Application of a neural network in modelling of hardenability of constructional steels

摘要

The paper presents a new, original method for hardenability modelling of the alloy constructionalsteels, both carburized and heat-treatable. The automatic steel classification based on steel chemicalcomposition was employed in this method. Three neural network based models of hardenabilitycurves pertaining to each of the investigated steel groups were employed to obtain calculation resultsas close as possible to experimental results. The tests performed on about 1500 experimentalhardenability curves gave, for the various heats indicated, a significant reliability of calculations madeaccording to the presented method. Knowledge of the hardness distribution, depending on thedistance from the specimen face and the distribution of the cooling rate of specimen cooled from theface, described by the Jominy hardenability curve, makes the rational selection of the alloyconstructional steels possible for the heat-treated or thermo-chemically treated machine parts.Therefore, the new method of the hardenability curve modelling presented here may be an importantelement of the computer based systems for selection of alloy constructional steels for machine parts.