Predicting haz hardness with artificial neural networks
The use of artificial neural networks (backpropagation networks) for predicting heat-affected zone hardness. given the 800-500 C cooling time and chemical composition, is investigated in this study. The experimental training data are taken from a database assembled by Yurioka et al. Network predicted hardness values are compared with experimental values from the entire Yurioka database and reasonable agreement is found (correlation factor = 0.98). The network results are also compared with values calculated from the regression relationships of Yurioka and Suzuki based on the same database. Finally, an optimal network architecture (1 hidden layer. 4 hidden nodes and 40 training patterns) is suggested.
|Journal||Canadian Metallurgical Quarterly|
Chan, B. (Billy), Bibby, M. (Malcolm), & Holtz, N. (1995). Predicting haz hardness with artificial neural networks. Canadian Metallurgical Quarterly, 34(4), 353–356. doi:10.1016/0008-4433(94)00028-I