A Feedforward Neural Network Based on Improved Particle Swarm Optimization Algorithm for Identification of User Stealing Electricity

In order to reduce the negative impact of electricity theft on power grid operation, improve the ability of electricity theft inspection and the accuracy of identifying electricity theft users, a feed-forward neural network model based on BFO-PSO is proposed to identify users behavior of stealing electricity. Based on the electricity consumption data of a city in recent four years, through feature extraction, four features that have a greater impact on electricity stealing behavior are obtained as input samples. A feedforward neural network recognition model based on BFO-PSO algorithm is constructed, and the optimal weight value of BP neural network model are calculated by using algorithm BFO-PSO. By comparing the recognition results of BP neural network model, genetic algorithm and the BFO-PSO algorithm, it is found that the feedforward neural network model based on BFO-PSO algorithm can better identify the power stealing users. The recognition accuracy is as high as 94%, and the training speed is increased by 5%. It is expected to be widely used in the power stealing user recognition..