Research On Three Phase Unbalance Additional Loss In Low Voltage Station Based On Machine Learning

Distribution network line loss rate is a very important technical and economic index for power supply enterprises,and it is the comprehensive embodiment of power supply enterprise management level,which directly determines the economic benefits of enterprises.For three-phase imbalance increases in low voltage power distribution line loss,at the same time,given the current three-phase unbalanced condition of additional line loss calculation,using assumptions simplify the algorithm,the results of the calculation error,the calculation result is not convergence problems,according to the structure features of the low voltage distribution network based on radial basis function is proposed based on improved K Means clustering algorithm and RBF neural network of low voltage distribution network unbalanced three-phase additional line loss analysis method.This paper introduces the concepts of three-phase balance and unbalance,establishes the mathematical model of additional loss of three-phase unbalance line,and analyzes the causes of additional loss.At the same time was carried out by the BP neural network used in unbalanced three-phase computing research,analysis of the basic model of BP network,and put forward the current BP neural network is applied to the circuit good computing that exist in the network model is easy to fall into local minimum value in the process of solving the solving process is terminated,the network model in the process of training the problem such as slow convergence speed.Secondly,based on the BP neural network is applied to the network model of line loss calculation in the process of training convergent speed is slow,this chapter will be carried out based on K Means clustering algorithm and improved RBF neural network is the study of the three-phase uneven loss calculation,analysis of BFF neural network algorithm in the application of line loss calculation,aiming at the problem of distribution network sample data diversity at the same time,with the improved K-Means of classifying data,finally proposed based on improved K Means clustering algorithm and RBF neural network of three-phase uneven loss calculation,and the flow chart of corresponding algorithm is designed.Finally,in order to validate the proposed based on improved K Means clustering algorithm and RBF neural network algorithm of quickness and accuracy,this chapter based on K Means clustering algorithm and improved RBF neural network algorithm,and to analyze a 101 sample,and compared with the traditional BP neural network analysis,through the calculation result shows that when set training goals of error is 0.01,the BP neural network iteration model convergence when 320 times,based on the improved K Means clustering algorithm and iterative 52 RBF neural network model of convergence,when training targets set error is 0.0001,the BP neural network model,based on the improved K Means clustering algorithm and RBF neural network convergent iterative 460 times model,instance analysis,using this article is based on K Means clustering algorithm and improved RBF neural network analysis of 101 samples calculated results,the relative error of 20%,only six samples through example analysis shows that the proposed based on improved K Means clustering algorithm and the convergence of RBF neural network model,the solving speed and precision are better than the traditional BP neural network model.