Definition of Statistical-based Data Time Windows for Predicting the Power System Post-contingency Dynamic Vulnerability Status

JAIME CEPEDA; D. GRACIELA COLOMÉ

Porto

Congreso; Intelligent Systems Applications to Power Systems ISAP 2015 Conference; 2015

A joint organization of the ISAP Society and INESC TEC

A data-mining-based approach for predicting the power system post-contingency dynamic vulnerability status (PCDVS) in real time was recently proposed by the authors. This proposal allows determining the system dynamic vulnerability regions (DVRs) by applying a pattern decomposition method based on empirical orthogonal functions (EOF), to a post-contingency data base obtained from phasor measurement units (PMUs) adequately located throughout the system. Afterwards, a support vector classifier (SVC), adequately trained, is employed to predict the post-contingency vulnerability status as regards three short-term stability phenomena, that is: transient stability, short-term voltage stability, and short-term frequency stability (TVFS). This novel methodology uses Statistical-based Data Time Windows (SDTW) that systematically offer the possibility of assessing the different phenomena as well as allow increasing the prediction accuracy. This paper focuses on highlighting the methodology employed for defining the Statistical-based Data Time Windows. In addition, some numerical results obtained by applying the methodology for predicting the PCDVS on the IEEE New England 39-bus test system, that demonstrate the importance of adequately defining the SDTW are presented.