Fault Location in Ungrounded Photovoltaic System Using Wavelets and ANN

Abstract

Identifying ground faults is a significant problem in ungrounded photovoltaic (PV) systems because such earth faults do not provide sufficient fault currents for their detection and location during system operation. If such ground faults are not cleared quickly, a subsequent ground fault on the healthy phase will create a complete short-circuit in the system. This paper proposes a novel fault location scheme in which high frequency noise patterns are used to identify the fault location. The high frequency noise is generated due to the switching transients of converters combined with the parasitic capacitance of PV panels and cables. Discrete wavelet transform (DWT) is used for decomposition of monitored signal (mid-point voltage of the converters) and features are extracted. Norm values of the measured waveform at different frequency bands give unique features at different fault locations and are used as the feature vectors for pattern recognition. Then, a three-layer feed- forward artificial neural networks (ANNs) classifier, which can automatically classify the fault locations according to the extracted features, is investigated. The proposed fault location scheme has been primarily developed for fault location in PV farm (PV panels and DC cables). The method is tested for ground faults as well as line-line faults. These faults are simulated with a real-time digital simulator (RTDS) and the data are then analyzed with wavelets. Finally, the effectiveness of the designed fault locator is tested with varying system parameters. The results demonstrate the proposed approach has accurate and robust performance even with noisy measurements and changes in operating conditions.