Estimation of Material Properties in Ferrous and AluminumAlloy Databases

Three numerical methods are proposed for estimating numerical properties in two metallic materials databases. When such methods are used on a complete database, it is found that the error on estimated properties is generally low for the majority of the properties (mechanical, thermal, electrical). However, errors of more than 100% in some cases are found for properties like fracture toughness, ductility, tensile strength or hardness. Those errors can be reduced when some details of the heat treatment and of the chemical composition of the alloys are introduced as properties in the database.