Uncertainty Analysis of Geomagnetically Induced Currents in UHV Power Grid Based on Compressed Sensing Optimization

Abstract

Geomagnetically induced currents (GICs) caused by geomagnetic storms pose a serious threat to the safe and stable operation of ultrahigh voltage (UHV) and extra-high voltage transmission systems. It is necessary to study the uncertainty of GICs affected by multidimensional input parameters to conduct a comprehensive assessment of the GICs levels in the power grid and their potential threats. The traditional polynomial chaos expansion (PCE) uncertainty analysis method incurs significantly higher computational cost in high-dimensional and high-order scenarios. Taking this into account, a PCE optimization method based on compressed sensing (CS) is proposed in this article. First, the Kriging method was employed to establish an earth conductivity model considering limited measurement data. Then, the proposed method was used to conduct the uncertainty analysis of GICs during storms in the North China UHV grid. The statistics of the maximum GICs in the North China power grid were obtained and compared with results from the Monte Carlo and PCE methods. Results indicate that the PCE method optimized by CS can efficiently analyze the uncertainty of GICs and substantially reduce the computational cost compared with the PCE method while ensuring accuracy. Finally, based on the total Sobol indices, the sensitivities of GICs to the input variables are analyzed to reasonably evaluate the threat of GICs to the power grid.