Information theoretic index for regime shifts in power systems

Abstract

A regime shift occurs when a dynamical system's state shifts from a nominal equilibrium to an alternative equilibrium. This concept is often used to describe the collapse of ecological systems, but it also appears in many engineering systems; most notably in the voltage collapse of power systems. This paper uses an information theoretic index known as the Fisher Information (FI) as a real-time statistical index detecting regime shifts in a stressed IEEE 118 bus system. These preliminary experiments demonstrate that the FI index indeed provide early warning of abnormal operations leading to voltage instability. As a data driven method the proposed FI index provides a useful tool supporting the real-time monitoring of power system operations.