A new robust approach for fault location in transmission lines using single channel independent component analysis

Reliable fault location schemes for Transmission Lines (TLs) are essential to ensure appropriate actions to keep the service continuity. In this sense, several fault location methods are found in the literature, with special emphasis on methods based on Traveling Waves (TW). Wavelet-based techniques are widely used together with TW because they provide a good response to aperiodic signals. However, methodologies based on orthogonal basis, such as Wavelet Transform (WT) and Principal Component Analysis (PCA), may not contain enough in-formation to interpret TW. Independent Component Analysis (ICA) is able to solve this type of problem, since it recovers the signals based on their statistical independence. The main objective of this work is to propose a fault location method based on Single Channel ICA (SCICA), a simpler variant of ICA that uses only one measured signal. The proposed method is simple, robust and efficient, and does not need a previous database for its development, such as some methods that use artificial intelligence and other ICA based approaches. Therefore, this method overcomes some drawbacks found in conventional solutions, such as accuracy for fault location and independence of load condition, requiring low computational burden. Several fault cases were simulated in EMTDC/PSCAD assuming two different TL models, fault and operating conditions. Using a practical sampling frequency the proposed method was compared to Discrete Wavelet Transform (DWT), always presenting an accurate performance regardless of the fault characteristics.