Model for Solidification Cracking in Low-Alloy Steel Weld Metals

K. Ichikawa, H. K. D. H. Bhadeshia and D. J. C. MacKay
University of Cambridge


This is the first application of a classification neural network based on a Bayesian framework. Data on the occurrence of solidification cracking in low alloy steel welds have been analysed using this method. It has been possible to express quantitatively the effect of variables such as the chemical composition, welding conditions, and weld geometry, on the tendency for solidification cracking during freezing. The ability of the network to express the relationship in a suitably non-linear form is shown to be vital in reproducing known experimental phenomena.

Science and Technology of Welding and Joining, 1 (1997) 43-50.

Download PDF file of paper

See the following url site for data appropriate to the present work:

PT Group Home Materials Algorithms Any Valid CSS!