Stator Current Harmonics Evaluation by Flexible Neural Network Method With Reconstruction Structure During Learning Step Based On CFE/SS Algorithm for ACEC Generator of Rey Power Plant

• Main Authors: Mohammad Reza Yousefi, Mohammad Teshnehlab
• Format: Article
• Language: English
• Published: Najafabad Branch, Islamic Azad University 2010-07-01
• Series: Journal of Intelligent Procedures in Electrical Technology
• Subjects: Finite element method; flexible neural network; capability curve and synchronous generators
• Online Access: http://jipet.iaun.ac.ir/pdf_4470_1f407a59233a2717e732cbab666f2471.html
• ISSN: 2322-3871; 2345-5594

One method for on-line fault detection in synchronous generator is stator current harmonics analysis. In this paper, the flexible neural network with reconstruction structure during learning has been used to evaluate the stator current harmonics in different loads. Generator modeling, finite element method and state space model make training set of flexible neural network. Many points from generator capability curve are used to complete this set. Flexible neural network that is used in this paper is a perception network with single hidden layer, flexible hidden layer neuron and back propagation algorithms. Results are show that the trained flexible neural network can identify stator current harmonics for desired load from the capability curve. The error is less than 10% in compared to the values obtained directly from the CFE-SS algorithms. The parameters of modeled generator are 43950(KVA), 11(kV), 3000(rpm), 50(HZ), (PF=0.5).