PSO trained ANN-based differential protection scheme for power transformers

Abstract

In this paper, a novel scheme is proposed which monitors and discriminates the different operating conditions (normal, magnetizing inrush current, over excitation of the core, internal faults and CT saturation due to external faults) of power transformers. The Particle Swarm Optimization (PSO) technique is used to train the multi-layered feed forward neural networks to discriminate the different operating conditions. The proposed scheme has been realized through two different Artificial Neural Network (ANN) architectures. One is used as an internal fault detector and the other one detects and discriminates the other operating conditions like normal, inrush, over excitation, and CT saturation due to external faults. These two ANNs were trained using Back Propagation Neural Network Algorithm (BPN) and PSO technique and the simulated results are compared. Simulations are performed for the practical power transformer ratings obtained from Tamilnadu Electricity Board (TNEB), India. Comparing the simulated results of the above two cases, training the neural network by PSO technique gives more accurate (in terms of sum square error) and also faster (in terms of number of iterations and simulation time) results than BPN. The PSO trained ANN-based differential protection scheme provides faster, accurate, more secured and dependable results for power transformers.