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

Grid resolution vs deep basin modeling

MA
Marina Amadori, modified 1 Year ago.

Grid resolution vs deep basin modeling

Youngling Posts: 2 Join Date: 8/7/18 Recent Posts

Hi, 

I am working on stationary wind-induced velocity profile (Coriolis force neglected) in un-stratified box domains, constant turbulence model and no thermal fluxes.  I am comparing the numerical results from boxes with different depths with the theoretical solution for the steady state (Heaps 1984). 

I verified that equilibrium is achieved (I am performing 60 days simulations) and I found good agreement for shallow boxes (H = 10 m, 30 m, 50 m, see Fig.v_res100m.png in the attachments). A mismatch between numerical and theoretical solution arises when I simulate deep boxes (H = 300 m, H = 500 m, always in Fig.v_res100m.png). 

As a first attempt I thought it was a time step problem and I reduced it to meet the conditions in Table 10.1 (FLOW manual) but I didn’t get any improvement.Then I tried changing the horizontal grid resolution and I got the results in Fig.v_allres.png (in the attachments), where I found that finer horizontal resolutions (DX = 20 m, 50 m, 100 m, square cells) lead to worse results. I got the best match with DX = 2000 m but still I am not there. What if I choose DX = 5000 m?

Now, my questions are:

1. Why are coarser resolutions giving better result?

2. If the answer is the "ADI accuracy" in Table 10.1, why is reducing time step not providing any change in the solution?

3. Is there a clear/formulated criterion for choosing the proper grid size according to the depth of the basin?

4. How is it this criterion related to the choice of the background vertical eddy viscosity (I am modeling with constant turbulence model, nuz = 0.0373 m2/s)?

 

Thank you!

Marina


 

Richard Measures, modified 1 Year ago.

RE: Grid resolution vs deep basin modeling

Jedi Knight Posts: 178 Join Date: 3/23/11 Recent Posts

Hi Marina,

Interesting question - Well explained and I really like your graphs.

Would you be able to give some more information about your vertical resolution e.g. how many layers do you have? Do you have the same number of layers when you increase the depth or do you have more vertical layers for the deeper models so the layer thickness is the same?

I don't have a lot of experience with constant vertical eddy viscosity but from my experience calibrating constant horizontal eddy viscosity I expect it could be quite sensitive to vertical resolution (the selection of appropriate horizontal eddy viscosity is sensitive to horizontal resolution). I expect that thinner vertical layers would require a smaller eddy viscosity to get the same result?

That doesn't help explain how horizontal resolution is influencing the vertical profile though! Have you checked to see if there are boundary effects propagating into your model horizontally? These could be affected by horizontal resolution and would be more significant for deeper domains. Maybe your horizontal eddy coefficient is too large for the models where you have a small cell size (I find the default horizontal eddy viscosity value is only suitable for very large cell sizes). As you are focused on vertical profiles I would try setting the horizontal eddy viscosity very small (but non-zero as that can cause weird instabilities) e.g. 0.001. Also you could try increasing the spatial size of the domain to see if that has any effect (i.e. reducing any horizontal boundary effect).

Good luck, I look forward to finding out what the problem was when you do solve it!

Cheers,

Richard

YJ
Yoan Jericko, modified 6 Months ago.

RE: Grid resolution vs deep basin modeling

Youngling Posts: 16 Join Date: 9/26/19 Recent Posts
Marina Amadori:

<p>Hi, </p>
<p>I am working on stationary wind-induced velocity profile (Coriolis
force neglected) in un-stratified box domains, constant turbulence
model and no thermal fluxes.  I am comparing the numerical
results from boxes with different depths with the theoretical solution
for the steady state (Heaps 1984). </p>
<p>I verified that equilibrium is achieved (I am performing 60 days
simulations) and I found good agreement for shallow boxes (H = 10 m,
30 m, 50 m, see Fig.v_res100m.png in the attachments). A mismatch
between numerical and theoretical solution arises when I simulate deep
boxes (H = 300 m, H = 500 m, always in Fig.v_res100m.png). </p>
<p>As a first attempt I thought it was a time step problem and I reduced
it to meet the conditions in Table 10.1 (FLOW manual) but I didn’t get
any improvement.Then I tried changing the horizontal grid resolution
and I got the results in Fig.v_allres.png (in the attachments), where
I found that finer horizontal resolutions (DX = 20 m, 50 m, 100 m,
square cells) lead to worse results. I got the best match with DX =
2000 m but still I am not there. What if I choose DX = 5000 m?</p>
<p>Now, my questions are:</p>
<p>1. Why are coarser resolutions giving better result?</p>
<p>2. If the answer is the &quot;ADI accuracy&quot; in Table 10.1, why
is reducing time step not providing any change in the solution?</p>
<p>3. Is there a clear/formulated criterion for choosing the proper grid
size according to the depth of the basin? filezilla uc browser rufus</p>
<p>4. How is it this criterion related to the choice of the background
vertical eddy viscosity (I am modeling with constant turbulence model,
nuz = 0.0373 m2/s)?</p>
<p> </p>
<p>Thank you!</p>
<p>Marina<br />
<br />
<br />  </p>

Good luck, I look forward to finding out what the problem was when you do solve it!
JS
john smith, modified 5 Months ago.

RE: Grid resolution vs deep basin modeling

Youngling Posts: 1 Join Date: 10/2/19 Recent Posts
Marina Amadori:

<p>Hi, </p>
<p>I am working on stationary wind-induced velocity profile (Coriolis
force neglected) in un-stratified box domains, constant turbulence
model and no thermal fluxes.  I am comparing the numerical
results from boxes with different depths with the theoretical solution
for the steady state (Heaps 1984). </p>
<p>I verified that equilibrium is achieved (I am performing 60 days
simulations) and I found good agreement for shallow boxes (H = 10 m,
30 m, 50 m, see Fig.v_res100m.png in the attachments). A mismatch
between numerical and theoretical solution arises when I simulate deep
boxes (H = 300 m, H = 500 m, always in Fig.v_res100m.png). </p>
<p>As a first attempt I thought it was a time step problem and I reduced
it to meet the conditions in Table 10.1 (FLOW manual) but I didn’t get
any improvement.Then I tried changing the horizontal grid resolution
and I got the results in Fig.v_allres.png (in the attachments), where
I found that finer horizontal resolutions (DX = 20 m, 50 m, 100 m,
square cells) lead to worse results. I got the best match with DX =
2000 m but still I am not there. What if I choose DX = 5000 m?</p>
<p>Now, my questions are:</p>
<p>1. Why are coarser resolutions giving better result?</p>
<p>2. If the answer is the &quot;ADI accuracy&quot; in Table 10.1, why
is reducing time step not providing any change in the solution?</p>
<p>3. Is there a clear/formulated criterion for choosing the proper grid
size according to the depth of the basin?</p>
<p>4. How is it this criterion related to the choice of the background
vertical eddy viscosity (I am modeling with constant turbulence model,
nuz = 0.0373 m2/s)?</p>
<p> </p>
<p>Thank you!</p>
<p>Marina<br />
<br />
<br />  </p>
I also have this question, if someone has a solution, please tell me
EC
eliot clake, modified 3 Months ago.

RE: Grid resolution vs deep basin modeling

Youngling Posts: 7 Join Date: 10/27/19 Recent Posts
Good luck, I look forward to finding out what the problem was when you do solve it!

jiofi.local.html tplinkloginis it down