3-d layered model - slope in more than 1 direction

A discussion forum for HYDRUS users. HYDRUS is a software package for simulating water, heat and solute movement in one-, two- and three-dimensional variably saturated media. Happy Posting!
Post Reply
jeff_k
Posts: 3
Joined: Fri Aug 26, 2022 10:01 pm
Location: USA

3-d layered model - slope in more than 1 direction

Post by jeff_k » Mon May 22, 2023 7:42 pm

I'm trying to build a 3d layered domain model to simulate recharge basins. Everything is fine: building the model, generating the mesh, defining boundary and initial conditions, running the model. The model has some irregular openings/interior surfaces to define mesh refinements.

Now, I'd like to skew the model geometry to represent a sloping water table. I can use the Slope... command to do so; with the slope defined by the x- or y- coordinate, and a non-zero value for the z-coordinate slope angle, I get the result I want.

The problem is, the direction of slope is at a 45 degree angle to the model grid. I can achieve this by running Slope... twice: once with slope defined by the x-coordinate, then with slope defined by the y-coordinate. But, after doing this, the extrusion that creates the 3-d domain only recognizes the interior openings, not the interior surfaces. That is, I get holes in my 3-d domain. The interior surfaces that used to be defines as "Base Surface" are now defined as "Planar". FWIW, this is only a problem with irregularly shaped interior openings - rectangular ones are retained as part of the base surface.

Again, it works fine when only applying Slope once on one axis. I also tried rotating the geometry 45 degrees, applying "Slope...", and rotating back. After the second rotation the "Base Surfaces" are converted to "Planar" and I have the same problem.

If that makes sense, are there any suggestions? I'd prefer not to work with a rotated grid as I have several monitoring locations where I need to make model predictions.

Thanks

Mirek
Posts: 527
Joined: Sat Mar 16, 2002 4:00 pm
Location: Czech Republic
Contact:

Re: 3-d layered model - slope in more than 1 direction

Post by Mirek » Tue May 23, 2023 5:42 am

I'm not sure what the problem is. Can you attach your project (before and after rotation) as a ZIP archive? Alternatively, you can send it to support@pc-progress.com.

jeff_k
Posts: 3
Joined: Fri Aug 26, 2022 10:01 pm
Location: USA

Re: 3-d layered model - slope in more than 1 direction

Post by jeff_k » Tue May 23, 2023 7:15 pm

Thank you for taking a look.

Project "...slope1" is after applying slope once - everything is fine. The screenshot in the .zip shows the Slope dialog settings.

Project "...slope2" is after applying slope twice - there are holes in the volume and I can't generate a mesh.

The 3 curves (ovals) in the center of the project are used both for defining grid refinements and setting boundary conditions. To me, it looks like Hydrus recognizes that they are contained in the surrounding surface (#22) for the purpose of creating openings, but they are not recognized as part of the base surface.

If I remove the openings from the slope2 project, I can define the boundary conditions on the mesh. That's an OK workaround, but it would be easier to define them on the geometry.

Thanks for any help. I'm also happy to take feedback if anything else in the model looks unusual. Other than that, It's working well for my purpose.
Attachments
HYD5_jk_slope2.zip
(4.78 MiB) Downloaded 739 times
HYD5_jk_slope1.zip
(4.79 MiB) Downloaded 639 times

Mirek
Posts: 527
Joined: Sat Mar 16, 2002 4:00 pm
Location: Czech Republic
Contact:

Re: 3-d layered model - slope in more than 1 direction

Post by Mirek » Wed May 24, 2023 9:57 am

Your domaim is relatively large (2 x 2 km) and the automatic detection of Base Surfaces stops working properly after its greater inclination. The solution is to turn off automatic detection and define Base Surfaces manually - see the first image. Then the FE=mesh can be generated correctly - see the second image.
Base_Surface.jpg
Base_Surface.jpg (389.64 KiB) Viewed 29296 times
Mesh.jpg
Mesh.jpg (829.46 KiB) Viewed 29296 times

jeff_k
Posts: 3
Joined: Fri Aug 26, 2022 10:01 pm
Location: USA

Re: 3-d layered model - slope in more than 1 direction

Post by jeff_k » Wed May 24, 2023 5:23 pm

Great! That was just what I was missing. Thank you very much.

Post Reply