AbstractA model for simulating the process of mould filling in castings is presented. Many defects in a casting have their origins at the filling stage. Numerical simulation of this process can be of immense practical benefit to the foundry industry, however a rigorous analysis of this process must model a wide range of complex physical phenomena. In order to contain the costs and complexity that would be necessary for such a model, certain simplifying assumptions have been made. These assumptions limit the scope of this model to only predicting realistic thermal fields during the filling process.A laminar regime has been assumed for the flow field, which is obtained by solving the incompressible Navier–Stokes equations using a velocity–pressure segregated semi‐implicit finite element method. The free metal surface is predicted by advecting a pseudo‐concentration function via the computed flow field. This involves an explicit finite element solution of a pure advection equation. The thermal field is calculated by solving the convective–diffusive energy equation by an explicit finite element method using the computed flow field and the location of the free surface. All the advection terms are discretized using a Taylor–Galerkin method. The interface between the metal and mould is modelled using special interface elements.The model is demonstrated by solving practical example problems. The results show that a sharp thermal front is maintained during the course of filling without excessive diffusion. The heat diffusion in the mould can be controlled by varying the metal mould heat transfer coefficient.