Fast contingency filtering based on linear voltage drop estimates

This paper deals with the filtering of harmless contingencies in voltage stability and security analyses. In many systems, a post-contingency load flow allows to identify the contingencies with severe impact on voltage stability, as causing either divergence or large voltage drops. However, for filtering purposes, accurate voltage drops need not be computed; linear estimates obtained from sensitivity formulas are appropriate. For a given contingency, the method used in this paper solves a sparse linear system to update the phase angles, assuming constant voltage magnitudes. Then, assuming constant active power flows in the branches, a second sparse system is solved to update the voltage magnitudes. This yields better accuracy than one full load flow iteration while retaining the sparse structure of a decoupled formulation. For contingency analysis, the incidents are filtered by comparing the voltage drops to a threshold. For secure power margin computations, the same procedure is used after stressing the system in the specified direction. The method has been tested on a real system where it has been found to combine simplicity of implementation, quality of filtering and computational efficiency.