| Abstract Scope |
Diffusion-controlled moving boundary problems underlie many industrially relevant materials processes, including carburization, transient liquid-phase bonding, and solid-state homogenization. Because analytical solutions exist only for simplified cases, numerical simulation is required to predict outcomes for real-world systems. Over the last 75 years, a wide variety of numerical methods have been developed to simulate the moving boundary problem; many early approaches had significant flaws, including failure to conserve mass. More robust algorithms addressing these issues have emerged over the last 30 years, but code implementing them is rarely available, hindering adoption. We have written Python implementations of several milestone algorithms and integrated them with the open-source diffusion and thermodynamics packages kawin and pycalphad. We compare the performance of these approaches on relevant test cases to identify the strengths and weaknesses of each, and demonstrate how they can guide and accelerate the design of a dissimilar diffusion bonding process. |