Open PhD position: “Improvement of a mean field model dedicated to the recrystallization simulation””

Countless products involved in our every-day life rely on vital metal parts. Optimizing these parts requires a knowledge of how material properties change during forming operations. Although the understanding of the underlying metallurgical phenomena has improved thanks to the continuous progress of experimental facil-ities, the interest for increasingly fine and predictive simulations has been recently growing. In this emerging context of “digital metallurgy”, the DIGIMU consortium has two main objectives. The first one is to develop an efficient multiscale numerical framework specifically designed to tackle such problems. The second one is to bring the corresponding numerical methods to an industrial level of maturity, by decreasing significantly their computational cost and by validating them against the industrial expertise existing in the DIGIMU consortium.

In this context, full field and mean field models dedicated to the simulation of recrystallization phenomena are developed. Full field approaches imply the description of the full topology of the considered microstructures (grains, phases...) whereas mean field models use approximated descriptions. Full field approaches imply the discretization of the microstructure where mean field models can be phenomenological or based on simplified equations predicting the evolutions of the microstructure attributes (mean values or even distribution of main caracteristics). If full field approaches remain more precise, their numerical cost is generally such that these methods cannot be envisaged at the scale of an industrial part. Here appears the interest of mean field meth-ods, whose numerical cost is much more reasonable and which can be used at the macroscopic scale of metal forming simulations.

The DIGIMU team develops both approaches in tandem, which allows mutual enrichment. Some main field equations can then be used in full field simulations and full field simulations can be seen as numerical tests to discuss/improve mean field models. In this context a new mean field model was recently developed [1]. If this model already proved its interest on some stainless steels undergoing discontinuous dynamic recrystallization (DDRX), it needs to be further improved and expanded.

Different items will be considered during this PhD:

• the equations dedicated to nucleation or appearance of ReX grains will be improved,
• post-dynamic and static mechanisms will be taken into ac-count,
• the range of validity if terms of strain and strain rate will be enlarged.

Moreover, the resulting developments will be prepared for integration in the FORGE® software package, that is a world-wide spread tool for thermomechanical simulation of industrial processes.