Abstract
A neural network model has been developed on the basis of published experimental data. This allows the creep rupture strength of bainitic and martensitic electric power plant steels with compositions based on Fe–2·25Cr–1Mo and Fe–(9–12)Cr to be estimated as a function of chemical composition, heat treatment and time at temperature. This model, together with a variety of thermodynamic and kinetic calculations, has been used to propose two new alloys related to the 10CrMoW steel which in theory have outstanding creep rupture properties.