Identifying inrush currents based on Bayesian recursive algorithm for a numerical overcurrent protection relay

Abstract

The overcurrent relays are one of the most important devices on the electrical power system because they help to eliminate the abnormal currents. However, the relays sometimes cause nuisance trips when the circuit is starting because they fail to discriminate between the short-circuit current and the inrush current. This paper introduces a new principle of discrimination between inrush currents and short-circuit currents based on Bayesian recursive algorithm. Information about the abnormal current patterns as well as the instance of the event can be represented in terms of probability density functions, Bayesian rules help combine the probability information to find the most probable prediction about the current states. Experiment results demonstrate that this method is effective and helps the relay avoid nuisance tripping of short-circuit protection.