Header menu link for other important links
X
Comparative studies of simulation, artificial neural network and fuzzy logic for laser shock peening
R.G. Kumar, A. Gopalakrishnan, A. Malav, H.S. Panjawani,
Published in Politechnica University of Bucharest
2018
Volume: 80
   
Issue: 1
Pages: 229 - 242
Abstract
The paper focuses on a development of various modeling techniques to obtain the value stress and deformation for the process of Laser peening on Titanium alloy (Ti6Al4V). The prediction techniques like simulation, artificial neural network (ANN) and fuzzy logic are discussed in this paper. All the three prediction techniques consider overlap of laser spot, power density and pulse duration of laser as input parameters. A design of experiment (DOE) was created according to the Taguchi's experimental design with L18 orthogonal array. Statistical values of the coefficient of determination and absolute percentage error were employed to compare the three developed models by considering initial experiments as well as additional experimental (validation) data. Analysis and results indicate that all three modeling approaches can be equally effectively used for the prediction of compressive stresses with laser peening. However, fuzzy logic model showed the best overall prediction results, while developed ANN model best generalization capability. © 2018 Politechnica University of Bucharest. All rights reserved.
About the journal
JournalUPB Scientific Bulletin, Series D: Mechanical Engineering
PublisherPolitechnica University of Bucharest
ISSN14542358