A method of distribution network line loss calculation based on neural network

摘要

A method for calculating the line losses in distribution system is presented using Artificial Neural Network (ANN) combined with Quantum Genetic Algorithm (QGA). Firstly, sample data is acquired through an actual measurement of the load for the ANN training. Secondly, the three parameters of ANN are determined, i.e., numbers of neurons in the hidden layer, learning rate and momentum constant. Lastly, BP models are applied to construct the model for line losses calculation. The simulation results for the practical case study demonstrate that the proposed method is effective and more accurate than other methods.