Inverse problem of photoelastic fringe mapping using neural networks

This source preferred by Venky Dubey

Authors: Grewal, G.S. and Dubey, V.N.

http://eprints.bournemouth.ac.uk/10712/

Journal: Measurement Science and Technology

Volume: 18

Pages: 1361-1366

ISSN: 0957-0233

DOI: 10.1088/0957-0233/18/5/024

This paper presents an enhanced technique for inverse analysis of photoelastic fringes using neural networks to determine the applied load. The technique may be useful in whole-field analysis of photoelastic images obtained due to external loading, which may find application in a variety of specialized areas including robotics and biomedical engineering. The presented technique is easy to implement, does not require much computation and can cope well within slight experimental variations. The technique requires image acquisition, filtering and data extraction, which is then fed to the neural network to provide load as output. This technique can be efficiently implemented for determining the applied load in applications where repeated loading is one of the main considerations. The results presented in this paper demonstrate the novelty of this technique to solve the inverse problem from direct image data. It has been shown that the presented technique offers better result for the inverse photoelastic problems than previously published works.

This data was imported from Scopus:

Authors: Grewal, G.S. and Dubey, V.N.

http://eprints.bournemouth.ac.uk/10712/

Journal: Measurement Science and Technology

Volume: 18

Issue: 5

Pages: 1361-1366

eISSN: 1361-6501

ISSN: 0957-0233

DOI: 10.1088/0957-0233/18/5/024

This paper presents an enhanced technique for inverse analysis of photoelastic fringes using neural networks to determine the applied load. The technique may be useful in whole-field analysis of photoelastic images obtained due to external loading, which may find application in a variety of specialized areas including robotics and biomedical engineering. The presented technique is easy to implement, does not require much computation and can cope well within slight experimental variations. The technique requires image acquisition, filtering and data extraction, which is then fed to the neural network to provide load as output. This technique can be efficiently implemented for determining the applied load in applications where repeated loading is one of the main considerations. The results presented in this paper demonstrate the novelty of this technique to solve the inverse problem from direct image data. It has been shown that the presented technique offers better result for the inverse photoelastic problems than previously published works. © 2007 IOP Publishing Ltd.

This data was imported from Web of Science (Lite):

Authors: Grewal, G.S. and Dubey, V.N.

http://eprints.bournemouth.ac.uk/10712/

Journal: MEASUREMENT SCIENCE AND TECHNOLOGY

Volume: 18

Issue: 5

Pages: 1361-1366

eISSN: 1361-6501

ISSN: 0957-0233

DOI: 10.1088/0957-0233/18/5/024

The data on this page was last updated at 04:42 on September 22, 2017.