Data Mining Applied to the Electric Power Industry: Classification of Short-Circuit Faults in Transmission Lines

Abstract

Data mining can play a fundamental role in modern power systems. However, the companies in this area still face several difficulties to benefit from data mining. A major problem is to extract useful information from the currently available non-labeled digitized time series. This work focuses on automatic classification of faults in transmission lines. These faults are responsible for the majority of the disturbances and cascading blackouts. To circumvent the current lack of labeled data, the Alternative Transients Program (ATP) simulator was used to create a public comprehensive labeled dataset. Results with different preprocessing (e.g., wavelets) and learning algorithms (e.g., decision trees and neural networks) are presented, which indicate that neural networks outperform the other methods.

An Overview of Data Mining Techniques Applied to Power Systems

“…For simplicity, the dependence on q is omitted hereafter. Some works in the literature use only one of the details or calculate the average power of the coefficients (Morais et al, 2007). In contrast, the framed wavelet front end keeps all the coefficients by taking in account that for 1 > γ they have different sampling frequencies and organizing them as matrix Z .…”

“…The notation is flexible enough to easily describe several wavelet front ends, such as the concatenated wavelet (wavelet-concat, for which min S min L = ) and wavelet-energy described in (Morais et al, 2007). Many recent works adopt the wavelet front end ((Saibal, 2008); ; (Costa, 2006); (Chia-Hun & Chia-Hao, 2006); (Pang & Ding, 2008)).…”

“…It consists in taking the minimum RMS value of each phase during the whole fault duration. As a first step, a normalization is adopted (Morais et al, 2007) to to represent the voltage values in p.u. Because the feature vectors have dimension 3, it is relatively simple to visualize them.…”

Classificação de Eventos em Redes de Distribuição de Energia Utilizando Transformada Wavelet e Modelos Neurais Autônomos

“…Tais eventos são caracterizados por alterar a forma de onda de tensão ou corrente, devido a curtos-circuitos ou má operação de equipamentos instalados nos consumidores de energia [2]. Alterações nas formas de onda podem por vezes resultar em danos aos consumidores, forçando as concessionárias a monitorar o fornecimento da energia para que esta seja entregue de forma adequada [3] nos vários estágios do sistema elétrico de potência.…”

“…O uso do ATP para esse tipo de modelagem jáé bastante conhecido e consolidado -espera-se que os resultados obtidos com essas simulações possam servir de base para o desenvolvimento de um classificador a ser aplicado em dados reais. Grande parte dos eventos reaisé resultante de faltas no sistema elétrico [2], sendo que cerca de 70% destas ocorrem no sistema de distribuição [6] e o restanteé devido a eventos de qualidade de energia. Essa proporção foi mantida nas simulações realizadas no presente trabalho.…”