 |
 |
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
Home > SCI Research: Evaporation
Back to Computational PDEs
Evaporation
Models of complex interactions between point-source air pollutants and endemic ozone over Western Europe. The hole toward the left and middle of the mesh, around which several of the worst mesh elements occur, is the location of the point source of several of the pollutants. The evaporation technique was used to investigate the quality of the mesh in terms of individual pollutants, and this technique was useful in that it provided strikingly different visualizations for the different pollutants. |
We have investigated several meshes used in simulations by Tomlin and her colleagues and by Speares and Berzins, to see how the shape, volume, position, and orientation of the tetrahedra relate to the local and overall quality of the meshes. A poor quality region of a mesh is one that, for reasons we wanted to discover, represents a poor discretization of the space, and one for which the simplifying assumptions (such as piecewise linearity) introduce errors into the solution. Currently this problem is solved by finding, and then remeshing the bad areas, but at some point in the future, the goal is to be able to generate meshes without having to go in and fix areas afterwards.
We built an interactive environment to analyze these meshes, and with it we were able to find all the poorly-meshed regions. We also could pull apart the worst areas of the mesh to look at everything that might affect the quality there: the tetrahedra's shapes, volumes, and orientations; local solution values; and any sharp increases in solution value from one tetrahedron to the next across a shared face. By comparing bad regions and bad tetrahedra to good ones, we were able to determine that, for these particular meshes, it was generally the size and orientation of one or two faces of a single tetrahedron that caused the regional problems. This experiment gives us direct answers on how to fix these two meshes. It also gives us potentially useful insight into how to change the software algorithms used in generating these meshes so that we can successfully avoid the formation of these few critical faces in regions of the mesh where a fairly fine-grained resolution is required.
Here at the SCI Institute, finite element meshes are utilized to develop improved solutions to:
- Inverse electroencephalography (EEG) problems
- Forward and inverse magnetoencephalography (MEG) problems
- Inverse electrocardiography (ECG) problems
Institute researchers have also been involved in mesh generation research, such as the development of adaptive methods and adaptive mesh generators, and in mesh quality analysis.
Principal Researchers:
