Massively Parallel Phase-Field Simulations for Ternary Eutectic Directional Solidification

Martin Bauer,Johannes Hötzer,Philipp Steinmetz,Marcus Jainta,Marco Berghoff,Florian Schornbaum,Christian Godenschwager,Harald Köstler,Britta Nestler,Ulrich Rüde

DOI: https://doi.org/10.48550/arXiv.1506.01684

2015-06-05

Abstract:Microstructures forming during ternary eutectic directional solidification processes have significant influence on the macroscopic mechanical properties of metal alloys. For a realistic simulation, we use the well established thermodynamically consistent phase-field method and improve it with a new grand potential formulation to couple the concentration evolution. This extension is very compute intensive due to a temperature dependent diffusive concentration. We significantly extend previous simulations that have used simpler phase-field models or were performed on smaller domain sizes. The new method has been implemented within the massively parallel HPC framework waLBerla that is designed to exploit current supercomputers efficiently. We apply various optimization techniques, including buffering techniques, explicit SIMD kernel vectorization, and communication hiding. Simulations utilizing up to 262,144 cores have been run on three different supercomputing architectures and weak scalability results are shown. Additionally, a hierarchical, mesh-based data reduction strategy is developed to keep the I/O problem manageable at scale.

Distributed, Parallel, and Cluster Computing,Computational Physics

What problem does this paper attempt to address?

This paper aims to solve the problem of microstructure formation during ternary eutectic directional solidification. Specifically, the author attempts to simulate the microstructure evolution of ternary eutectics (such as Ag - Al - Cu alloys) during directional solidification through an improved phase - field model. The following are the main problems that this paper attempts to solve:

1. **The influence of microstructure on macroscopic properties**:
- Microstructure significantly affects the macroscopic mechanical properties of metal alloys. In order to predict and optimize these macroscopic material parameters, an in - depth understanding of the microstructure formation process is required.

2. **Limitations of existing models**:
- Most previous studies used relatively simple phase - field models or smaller simulation areas. These models cannot accurately capture the complex physical phenomena during ternary eutectic directional solidification.

3. **Computational complexity and efficiency**:
- The new phase - field model introduces temperature - dependent diffusion concentrations, which makes the calculation very intensive. Therefore, an efficient parallel computing framework needs to be developed to handle large - scale simulations.

4. **Data processing and storage**:
- Large - scale parallel simulations will generate a large amount of output data. How to efficiently process and store this data is a challenge. The author proposes a hierarchical, grid - based data reduction strategy to address this problem.

### Solutions

To solve the above problems, the author has taken the following measures:

- **Improved phase - field model**: A new free - energy formula, especially the grand potential formulation, is introduced to describe the concentration evolution more accurately.

- **Efficient parallel computing framework**: The waLBerla framework is used to achieve large - scale parallel computing, making full use of the performance of modern supercomputers. The computational efficiency is improved through various optimization techniques (such as buffering techniques, explicit SIMD kernel vectorization, and communication hiding).

- **Data processing optimization**: A hierarchical grid data reduction strategy is developed to ensure that the I/O problem is controllable in large - scale simulations.

### Simulation results

Finally, the author shows good agreement between their simulation results and experimental two - dimensional micrographs and three - dimensional tomography results, verifying the effectiveness and accuracy of the new model.

In summary, this paper successfully solves the complex simulation problem of microstructure formation during ternary eutectic directional solidification through an improved phase - field model and an efficient parallel computing method, and provides a powerful tool for research in related fields.