New Approach Of Calculating Energy Losses Of Distribution Systems Based On RBFNN With Dynamic Adaptive Clustering Algorithm

The energy losses in distribution system is an important index for evaluating economy performance of power system. Accurate and convenient methods of calculating the energy losses in distribution systems are advantageous to work out reasonable measures to decrease the energy losses and improve the power supply quality.This paper presents a new method that applys the radial basis function (RBF) network to calculate energy losses in distribution systems based on analysis present researches in this field. Firstly, an adaptive dynamic clustering algorithm is proposed to determine the number of hidden neurons of RBF network and makes the network have the characteristics of speedy learning and optimal number of hidden neurons; Secondly, RBF network, due to its nonlinear processing ability, is used to map complex nonlinear relationship between the energy losses and feature factors in distribution systems, the trend of energy losses varying with distribution net structure and operation parameters is memorized accurately. A distribution net with 68 typical lines is used as an example, the simulation results verified that the method presented in this paper has the advantages of simple network model, fast learning process and high precision.In order to improve precision of energy losses calculation and nonlinear processing ability of RBF network, this paper presents a new expand RBF nonlinear network, which is one more nonlinear layer between hidden layer and output layer than traditional RBF network. Because of the added nonlinear layer, the nonlinear processing ability of expend RBFNN model is better than traditional RBF network, so it can maps complex nonlinear relationship between the energy losses and feature factors of distribution line. A distribution net with 68 typical lines is also used as an example, the simulation results verified the calculation precision is improved largely; in addition, the distribution system of Binhai power company of Tianjin is used as an practical example, expand RBF nonlinear network is used to analyze and calculate energy losses of this distribution net. In allusion to this distribution net, this paper presents variety of approach to calculate energy losses, such as total energy losses in distribution system, daily energy losses of a single distribution line and real-time...