INTRODUCTIONSimulations are continuously becoming more and more important to predict the behaviour of materials, components and structures. By applying advanced material modelling as a supplement to experiments, it is often possible to shorten lead times during development considerably. One example is surface densification of sintered gears through gear rolling, see for instance [1][2][3]. By using finite element simulations it is possible to simulate the process and optimize for instance the tool geometry or the material stock on the work pieces.However, the accuracy of the simulations will depend on how well the material model predicts the material behaviour. Classic metal plasticity models typically rely on von Mises plasticity, and results in zero volumetric strain. This is clearly not correct for densification of PM steels, and special plasticity models for porous materials should be used instead.One model that is often used is the Gurson plasticity model [4], that is also found in some commercial finite element software such as Abaqus [5]. The Gurson yield criterion combines the von Mises stress with the hydrostatic stress, while accounting for different porosity levels, into a model suitable for porous materials. Fig. 1 illustrates the yield surfaces for different porosity levels, note that for zero porosity (f=0) the model is identical to classic von Mises plasticity.The purpose of this paper is first to investigate how the necessary input to the Gurson model can be calibrated from tensile tests on PM materials. Next, to simulate surface densification, indentation tests are made where a steel ball is pressed into the surface of a PM material. The experimental indentation tests are then compared to calculated ones, testing the application of the Gurson model.