Detection Method of Distribution Transformer Capacity Increase based on Neural Network and Short Circuit Impedance

Abstract

This paper proposes an online identification method of transformer capacity increase based on neural network combined with short-circuit impedance method to address the problems of distribution transformer capacity increase done by customers and discrepancy between nameplate capacity and actual capacity in the operation of distribution transformers for large customers. Using the actual data from the electric energy information acquisition system, a recurrent neural network algorithm is built and simulated. In the same time, a short-circuit impedance based model is set and tested using S9 series 10kV class distribution transformers. The experimental results show that the algorithm as well as the model can accurately indicate the status of the distribution transformer, which verifies the accuracy and the effectiveness of this method.