Designing of a New Transformer Ground Differential Relay Based on Probabilistic Neural Network

Low-impedance transformer ground differential relay is a part of power transformer protection system employed for detecting the internal earth faults. This is a fast and sensitive relay, but in some situations such as external faults and inrush current conditions, it may be exposed to maloperation due to current transformer (CT) saturation. In this paper, a new intelligent transformer ground differential relay based on probabilistic neural network (PNN) is presented. To do so, a real power transformer is simulated under a large number of different operating conditions, including internal fault, external fault, and inrush current by using PSCAD/EMTDC software. Then, one cycle data of differential current obtained from each simulation case of mentioned operation conditions is used to provide exemplar patterns. Then, a probabilistic neural network is trained using them. Finally, the trained network is employed as a detection core of the new relay. A comparative evaluation proves the absolute superiority of the proposed method in comparison with some other methods regarding its immunity against maloperation.