Accurate Time-Dependent Computations and Reduced-Order Modeling for Multiphase Flows

In the present study, we initiate development of a non-iterative multiphase algorithm by enhancing the Pressure Implicit with Splitting of Operators (PISO) algorithm. The Gallium fusion problem, which is characterized by a solid-liquid phase front and natural convection effects, is employed as a test case for validation. The problem poses serious computational issues in form of a non-linear energy equation and a strong pressure-velocity-temperature coupling. The single-fluid modeling approach is adopted in conjunction with the enthalpy-based formulation for the temperature equation. The current algorithm computes the solution through a series of predictor-corrector steps with special treatment to achieve rapid convergence of the energy equation. The algorithm demonstrates an enhanced performance for the highly unsteady, chosen test problem. A reduced-order analysis of the simulated data is also performed by Proper Orthogonal Decomposition (POD). Specifically, impact of the constantly changing flow domain, and flow scales, on the POD implementation is highlighted.