Control Inversion: A Clustering-Based Method for Distributed Wide-Area Control of Power Systems

TL;DR

This paper introduces control inversion, a clustering-based method for scalable and simpler wide-area control in power systems, by projecting the system into a lower-dimensional space for easier LQR controller design.

Contribution

It proposes a novel clustering-based projection approach that simplifies wide-area control design and implementation in power systems.

Findings

Effective damping of low-frequency oscillations demonstrated on NPCC 48-machine model

Control inversion achieves comparable performance to traditional LQR control

Method enhances scalability and reduces complexity of wide-area control

Abstract

Wide-area control (WAC) has been shown to be an effective tool for damping low-frequency oscillations in power systems. In the current state of art, WAC is challenged by two main factors - namely, scalability of design and complexity of implementation. In this paper we present a control design called control inversion that bypasses both of these challenges using the idea of clustering. The basic philosophy behind this method is to project the original power system model into a lower-dimensional state-space through clustering and aggregation of generator states, and then designing an LQR controller for the lower-dimensional model. This controller is finally projected back to the original coordinates for wide-area implementation. The main problem is, therefore, posed as finding the projection which best matches the closed-loop performance of the WAC controller with that of a reference LQR controller for damping low-frequency oscillations. We verify the effectiveness of the proposed design using the NPCC 48-machine power system model.