基于改进子空间模型辨识法的交直流系统降阶模型在线辨识

Online Identification for Low-order Model of AC/DC System Based on Modified Subspace Model Identification Method

广域测量系统(WAMS)的发展使得交直流系统模型在线辨识成为可能。基于加快辨识速度以及准确自适应定阶的考虑,提出了改进子空间辨识方法。通过交直流系统机侧、网侧最佳控制点以及最佳观测点的选择确定输入输出信号点,并在输入点注入带通信号以激发低频振荡模态,从而得到系统振荡模态的输出响应信号。改进子空间方法可同时根据输入输出信号得到系统降阶模型,而多种故障条件下Kalman预估器的估计信号与实际信号的接近程度则可验证方法的鲁棒性。8机36节点交直流系统的仿真结果表明,该方法能够准确、有效辨识出交直流系统降阶模型,为交直流系统中央协调阻尼控制器的设计奠定了基础。

Online identification of AC/DC system model becomes realizable due to the development of wide area monitoring systems （WAMSs）. We proposed a modified subspace identification method （MSIM） which accelerated the identification of model parameters and selected model order adaptively and accurately. After determining the optimal controllable points and the optimal observable points in machine/grid sides, we obtained the signal input and output points. At the same time, injecting bandpass signals into the input points to make the system to get into low frequency oscillation modes, we obtained the output response signals. The proposed MSIM helps to identify the system model by using input/output signals. The similarity between estimated signals by Kalman predicted observer and the actual measured signals validates the robustness of the proposed method. The simulation of a AC/DC system with 8 machines and 36 nodes shows that using the proposed method can identify the low-order model of the system accurately, thus the method can be used for designing central coordinating damping controllers for the relevant controllable devices of the AC/DC system.