Search

A Cartesian grid solver is developed that is capable of simulating the convection-dominated melting processes in a latent-heat thermal storage device (TSD). The Navier-Stokes equations are solved using a dynamically refined mesh. The phase boundary is tracked using the enthalpy method. Conjugate heat transfer is calculated with a strongly coupled implicit scheme. The approach does not require the creation of a geometry-specific grid, and so allows for efficient prototyping of different complex geometric designs. Systematic benchmarking of the results against other numerical approaches is conducted, followed by tests of two basic prototypes for the design of a TSD. [ABSTRACT FROM PUBLISHER]

We present and validate a numerical technique for computing dendritic growth of crystals from pure melts. The solidification process is computed in the diffusion-driven limit. The governing equations are solved on a fixed Cartesian mesh and a mixed Eulerian-Lagrangian framework is used to treat the immersed phase boundary as a sharp solid-fluid interface. A conservative finite-volume discretization is employed which allows the boundary conditions to be applied exactly at the moving surface. The results from our calculations are compared with two-dimensional microscopic solvability theory. It is shown that the method predicts dendrite tip characteristics in good agreement with the theory. The sharp interface treatment allows discontinuous material property variation at the solid-liquid interface. Calculations with such discontinuities are also shown to produce results in agreement with solvability and with other sharp interface simulations. [ABSTRACT FROM AUTHOR]