Title :
A Practical Method of Calculating the Energy Losses in Distribution Systems Based on RBF Network
Author :
Hui-Lan, JIANG ; Min, AN ; Jie, LIU ; Jian-Qiang, XU
Author_Institution :
Tianjin Univ.
Abstract :
This paper presents a method of calculating energy losses in distribution systems based on RBF network. For representational samples of reflecting the relation between energy losses and characteristic parameters of distribution net, RBF network with strong regression characteristic is adopted to map complex non-linearly relation between energy losses and characteristic parameters of distribution net, the trend of energy losses varying with distribution net structure and operation parameters is memorized accurately. Meanwhile this paper proposes an optimal clustering criterion of being used to determine the number of nodes of hidden layer, and therefore the use efficiency of the RBF network is improved. In addition, there are some special samples which are difficult to be attracted strongly by network, so the computational error of energy losses for these lines is larger than others. This paper proposes a mechanism of reconstructing network to enhance the precision for samples with larger error by using the fuzzy processing. The distribution net with 68 lines is used as an example, the simulation results have verified that the way presented in this paper has the advantages of simple model, speedy learning and high precision
Keywords :
distribution networks; fuzzy systems; power engineering computing; radial basis function networks; regression analysis; RBF network; characteristic parameters; computational error; distribution net structure; distribution systems; energy loss calculation; fuzzy processing; high precision; nonlinearly relation; optimal clustering; regression characteristics; speedy learning; Artificial neural networks; Clustering algorithms; Computational modeling; Computer networks; Energy loss; Equations; Neural networks; Petroleum; Radial basis function networks; Regression analysis; Distribution systems; Energy losses; RBF network; Self-adaptive clustering algorithm;
Conference_Titel :
Transmission and Distribution Conference and Exhibition: Asia and Pacific, 2005 IEEE/PES
Conference_Location :
Dalian
Print_ISBN :
0-7803-9114-4
DOI :
10.1109/TDC.2005.1546858