Detrending and Characterizing System Frequency Oscillations Using an Adapted Zhou Algorithm

Electro-mechanical oscillations between interconnected synchronous generators and oscillations in system frequency are an inherent part of the operation of large power systems. Very Low Frequency (VLF) oscillations are usually classified as oscillations in the 0.01-0.1 Hz range. With the move towards variable renewable energy sources and low-inertia power systems, VLF oscillations are being observed with increasing regularity in many small and island grids. If left undamped, these can present a threat to system stability. However, finding the root cause and source(s) of VLF oscillations is an extremely challenging task for network operators. Recent work has identified a need for improved tools for identifying and characterising VLF oscillations, in order to determine the combination of system conditions that can be used as predictors for VLF events. A suitable small signal model is also required in order to enable verification of the root cause of VLF events and study of mitigation measures. Accordingly, this paper presents a new approach for detrending and characterizing system frequency oscillations using an adapted Zhou algorithm. The paper also describes a method for applying this algorithm for the detection/location of oscillations, and for their detrending and characterization. Finally, an approach for relating detected oscillation events to power system operating conditions for diagnostic purposes is described. The effectiveness of the proposed approach is demonstrated using a single frequency power system model, and using system frequency oscillations recorded from the Irish power system.