Predicting the Line Loss for a 10 kV Distribution Network Using AGA-BPNN

Abstract

With the rapid development of power grids, it is essential for these grid enterprises to pay more and more attention to the comprehensive reduction of power loss. Due to a large number of medium voltage power network based on distribution phenomenon nodes and power loss, it is difficult to collect accurate data of more than 20% of the total power loss of the network based on distribution phenomenon. Based on the above problems, it is significant to propose a method to quickly and accurately predict the losses in line of the 10 kV network based on distribution phenomenon. The improved BP neural network model and adaptive evolutionary algorithm programming can effectively analyze the 10kV transmission and distribution problems and reduce the line loss in the transmission project. The line loss prediction model mainly includes data cleaning, electrical characteristic index system, determination of the number of nodes in the BPNN hidden layer, and losses in line prediction. For this purpose, AGA-BPNN is proposed in this paper. Additionally, cable quality, power factor management, and reduction resistance are some parameters that can help lower the losses in lines. The author studies the performance of the losses in the line prediction model before and after the improvement of BPNN to validate the application impact of the AGA-BPNN algorithm in losses in line prediction of a 10 kV network based on distribution phenomenon. This technique has benefits of rapid convergence and great accuracy over ordinary BPNN. The simulation and computation of the example validate the suggested approach.