High-impedance fault detection system using an adaptive element model

An adaptive-element model approach for high impedance fault detection is presented. The system employs multiple detection algorithms using different frequency parameters. To obtain a final fault/nonfault classification for a power distribution feeder, a revised adaptive-element model is applied using expert weights associated with each algorithm-parameter pair. The adjustment of the weight, assigned by experts, is performed by an elliptic formula. This revised adaptive-element model is different from the original adaptive-element scheme in two components: the combination method for multiple inputs and the learning algorithm for weight correction. This system when tested using examples shows good performance, especially in terms of security against false identification.