Phase field as a front propagation method for modeling grain growth in additive manufacturing
Murali Uddagiri 1,∗, Pankaj Antala 2, and Ingo Steinbach 1 1 ICAMS, Ruhr University Bochum, Universitätsstr. 150, 44801 Bochum, Germany 2 Department of Materials Engineering, Katholieke Universiteit Leuven, Arenberg 44, Leuven, 3001, Belgium · 2026
Read Phase field as a front propagation method for modeling grain growth in additive manufacturing by Murali Uddagiri 1,∗, Pankaj Antala 2, and Ingo Steinbach 1 1 ICAMS, Ruhr University Bochum, Universitätsstr. 150, 44801 Bochum, Germany 2 Department of Materials Engineering, Katholieke Universiteit Leuven, Arenberg 44, Leuven, 3001, Belgium (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-06627.txt.
Introduction
Uddagiri and team develop a mid-scale grain-envelope model for grain growth under additive manufacturing conditions, showing dendritic grains' surface through a diffuse interface. A heat-conduction model with moving heat sources and latent-heat release supplies the thermal field, while shell spread follows a tiny-melt-based kinetic law. Tested in two- and three-dimensional simulations, the model is applied to parameter effects on grain mix evolution. The authors say it offers an efficient, forward-looking approach for grain growth during multi-pass, multi-layer build-up.
Catalog record
| CREATOR | Murali Uddagiri 1,∗, Pankaj Antala 2, and Ingo Steinbach 1 1 ICAMS, Ruhr University Bochum, Universitätsstr. 150, 44801 Bochum, Germany 2 Department of Materials Engineering, Katholieke Universiteit Leuven, Arenberg 44, Leuven, 3001, Belgium |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23775 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.06627
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Phase field as a front propagation method for modeling grain growth in additive manufacturing Murali Uddagiri 1,∗, Pankaj Antala 2, and Ingo Steinbach 1 1 ICAMS, Ruhr University Bochum, Universitätsstr. 150, 44801 Bochum, Germany 2 Department of Materials Engineering, Katholieke Universiteit Leuven, Arenberg 44, Leuven, 3001, Belgium Abstract Abstract: A mesoscopic grain-envelope model applying a phase-field front-propagation method is developed to simulate grain growth under additive manufacturing process conditions. The envelope represents the outer surface of dendritic grains through a diffuse interface. While a modified heat-conduction model that incorporates moving heat sources and latent-heat release provides the evolution of local thermal field. Envelope propagation is determined from microscopic-solvability-based kinetic law. The model is validated through two- and three-dimensional simulations and subsequently applied to examine the influence of material and process parameters on microstructure evolution. The results demonstrate that the proposed mesoscopic model offers an efficient and predictive approach for modeling grain growth during multi-pass and multi-layer build-up in additive manufacturing. Keywords: Phase field, Mesoscale, Graim growth, Texture, Additive Manufacturing. 1 Introduction Additive manufacturing (AM), particularly powder-bed fusion AM (PBF-AM), has gained significant attention for producing complex metallic components with tailored microstructures and mechanical properties. Microstructure development during PBF-AM is highly sensitive to process parameters and strongly influenced by rapid solidification driven by large thermal gradients and high interface velocities. These conditions promote non-equilibrium microstructural features, including epitaxial columnar grains with strong \langle 001\rangle fiber texture, which contribute to mechanical anisotropy [ 10 ] . Repeated thermal cycling during layer-by-layer deposition further induces phase transformations and intermetallic formation, complicating microstructure–property relationships. Process parameters such as beam power, scan speed, and layer thickness directly affect local solidification behavior, and variations in heat input across AM systems lead to significant differences in grain morphology even for the same alloy [ 3 , 8 ] . These considerations highlight the need for computationally efficient and predictive microstructure models to support process optimization and material design in AM. A number of numerical techniques including front-tracking, phase-field (PF), level-set (LS), lattice Boltzmann (LB), Potts kinetic Monte Carlo, cellular automata (CA), and Johnson–Mehl–Avrami–Kolmogorov-type approaches have been employed, each offering distinct advantages and limitations depending on computational cost and physical fidelity [ 6 , 19 ] . Cellular automata (CA) models pioneered mesoscale simulations of competitive grain growth and texture evolution at modest computational cost [ 5 , 18 ] . Their coupling with finite element thermal solvers in cellular automata–finite element (CAFE) frameworks has been widely used for AM microstructure modeling, capturing epitaxial grain growth, competitive selection, and texture evolution [ 24 , 18 , 12 ] . More recent CA implementations incorporate temperature-gradient tracking, grain competition, and probabilistic nucleation. For example, [ 23 ] combines a finite-difference thermal solver for electron-beam melting with