Resumen

SORIA, Mauricio; TOVAR, Antonio; MALDONADO, Diego  y  FABARA, Cristian. Data Mining for Patterns Recognition of Power Systems Static Security Assessment with Contingency Events. Revista Técnica energía [online]. 2019, vol.16, n.1, pp.17-22. ISSN 2602-8492.  https://doi.org/10.37116/revistaenergia.v16.n1.2019.331.

This paper analyzes the static security of the power system, applying advanced data mining techniques that allow the evaluation of safety patterns of a power electrical system in a steady state analysis with contingency events N-1. Data are obtained through power flows, to perform Monte Carlo simulations with scripts developed in Python. Using the DIgSILENT PowerFactory simulation software, 10,000 scenarios are analyzed, which allows us to consider the uncertainty of the system according to the probabilistic nature of the system. The static security indexes of the system are calculated to classify the types of contingencies as safe, critically safe, insecure and highly unsafe. Data mining is developed by means of an algorithm programmed in Python language with which the design of the multiclass support vector machine classifier (SVM Multiclass) is carried out. It is trained to determine if a contingency is safe or unsafe. The parameters of the SVM were obtained through an optimization with a differential evolution algorithm (Differential Evolution). The results of the validation of the classifier showed that the technique is very effective in classifying new contingencies. The methodology is applied to an IEEE test system of 39 buses.

Palabras clave : Contingency events; data mining; pattern recognition; Monte Ca rlo simulation; static security assessment.

        · resumen en Español     · texto en Español     · Español ( pdf )