工业过程的子空间模型辨识

Subspace model identification for industrial processes

子空间模型辨识方法(SMI)是一类新兴的直接估计线性状态空间模型的黑箱建模方法,近年来获得了广泛关注.和传统的线性建模方法相比,SMI的优势不仅在于算法本身的简单可靠,也在于它的状态空间表达.本文首先简要介绍了SMI的基本思想以及3种基本算法(N4SID,MOESP,CVA).然后将这类方法应用于一个实际的工业过程建模,同时对3种SMI基本算法和一种传统辨识算法—预测误差方法(PEM)进行了研究对比.

Subspace model identification （SMI） methods, a new class of black-box algorithms to setting up a linear state space model directly from input-output data, have drawn much research attention recently. Compared with the classical linear system identification methods, SMI methods are attractive not only because of their numerical simplicity and stability, but also their availabilities for the state space form. In this paper, the basic idea of SMI methods and three basic algorithms, i.e., N4SID, MOESP and CVA, are briefly addressed at first. Then comparison between the SMI methods and the classical system identification method-PEM （prediction error methods） is made based on the same data sets collected from a real industrial process.