CA to predict grain structures under varying process conditions. While efficient, CA-based methods rely on empirical rules for interface kinetics and cannot inherently capture curvature effects, solute redistribution, or thermodynamically consistent interface motion. Their predictive accuracy diminishes when solidification becomes strongly transient, such as under keyhole-mode melting or rapid inter-layer reheating. Phase-field (PF) methods provide a thermodynamically rigorous description of diffuse interfaces and have become a leading approach for modeling alloy solidification [ 15 , 16 ] . PF models enable quantitative simulation of dendritic growth, solute redistribution, and grain morphology in dilute and multicomponent alloys. Numerous studies have applied PF to AM-like rapid solidification, including melt-pool simulations in Ni-based superalloys, Ti–6Al–4V, steels, and Al alloys. In our previous work [ 21 ] , a multi-phase-field approach with constant heat extraction was used to investigate layer remelting, epitaxial growth, and solute-driven nucleation. However, AM microstructures are dictated by the tight coupling of heat transfer, melt flow, and solidification, motivating the development of concurrent multiscale frameworks. For example, thermal–lattice Boltzmann approaches have been used (in 2D) to jointly solve fluid flow and heat transfer [ 11 ] , while [ 2 ] employed an analytical Rosenthal solution to couple PF simulations with highly non-equilibrium solidification conditions. Because the temperature field forms the foundation of any solidification model, AM simulations must resolve rapid, transient thermal evolution within small melt pools, which imposes severe computational constraints. Hierarchical strategies therefore remain common, in which macroscopic CFD or finite element models provide thermal gradients and cooling rates for mesoscale PF or CA simulations [ 7 , 1 , 17 ] . Although computationally feasible, these approaches require assumptions about interface kinetics and generally cannot resolve dendritic-scale curvature or solutal effects. Despite significant progress, fully resolving solute diffusion and dendrite morphology over millimeter-scale AM melt pools remains computationally prohibitive. To address this, mesoscopic envelope models were proposed as an efficient alternative [ 13 , 14 ] . These models represent a dendritic grain by a smooth envelope surface that tracks the outer boundary of active dendrite tips while incorporating physically meaningful growth kinetics. When integrated with PF front propagation, envelope models enable computationally tractable simulations of competitive grain growth and texture evolution under realistic AM thermal fields. Building upon this concept, the present work develops a mesoscopic grain-envelope model coupled with a phase-field front-propagation method for PBF-AM. The framework incorporates a modified heat-conduction formulation including moving heat sources and latent heat, and employs a microscopic-solvability-based kinetic relation to compute interface velocities from local undercooling. The model is validated through two- and three-dimensional simulations and used to investigate the influence of material properties and temperature gradients on microstructural selection during AM. 2 Model description The phase-field based grain-envelope model, originally introduced by on of the authors and co-workers [ 13 , 14 ] provides a scale-bridging approach for dendritic solidification. This approach connects the microscale (dendrite tip), grain-internal morphology, and the mesoscale interaction of multiple grains. As shown in the Fig. 1 , instead of resolving multiple dendrite arms within a grain, the entire multi-branched dendritic structure of a grain is represented by a smooth envelope surface that connects the active dendrite tips. The region inside the envelope is assumed to be either fully solid or mushy, while the envelope surface evolves according to dendrite-tip growth kinetics. Figure 1 : Schematic illustration of a) Envelop around dendrite branches and b) Equivalent envelop surface, adapted from [ 13 ] In the present model framework, which is adopted from the original envelope model, the evolving temperature field is obtained from a transient heat conduction equation that incorporates the moving heat source and the latent heat released inside the mushy zone, following the treatment in [ 22 ] . The envelop surface is not treated as an iso-surface but rather in terms of diffuse phase field interface and velocity of the envelop surface is calculated by drawing an analogy between equation describing evolution of Phase-field and Gibbs-Thomson relation. The envelope motion is driven by the local undercooling, which contains thermal and solutal components. A full solution of solute diffusion at the scale of the dendrite tip would require spatial and temporal resolutions far below those feasible for melt pool simulations. Instead, a reduced microstructural description is employed, based on microscopic solvability theory for dendritic growth in dilute binary alloys. This model provides the dendrite-tip velocity as a function of total undercooling, which is subsequently converted into a normal envelope velocity for the phase-field