Classification of faults in a TCSC compensated transmission line using data mining algorithm

This article presents the decision tree (DT) as an effective data mining tool for classification of fault in a thyristor-controlled series capacitor (TCSC) compensated line. The presented DT-based fault classifier uses discrete wavelet transform (DWT) as feature extraction tool. The presented technique offers fast response as it utilizes 1/2 cycle post-fault sample of current signal measured on relaying end. After being trained with 2688 cases, the technique has been validated on large data set (i.e. 4032 cases). Based on the studies carried out in this article, the proposed technique is accurate, fast and immune to noise and compensation level.