LAUSR

Geomagnetically Induced Currents (GICs) are induced after a complex interaction between the sun and earth’s magnetic field. GICs may cripple power systems leading to voltage collapse and a drop in power quality. While the commonly-used dc model for GICs is valid for steady-state analyses of power systems and transformer responses, GICs must be represented with their continuously varying magnitudes and frequency components to understand the dynamic power system’s response to solar storms. In this paper, we used measured GIC data from the 2003 Halloween storm to compare several signal processing techniques. The Fast Fourier Transform (FFT) with a fixed window size, Short-Time Fourier Transform (STFT) with a variable window size, wavelet transform, and Particle-Swarm Optimization (PSO) methods were used to identify the GIC frequency characteristics. To understand the effects of these GIC characteristics on voltage stability, a multi-machine network was modelled on a digital real-time simulator (OPAL-RT). The results reveal the limitations of fixed window FFT in capturing the dominant high-frequency GIC bandwidth arising during peak solar activity. Wavelet and STFT are recommended for GIC data analysis as they provide time-frequency localization. High-frequency GICs during sudden storm commencements and pulsation activity are more of a concern for voltage stability than lower frequency components. Novel results from this study suggest that limiting contingency analysis to only high magnitudes of dc/GIC does not cater for the underlining GIC dynamics. In addition to magnitude thresholds, contingency analysis should incorporate GIC characteristics such as frequency which affect voltage dynamics and thus power system stability.