Please use this identifier to cite or link to this item: https://dspace.upt.ro/xmlui/handle/123456789/7053
Title: Analog circuits fault diagnosis using genetic algorithm based wavelet neural network [articol]
Authors: Gunasundari, A.
Bhuvaneswari, M.C.
Nithyasree, G.
Subjects: Wavelet Neural Network
Fault diagnosis
Monte Carlo analysis
Genetic Algorithm
Issue Date: 2019
Publisher: Timișoara : Editura Politehnica
Citation: Gunasundari,A.; Bhuvaneswari,M.C.; Nithyasree,G.: Analog circuits fault diagnosis using genetic algorithm based wavelet neural network. Timişoara: Editura Politehnica, 2019.
Series/Report no.: Journal of Electrical Engineering;Vol 19 No 2
Abstract: In analog circuits, fault diagnosis is a complex task due to lack of simple fault models and the presence of component tolerances and circuit non- linearity. The method of fault diagnosis in analog circuits using Genetic Algorithm (GA) based Wavelet Neural Network (WNN) is presented in this paper. Wavelet Neural Networks are a new class of networks that combine the classic Neural Networks (NNs) and Wavelet Transform (WT) which inherits the advantages of the Neural Network and wavelet transformation. In this work Wavelet Neural Network is constructed and the network weights are updated to get the desired value at the output nodes of WNN. The constructed WNN is trained with training set until the objective function is minimized and the faults are classified. Then the parameters of WNN are optimized using Genetic Algorithm (GA). The simulation results obtained for WNN and GA based WNN are compared. A comparison of the proposed work of GA based WNN with WNN reveals that the proposed method performs significantly better in fault diagnosis of analog circuits with improved classification accuracy.
URI: https://dspace.upt.ro/xmlui/handle/123456789/7053
ISSN: 1582-4594
Appears in Collections:Articole științifice/Scientific articles

Files in This Item:
File Description SizeFormat 
BUPT_ART_Gunasundari_f.pdf1.07 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.