Inverse Heat Transfer Problems rely on temperature measurements for the estimation of unknown quantities (e.g. boundary or initial conditions, thermophysical properties); this kind of problems is classified as ill-posed. An application of the inverse analysis in the continuous casting process of steel is here presented. The aim is the estimation of the mould heat transfer starting from temperature measurements, recorded using thermocouples buried inside the copper mould wall. The mould is water-cooled to solidify the hot metal directly in contact with it. The direct stationary conduction problem was solved both on a 2D and a 3D domain. The inverse problem was solved using Gradient algorithms, Genetic Algorithms and SIMPLEX. For both geometries, a good agreement between numerical and experimental temperature values is observed; moreover, the 3D model provides a better estimate of the outlet temperature of the cooling water.
