应用智能预测控制器,系统的稳态性能和动态性能都得以提高:在稳态跟踪过程中,速比的波动更小;在阶跃响应中,超调量和过度时间均显著下降。

The continuously variable transmission（CVT） speed ratio system was described with a nonlinear characteristics with pure time- delay by analyzing its basic control principle. The characteristics of time--delay were the basic reason that induced speed ratio tracking fluctuation when the common--used control algorithm was applied in this system. Aiming at this problem,an intelligent predictive controller（IPC） was designed according to the requirements of CVT speed ratio control. The proposed controller utilized the support vector machine（SVM） predictive model to forecast the future control effect; the introduction to this SVM predictive model improves the forecasting precision for its nonlinear behaviors. The optimal control was determined by using a simple chaos optimal algorithm （COA）; the introduction to the chaos optimal algorithm improves the optimal control searching efficiency effectively and solves the problem of real--time performance in control process. It is shown from the bench test that the proposed control algorithm accomplishes the better steady and dynamic performances than those of general PID： the fluctuation reduces in static tracking; the overshoot and the transition time decrease obviously in step--response.