预测控制的预报误差校正法研究

基于时间序列方法对模型误差过程建立模型，用递推最小二乘法估计模型误差过程的参数，由最小方差预报原理给出模型误差的预报，在上原理的基础上，对误差预报校正器进行了设计，并在动态知控制（DMC）中，用预报误差校正代替一般的误差校正，经过系统仿真比较，可以看出误差预后和误差预报校正的结果令人满意。

A Study of Prediction Error Correction Method of Predictive Control

This paper establishes ARMA model of model error process based on time series analysis method, estimates error model parameter by recursion least square, presents the prediction of model error based on prediction principle of least variance. It designs corrector of prediction error on the basis of above-mentioned principle, and uses this corrector to replace general error correction in dynamic matrix control. By the comparison of system simulations, the results of error prediction and error prediction correction are satisfactory.