Reduced-Order Method for Detecting the Risk and Tracing the Sources of Small-Signal Oscillatory Instability in a Power System With a Large Number of Wind Farms

Abstract

A power system with a large number of wind farms may consist of several thousand or more wind turbine generators (WTGs). Owing to the high dimension of the linearized model of such a power system, applying conventional modal analysis may be impossible due to the numerical complexity. This study proposes a reduced-order method of modal analysis to detect the risk and trace the sources of oscillatory instability within the power system. The proposed method does not use the representation of wind farms by aggregating WTGs to reduce model dimension. The maximum dimension of the matrix involved in the reduced-order modal analysis is either the total number of WTGs in the power system or the order of dynamic models of individual WTGs. Hence, the proposed method effectively avoids the numerical complexity of applying modal analysis to the high-dimensional model. Improvements of the reduced-order method in practical applications are suggested, particularly for situations in which detailed parametric information for every WTG is not available. Initially, a few representative WTGs are selected to derive their parametric models via field or laboratory tests. Then, reduced-order modal analysis is applied to detect the risk and trace the sources of oscillatory instability.