一种非线性系统的自适应无模型预测控制方法

摘要

In order to avoid modeling difficulties and model mismatch problems of control method based on model which control the nonlinear system , an adaptive model-free predictive control algorithm of nonlinear sysytem is proposed .In the algorithm , nonlinear systems are converted into linear systems which are described by a series of pseudo-partial-derivatives.Then a novel project algorithm is used to estimate the pseudo-partial-derivatives , and the general model of the controlled object is given .By sol-ving quadratic objective function through the receding horizon optimization strategy , the optimized control law is obtained .Simulation result of CSTR process shows that the proposed algorithm is an effective strat -egy with excellent tracking ability and strong robustness .%为了避免基于模型的控制方法在控制非线性系统时存在建模困难和模型失配的问题，提出一种非线性系统的自适应无模型预测控制方法。该方法首先将非线性系统转化为由一组伪偏导数描述的线性系统，然后利用一种改进的投影算法在线估计这组伪偏导数，得到被控系统的泛模型。根据得到的泛模型，推导出预测模型，在此基础上根据预测控制滚动的优化策略求解二次目标函数得出最优控制律。通过对CSTR过程进行仿真验证，结果表明该方法具有良好的跟踪性能和较强的鲁棒性。