Ph.D. Proposal Presentation by Seungwon Shin
Friday, May 18, 2001

(Dr. Damir Juric, advisor)

"A Level Contour Interface Reconstruction Method for Three-Dimensional Simulations of Boiling Flows"


The detailed analysis of multiphase flows, including the deformable phase boundary, is a very important aspect of many emerging engineering technologies in a variety of disciplines. Computations of multiphase flows have been performed since the early days of computational fluid dynamics however the coupled solution of the Navier-Stokes equations in the presence of deforming phase boundaries is still quite difficult. The problem is even more challenging when combined fluid flow and heat transfer for phase change problems is considered.  Full three-dimensional simulations are computationally demanding and efficient, robust numerical methods are required as much as recent advances in computer processing speed.  The proposed research is an attempt to achieve a high resolution three-dimensional direct numerical modeling capability for general boiling flows.

The objectives of the proposed research are 1) to develop and validate a computational method that can be applied to general three-dimensional multiphase problems including phase change 2) to model complex three-dimensional boiling flows where repeated merging and breakup is an inherent feature of the interface dynamics 3) to develop a model for contact line dynamics for nucleate boiling and 4) to implement the three-dimensional method on parallel computer architectures.
In this proposal our recent work is shown. Three-dimensional multiphase flow and flow with phase change are simulated using a simplified method of tracking and reconstructing the phase interface. The new level contour reconstruction technique presented here enables front tracking methods to naturally, automatically and robustly model the merging and breakup of interfaces in three-dimensional flows. The method is designed so that the phase surface is treated as a collection of physically linked but not logically connected surface elements. Eliminating the need to bookkeep logical connections between neighboring surface elements greatly simplifies the Lagrangian tracking of interfaces, particularly for 3D flows exhibiting topology change.

Validation tests are conducted for drop oscillation and bubble rise. The susceptibility of the numerical method to parasitic currents is also thoroughly assessed. The capabilities of the new interface reconstruction method are also tested in a variety of two-phase flows without phase change. Three-dimensional simulations of bubble merging and droplet collision, coalescence and breakup demonstrate the new methodís ability to easily handle topology change by film rupture or filamentary breakup. We also perform simulations of 3D film boiling from a horizontal surface with multiple interacting bubbles.

The motivation for this new method is the modeling of complex three-dimensional boiling flows.  Tasks still to be undertaken include the development and implementation of a contact line model whereby the liquid-vapor interface is attached to a solid surface. This capability will allow numerical modeling of nucleate boiling, the most desirable boiling operating regime and the one of most engineering interest. In order to enable three-dimensional simulations with sufficient grid resolution to model all of the small scale features of boiling flows, the implementation of this method on multiprocessor parallel architectures is necessary. Once parallelized the code will be used in higher resolution simulations and analysis of boiling on flat and cylindrical surfaces.