Skip to content

Richards mechanics linear-linear shape functions#2843

Merged
endJunction merged 3 commits intoufz:masterfrom
endJunction:RichardsMechanicsLinearShapeFunctions
Mar 9, 2020
Merged

Richards mechanics linear-linear shape functions#2843
endJunction merged 3 commits intoufz:masterfrom
endJunction:RichardsMechanicsLinearShapeFunctions

Conversation

@endJunction
Copy link
Copy Markdown
Member

@endJunction endJunction commented Mar 5, 2020

Add ORDER to shape functions.
Allows linear-linear shape functions in the RichardsMechanics process. So far only quadratic-linear combinations were possible.

Compilation times (for the RM process) are skyrocketing, as well as memory usage while compilation.

  1. Feature description was added to the changelog
  2. Tests covering your feature were added?
  3. Any new feature or behavior change was documented?

@nagelt
Copy link
Copy Markdown
Member

nagelt commented Mar 5, 2020

I ran a test on the B3 saturation benchmark. The two-fold DOF reduction from 45803 (Q2P1) to 19803 (Q1P1) yielded a reduction in simulation time by about a factor of 4. In addition to the DOF reduction I reduced IP number from 4x4 to 2x2. Results show only a slight shift, simulation is stable against pressure oscillations (storage and drainage sufficient). This can be a feature in HM, THM, etc. also!

The constexpr value allows some decisions to be made
at compile time.
The reference files are converted from the second
order mesh results with the 'convertToLinearMesh'
utility.
Note, the sigma, and the NodalForces fields do not
compare well to the higher order results, but other
variables do very well.
sigma, probably due to extrapolation;
NodalForces because of the missing middle nodes which
would take some of the load.
@endJunction
Copy link
Copy Markdown
Member Author

FYI; Compilation time roughly doubles, so does the memory usage.

@endJunction endJunction marked this pull request as ready for review March 6, 2020 17:42
Copy link
Copy Markdown
Member

@wenqing wenqing left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good.

@endJunction endJunction merged commit 480e59e into ufz:master Mar 9, 2020
@endJunction endJunction deleted the RichardsMechanicsLinearShapeFunctions branch March 9, 2020 12:17
@ogsbot
Copy link
Copy Markdown
Member

ogsbot commented Jun 19, 2020

OpenGeoSys development has been moved to GitLab.

See this pull request on GitLab.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants