TY - GEN

T1 - Neuromelt model for estimating mold flux melting behaviour

AU - Vargas Hermandez, M

AU - Mapelli, Carlo

AU - Cho, Jungwook

AU - Kölbl, Nathalie

AU - Marschall, Irmtraud

AU - Alloni, Marco

AU - Carli, Ricardo

PY - 2021/10/14

Y1 - 2021/10/14

N2 - The existing models and methods used to determine the melting temperature of the mold powders used in thecontinuous casting process remain inaccurate in the case of equations reported in current literature or considerfor commercial software only an equilibrium state. In this work, a new approach has been implemented usingneural networks, which will act as a “black box” to predict the melting temperature of a mold powder with agiven composition within an acceptable range. The proposed neural network will be working as a regressionneural model whose inputs will be the composition of each of the chemical species contained within thepowder. A database provided by a research net comprising multiple countries' research institutes will be fedas a training set for network learning. Such data comes from experimental measurements performed mainlythrough the high-temperature microscope test. The correct implementation and training of the network shouldprovide a new alternative to develop new products and verify existing products' melting properties. In futuremodels, further considerations should be made towards a better understanding of these phenomena, whichshould consider factors such as the formation of mineral phases, interaction among some specific componentsof the powder, or even the parameters used at the time of experimental measurement.

AB - The existing models and methods used to determine the melting temperature of the mold powders used in thecontinuous casting process remain inaccurate in the case of equations reported in current literature or considerfor commercial software only an equilibrium state. In this work, a new approach has been implemented usingneural networks, which will act as a “black box” to predict the melting temperature of a mold powder with agiven composition within an acceptable range. The proposed neural network will be working as a regressionneural model whose inputs will be the composition of each of the chemical species contained within thepowder. A database provided by a research net comprising multiple countries' research institutes will be fedas a training set for network learning. Such data comes from experimental measurements performed mainlythrough the high-temperature microscope test. The correct implementation and training of the network shouldprovide a new alternative to develop new products and verify existing products' melting properties. In futuremodels, further considerations should be made towards a better understanding of these phenomena, whichshould consider factors such as the formation of mineral phases, interaction among some specific componentsof the powder, or even the parameters used at the time of experimental measurement.

KW - Neural network

KW - mold powder

KW - liquidus temperature

M3 - Conference contribution

BT - Proceedings of the 10th European Conference on Continuous Casting

PB - Associazione Italiana di Metallurgia

ER -