From mean field to polycrystal modelling through digital microscopy

Predicting microstructural evolution during thermomechanical processing remains a central challenge in metallurgical engineering. Traditional approaches range from phenomenological JMAK models, which provide only mean values with limited coupling between phenomena, to class models that track grain size distributions but require assumptions about grain shapes and neighborhood interactions. While computationally efficient, these methods struggle to capture the complex interplay of competing mechanisms that govern industrial forming operations.

This presentation introduces a digital microscope called DIGIMU®, a full-field polycrystal simulation software, which advances microstructure prediction through explicit topological description of grain structures. Unlike mean field approaches, this approach resolves physical neighborhoods, grain morphologies, and coupling between concurrent phenomena at the mesoscale—the ideal modeling scale for comprehensive studies and process optimization.

The solution framework addresses the complete spectrum of microstructural mechanisms encountered in industrial practice: grain growth driven by capillarity, Zener pinning by second-phase particles, strain-induced boundary migration, and multiple recrystallization modes including discontinuous dynamic (DDRX), post-dynamic (PDRX), static (SRX), and continuous dynamic recrystallization (CDRX). Recent developments incorporate heterogeneous hardening distributions, nuclei size variability, precipitate evolution through growth, dissolution and Ostwald ripening, and orientation-dependent grain boundary energies following the Read-Shockley formulation.

The polycrystal generator enables microstructure initialization from EBSD data or statistical distributions, while extensive analysis tools provide outputs directly comparable to experimental characterization. Industrial validation studies demonstrate predictive accuracy across diverse applications: Inconel 718 forging for aerospace components both sub-solvus and super-solvus, and 304L stainless steel rolling sequences. Results show excellent agreement between simulated and experimental grain sizes, recrystallized fractions, and microstructural morphologies across multiple process conditions, confirming DIGIMU®'s capability to bridge fundamental materials science with industrial process design requirements.