Open AccessNeural network model for predicting ferrite number in stainless steel welds
Author(s) -
J.M. Vitek,
Y.S. Iskander,
E.M. Oblow,
S. S. Babu,
S. A. David
Publication year - 1998
Language(s) - English
Resource type - Reports
DOI - 10.2172/290929
Subject(s) - ferrite (magnet) , materials science , metallurgy , artificial neural network , composite material , computer science , artificial intelligence
Predicting the ferrite content in stainless steel welds is desirable in order to assess an alloy`s susceptibility to hot cracking and to estimate the as-welding properties. Several methods have been used over the years to estimate the ferrite content as a function of the alloy composition. A new technique is described which uses a neural network analysis to determine the ferrite number. The network was trained on the same data set that was used to generate the WRC-1992 constitution diagram. The accuracy of the neural network predictions is compared to that for the WRC-1992 diagram as well as another recently proposed method. It was found that the neural network model was approximately 20% more accurate than either of the other two methods. In addition, it is suggested that further improvements to the neural network model, including the consideration of process variables, can be made which lead to even better accuracy