Calculation Method of the Line Loss Rate in Transformer District Based on Neural Network with Optimized Input Variables

The line loss rate of transformer district has been paid close attention by power supply companies. With the accumulation of line loss sample data, artificial neural network algorithm has been widely used in line loss calculation. In view of the influence of the type and quantity of input variables on the calculation accuracy of artificial neural network model, this paper proposes an artificial neural network line loss calculation method based on improved k-means clustering and grey relational analysis. Aiming at the problem that the line loss characteristics of different transformer district are significantly different, the improved k-means algorithm which optimizes the initial clustering center is used to classify the transformer districts. Further, in order to reduce the network training burden caused by too many input variables, the correlation between electrical characteristic indicators and line loss rate is determined by grey relational analysis, and the number of input variables in each type of transformer district is determined by K-fold cross validation method. The simulation results show that the method can improve the accuracy of line loss calculation, and provide a theoretical basis for scientific and reasonable loss reduction.