Welcome to the InfoSci Platform
Reference Hub3
Prediction of Weld Bead Geometry in Chromium-Manganese Stainless Steel Gas Tungsten Arc Welded Plates Using Artificial Neural Networks

Prediction of Weld Bead Geometry in Chromium-Manganese Stainless Steel Gas Tungsten Arc Welded Plates Using Artificial Neural Networks

R. Sudhakaran, P. S. Siva Sakthivel
Copyright: © 2013 |Volume: 3 |Issue: 3 |Pages: 29
ISSN: 2156-1680|EISSN: 2156-1672|EISBN13: 9781466633681|DOI: 10.4018/ijmmme.2013070102
Cite Article Cite Article

MLA

Sudhakaran, R., and P. S. Siva Sakthivel. "Prediction of Weld Bead Geometry in Chromium-Manganese Stainless Steel Gas Tungsten Arc Welded Plates Using Artificial Neural Networks." IJMMME vol.3, no.3 2013: pp.13-41. http://doi.org/10.4018/ijmmme.2013070102

APA

Sudhakaran, R. & Sakthivel, P. S. (2013). Prediction of Weld Bead Geometry in Chromium-Manganese Stainless Steel Gas Tungsten Arc Welded Plates Using Artificial Neural Networks. International Journal of Manufacturing, Materials, and Mechanical Engineering (IJMMME), 3(3), 13-41. http://doi.org/10.4018/ijmmme.2013070102

Chicago

Sudhakaran, R., and P. S. Siva Sakthivel. "Prediction of Weld Bead Geometry in Chromium-Manganese Stainless Steel Gas Tungsten Arc Welded Plates Using Artificial Neural Networks," International Journal of Manufacturing, Materials, and Mechanical Engineering (IJMMME) 3, no.3: 13-41. http://doi.org/10.4018/ijmmme.2013070102

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

The quality of the weld joint is highly influenced by the welding parameters. Hence accurate prediction of weld bead parameters is highly essential to achieve good quality joint. This paper presents development of neural network models for predicting bead parameters such as depth of penetration, bead width and depth to width ratio for AISI 202 grade stainless steel GTAW plates. The use of this series in certain applications ended in failure of the product as there is no adequate level of user knowledge. Hence it becomes imperative to go for detailed investigations on this grade before recommending it for any application. The process parameters chosen for the study are welding current, welding speed, gas flow rate and welding gun angle. The chosen output parameters were depth of penetration, bead width and depth to width ratio. The experiments were conducted based on design of experiments using fractional factorial with 125 runs. Using the experimental data feed forward back propagation neural net work models were developed and trained using Levenberg Marquardt algorithm. The training, learning, performance and transfer functions used are trainlm, learningdm, MSE and tansig respectively. Four networks were developed with four neurons for the input layer, 3 neurons for the output layer and different nodes for the hidden layer. They are 4 – 2 – 3, 4 – 4 – 3, 4 – 8 – 3 and 4 – 9 – 3. It was found that ANN model based on network 4 – 9 – 3 predicted the bead dimensions more accurately than the other networks. The prediction of weld bead geometry parameters helps in identifying the recommended combination of process parameters to achieve good quality joint.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.