New smart fault locator in compensated line with UPFC

In this paper, a new smart fault locator in a compensated transmission line with a Unified Power Flow Controller (UPFC) is proposed. Three types of features are extracted from the captured fault signals at one-end of the compensated line by using a time-frequency signal processing tool known as the Hyperbolic S-Transform (HST). The HST, with an asymmetrical window, is an improved version of the S-transform. Then, the regression model of the Support Vector Machines (SVMs) with a non-linear kernel function is applied for the fault location estimation. The proposed smart fault locator gives accurate estimation results by involving hidden statistical features which comprise a new type of time-frequency features. The evaluation of the features and the estimation error obtained under different conditions confirm the efficacy of the proposed smart fault locator.