CONTINGENCY SEVERITY ASSESSMENT FOR VOLTAGE SECURITY USING NONPARAMETRIC REGRESSION TECHNIQUES

This paper proposes a novel approach to voltage security assessment ex ploiting non-parametric regression techniques to extract simple and at the same time reliable models of the severity of a contingency, defin ed as the difference between pre- and post-contingency load power marg ins. The regression techniques extract information from large sets of possible operating conditions of a power system screened off-line via massive random sampling, whose voltage security with respect to contin gencies is pre-analyzed using an efficient voltage stability simulatio n. In particular, regression trees are used to identify the most salie nt parameters of the pre-contingency topology acid electrical state wh ich influence the severity of a given contingency, and to provide a fi rst guess transparent approximation of the contingency severity in ter ms of these latter parameters. Multilayer perceptrons are exploited to further refine this information. The approach is demonstrated on a re alistic model of a large scale voltage stability limited system, where it shows to provide valuable physical insight and reliable contingenc y evaluation. Various potential uses in power system planning and oper ation are discussed.