intro story D-Flow FM

 

D-Flow Flexible Mesh

D-Flow Flexible Mesh (D-Flow FM) is the new software engine for hydrodynamical simulations on unstructured grids in 1D-2D-3D. Together with the familiar curvilinear meshes from Delft3D 4, the unstructured grid can consist of triangles, pentagons (etc.) and 1D channel networks, all in one single mesh. It combines proven technology from the hydrodynamic engines of Delft3D 4 and SOBEK 2 and adds flexible administration, resulting in:

  • Easier 1D-2D-3D model coupling, intuitive setup of boundary conditions and meteorological forcings (amongst others).
  • More flexible 2D gridding in delta regions, river junctions, harbours, intertidal flats and more.
  • High performance by smart use of multicore architectures, and grid computing clusters.
An overview of the current developments can be found here.
 
The D-Flow FM - team would be delighted if you would participate in discussions on the generation of meshes, the specification of boundary conditions, the running of computations, and all kinds of other relevant topics. Feel free to share your smart questions and/or brilliant solutions! 

 

=======================================================
We have launched a new website (still under construction so expect continuous improvements) and a new forum dedicated to Delft3D Flexible Mesh.

Please follow this link to the new forum: 
/web/delft3dfm/forum

Post your questions, issues, suggestions, difficulties related to our Delft3D Flexible Mesh Suite on the new forum.

=======================================================

** PLEASE TAG YOUR POST! **

 

 

Sub groups
D-Flow Flexible Mesh
DELWAQ
Cohesive sediments & muddy systems

 


Message Boards

Morphological Model Setup | Area of Interest

OL
Oliver Lojek, modified 7 Years ago.

Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Hi there,

I am setting up a model to simulate morphological long term processes.

Currently I am looking at detailed meshing of tidal channels which end within the domain. Does anybody have any suggestions on how to mesh this?


Kind regards,


Oliver
Qinghua Ye, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Council Member Posts: 612 Join Date: 3/2/11 Recent Posts
Hi Oliver,

Looks like the area you are interested in has a tidal channel flowing in the area in an angle.

I would suggest a model covering bigger area first. Then you may nest the smaller model to get correct boundary conditions. You may have to tune the model very careful about sediment transport at very shallow water.... The Flooding scheme is definitely sth you should try. Maybe also cutting cell technique.

Regards,

Qinghua
OL
Oliver Lojek, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Hi Qinghua,


you are correct, there is a tidal channel flowing through my area of interest.

However, I need to get the model to run properly first before I can add Sediment and so on - currently I seem to get a lot of CFL related issues.
I have removed all physical processes and so far just have a tidal boundary forcing and that is it. I am modelling one week which takes about 2 minutes on my small server.

The GUI suggested that the resolution was fairly high and the aspect ratio was fine as was the orthogonality. At some point however the Model seems to become instable and suggests Time steps of -0.147E01 Minutes.

I reduced the timestep from 5 min to 2.5, 1, 0.5 and 0.25 min. This usually happens around 80 to 90 % of the calculation.

*** MESSAGE Using UNESCO density formulation by default
*** MESSAGE Momentum solver cyclic method is specified
*** MESSAGE DRYFLP and DPSOPT both specified in MD-file. Using DPSOPT: MAX
*** MESSAGE Transport solver cyclic-method method is specified
*** WARNING File will not be created - Inconsistent comm. file times
*** MESSAGE Nodal factors updated at 2007-01-01 00:00:30.0000
*** MESSAGE Courant number for U-velocity in U-point equals     1.04 for (m,n,k) = (1,21,1),  at nst = 10. Advised time step: 0.966E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.08 for (m,n,k) = (1,21,1),  at nst = 11. Advised time step: 0.928E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.02 for (m,n,k) = (2,20,1),  at nst = 11. Advised time step: -.980E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.18 for (m,n,k) = (2,17,7),  at nst = 11. Advised time step: 0.848E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.10 for (m,n,k) = (3,16,7),  at nst = 11. Advised time step: -.913E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.16 for (m,n,k) = (1,21,1),  at nst = 12. Advised time step: 0.859E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.12 for (m,n,k) = (3,16,7),  at nst = 12. Advised time step: -.896E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.40 for (m,n,k) = (2,18,9),  at nst = 12. Advised time step: 0.713E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.24 for (m,n,k) = (3,16,8),  at nst = 12. Advised time step: -.807E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.29 for (m,n,k) = (2,16,8),  at nst = 13. Advised time step: 0.775E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.24 for (m,n,k) = (3,16,7),  at nst = 13. Advised time step: -.806E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.37 for (m,n,k) = (2,16,9),  at nst = 13. Advised time step: 0.729E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.32 for (m,n,k) = (3,15,1),  at nst = 13. Advised time step: -.758E+00 minutes.
*** MESSAGE Courant number for U-velocity in U-point equals     1.33 for (m,n,k) = (3,16,1),  at nst = 14. Advised time step: 0.754E+00 minutes.
*** MESSAGE Courant number for U-velocity in V-point equals     1.28 for (m,n,k) = (3,15,1),  at nst = 14. Advised time step: -.780E+00 minutes.


