Error-driven and predictive mesh control for adaptive simulations of incompressible two-phase flows

Zhang, Alvin
Thumbnail Image
Other Contributors
Sahni, Onkar
Shephard, M. S. (Mark S.)
Oberai, Assad
Kees, Christopher, E.
Issue Date
Mechanical engineering
Terms of Use
Attribution-NonCommercial-NoDerivs 3.0 United States
This electronic version is a licensed copy owned by Rensselaer Polytechnic Institute, Troy, NY. Copyright of original work retained by author.
Full Citation
adaptive approach becomes highly desirable. This is particularly the case for two-phase flows where interface evolution and dynamics are of interest.
These techniques are implemented and coupled with the Proteus toolkit, which includes an open-source multiphase flow solver. The overall implementation of the adaptive workflow is designed with parallel and large-scale applications in mind. To demonstrate its effectiveness, a number of complex two-phase flow cases are considered: a broad-crested weir case, a dam break case, and a 3D dam break case with an obstacle. Significant cost savings in computational time are shown for all three cases.
is derived from the variational multi-scale (VMS) paradigm and is then used to drive mesh adaptation. For problems with a highly dynamic interface, mesh adaptivity must be applied judiciously over time to control the computational cost, i.e., the rate or frequency of mesh adaptation must be controlled. Therefore, this work predictively defines the mesh resolution such that it is sufficient for a user-specified duration of time, where the trade-off is between the additional spread/width of regions with refined mesh resolution and the rate of mesh adaptation.
In this work, we present a robust, in-memory, and parallel mesh-adaptive workflow for incompressible two-phase flows that combines a two-phase flow solver with a posteriori error estimation and mesh adaptation including predictive mesh resolution. The error estimator
A common bottleneck in simulation-based workflows is the control of the discretization error, which dictates the cost and accuracy of the simulation. Since the location and evolution of important solution features is unknown in advance for complex problems of interest, an
December 2019
School of Engineering
Dept. of Mechanical, Aerospace, and Nuclear Engineering
Rensselaer Polytechnic Institute, Troy, NY
Rensselaer Theses and Dissertations Online Collection
CC BY-NC-ND. Users may download and share copies with attribution in accordance with a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License. No commercial use or derivatives are permitted without the explicit approval of the author.