Abstract
This work compares a few attempts based on Wavelet and Neural networks, for extracting the Jominy hardness profiles of steels directly from chemical composition. That is essentially a black-box modeling problem: Wavelet and Neural networks seem powerful, especially when compared with classical methods commonly found in literature. In particular, the paper proposes a multi-network architecture, where a first network is used as a parametric modeler of the Jominy profile, while a second one is used as a parameter estimator from the steel chemical composition. Several combinations of Wavelet and Neural networks have been compared. The paper also proposes an innovative method to train the activation function which significantly improves network performance.
Get full access to this article
View all access options for this article.