No abstract

No abstract

Abstract. For a class of process which had characteristics of nonlinear time-varying, dynamic characteristic varing significantly and can not be accurately modeled mechanism, this paper proposed a adaptive subspace predictive controller for stable operation. The method compared the prediction error to update prediction model online, and adjuste the length of the scroll window automatically according to the feedback error, which enhanced the adaptability for nonlinear time-varying characteristics, and strengthened its ability to inhibit unpredictable interference. Finally, simulated to verify the effectiveness of the method.