I have simulated waves with a 3G spectral model coupled to it on unstructured grids and unfortunately am not too familiar with curvilinear grids - are there any help resources for Delft3d on how to generate an appropriate mesh?
OL
Oliver Lojek, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
I just tried to check the interpolation in RGFGRID. But RGFGRID crashes every time I click the button.
Does this routine only support a limited number of interpolated points aka. did I set the resolution too high for it to work recursively?


Kind regards,


Oliver
OL
Oliver Lojek, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Okay - for what ever reason the previous GUI version (4.00.01) works the Triangulation Check independant of the size.
For the homogeneously fine grid I got the following values.

Thus the grid comprises 235 X 91 = 21'385 Cells - I usually introduce 10 vertical layers given a typical water level of about 4 m. Eventually it holds 213'850 Cells to be calculated per step.
Reading the manual and the modelling guidelines I found that a typical grid comprises 20'000 cells. Therefore, I guess I will have to coarsen the grid quite a bit and add some nesting for the area of interest as suggested by Qinghua earlier.


Interpolation Check by Reverse Interpolation
*************************************************
Mean Interpolation Error : 0.823365E-01 [m]
Mean Abs Interpolation Error : 0.365152E+00 [m]
Maximum Error Value : 0.528059E+03 [m]
Number of Samples Checked : 80615 [-]

Actual and Maximum Dimensions of Data
****************************************************************
Data Type : Actual Maximum
Grid M-Direction : 235 235
Grid N-Direction : 91 91
Samples : 102370 102445
Land Boundary : 0 1
Polygon : 0 1
Depth Points in Polygon : 21381 21385
Second Depth Points in Polygon : 0 21385
Samples in Polygon : 102370 102445

Volumes and Areas below Reference Level
*************************************************
Volume Grid below 0.0[m]: 4.578562E+08[m3]
Grid Area below 0.0[m]: 1.368022E+08[m2]
Polygon Area : 0.000000E+00[m2]
Dmitry Kushnir, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 103 Join Date: 11/9/12 Recent Posts
Hi Oliver,

I think that modeling guide about computational grid refer to 20'000 cells per one layer. I'm now running simulations on a domain consisting of 244x52 grid cells and with 15 sigma-layers this will be similar to your model. For now it was running without errors. Probably you don't need to coarse your grid. Have you tried to check the courant number in quickin for your bathymetry and time step?

Regards,

Dmitry
OL
Oliver Lojek, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats (Answer)

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Hi Dmitry,


good to know - I was wondering already seemed way to small.

I had the model running but it gave me multiple errors on too harsh velocity changes. I checked the courant numbers but QUICKIN suggested negative values.

e.G.:

*** WARNING Velocity change too high > 5.00 m/s (per 0.5 DT) after 18020
*** timesteps in the following points:
*** WARNING Velocity change too high > 5.00 m/s (per 0.5 DT) after 18021



That is why I assumed I better coarsen the grid a bit. I did coarsen it and othogonalized it and checked the triangulation and so forth - now I get courant numbers around 0.276 all over the domain (see PNG).
I'll reduce the runtime and see if the inconsistencies are gone.


Regards,


Oliver
Qinghua Ye, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Council Member Posts: 612 Join Date: 3/2/11 Recent Posts
Hi Oliver and Dmitry,

Yes, I would say to decrease the time step, to make the grid coarser, or to increase the grid orthogonality, and so on, could be helpful to remove the warning about the big water level gradient and velocity gradient.

Delft3D used a lot of numerical tips to stabilize the solution. Officially the Cr number can be as big as 10. I would not trust the results if the Cr number is bigger than that.

But there is no general solution to solve everybody's modeling problem. So in that sense, we still have to tune the model and modelling stays as an art. The modeller stays as artists. emoticon

Keep us update and see what comes out of your work as an artist...

Regards,

Qinghua
OL
Oliver Lojek, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Hi Qinghua,

it worked - the model runs smoothly and the output looks quite good.

Regards,

Oliver
OL
Oliver Lojek, modified 7 Years ago.

Morphological Model Setup | Tidal Flats | Domain size

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
And a totally different question in general - are 8 km of tidal flat in both directions enough for the tide to realistically simulate the morphological processes regarding the different modes of entrainment?

Given is fine, non-cohesive sediment < 63 µm comprise between 100 - 70 % of sediment probes taken on site.

Average water level amounts to 5.00 m with a tidal amplitude of 2.75 m.

Current velocities range between 0.3 - 1.0 m/s within the domain over the tide.


Regards,


Oliver
Qinghua Ye, modified 7 Years ago.

RE: Morphological Model Setup | Tidal Flats | Domain size

Jedi Council Member Posts: 612 Join Date: 3/2/11 Recent Posts
Hi Oliver,

I would say it should be ok, but still depends on the situation.

Regards,

Qinghua
OL
Oliver Lojek, modified 7 Years ago.

Morphological Model Setup | Area of Interest

Jedi Knight Posts: 111 Join Date: 4/30/13 Recent Posts
Hi there,


I am progressing on the model and started looking at the NESTING possibilities for the Model.

I have supplemental high resolution bathymetry data for the area of interest in my domain.

Currently I am a little puzzled as to how one could mesh it in order for it to work smoothly.

I attached an overview of the area. It features a coastline with artificial structures and adjacent tidal flats.

I am thankful for suggestions on how to mesh this the best possible way.

Cheers,


Oliver