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

Thatcher - Harleman time lag

KM
Konstantinos Matsoukis, modified 2 Years ago.

Thatcher - Harleman time lag

Youngling Posts: 3 Join Date: 1/19/17 Recent Posts
Hi,

I am trying to model salinity flow in a rectilinear grid with DELFT3D-FLOW. The model is forced only by a river discharge coming from the western boundary.
I want my north,south and offshore boundaries to be open so in my first test case, I prescribed the Riemann invariant (taken zero) at the north and south boundaries and water level equal to zero at the offshore boundary. For salinity calculations, at first a uniform value equal to 30psu is set as initial conditions and 30 at the other boundaries too (north,south,offshore). This simple test seemed to work fine at first with the flow vectors extracting from the channel to the outer field and developing a logical pattern.
But the transport boundaries are required to be of a free exit as well. Following the instructions given by the manual, this is done by activating the Thatcher-Harleman boundary condition and introducing a negative value of the required time of residence in the relationship as this ensures salinity transport to be calculated exclusively from the model and not influenced from the outside (ie free exit). In this case though, this transport condition results in a flow parallel to the north and south boundaries instead of being an exit flow, while a reverse flow occurs at the offshore boundary caused obviously by inappropriate density gradients.
In other words, applying a free exit condition for the salinity equation at the boundaries makes the flow to become parallel to the boundary or reversed back into the model area instead of getting through the boundary unobstructed. Is there a reason for this to happen ?
Any suggestions of how this could be achieved, that is, to have both flow and salinity transport free exit conditions at the open boundaries and the proper physical results as well?
Are there any other conditions that could be applied with or without the Thatcher-Harlemann option to overcome this issue?


Thank you,
Kostas
Erik de Goede, modified 2 Years ago.

RE: Thatcher - Harleman time lag

Youngling Posts: 13 Join Date: 1/7/11 Recent Posts
Dear Kostas,
The Thatcher Harleman boundary condition is meant to suppress unphysical currents near open boundaries. At first, try a simulation without TH boundary conditions. If unphysical currents occur, try positive TH boundary values; for example of 60 (minutes). At a last alternative try a negative TH value, which corresponds to a so-called Neumann boundary condition (d c / d x = 0).

In general, the TH boundary conditions have a marginal effect on the model results. If all options for the TH boundary condition give unrealistic results, then other options come into play (e.g., locally adaptation of the depth or the horizontal diffusivity). This is of later interest, At first, try zero or positive values for the TH boundary conditions. Success!

Erik de Goede, Deltares
KM
Konstantinos Matsoukis, modified 2 Years ago.

RE: Thatcher - Harleman time lag

Youngling Posts: 3 Join Date: 1/19/17 Recent Posts
Dear Erik,

Thank you very much for your reply.It is very explanatory.
I did a simulation without TH and the results look good (attached an example of salinity contours and flow vectors I get without TH).
However, it seems that the plume is forced in this case by the prescribed boundary conditions that I give for salinity equal to 30.
My flow boundaries are: Riemann at the north and south, water elevation equal to zero at the offshore.
I was looking for an appropriate way to free the transport at these boundaries and allow the contours to surpass if possible the boundaries.
Could this be done by implementing the TH condition and what are the restrictions in this case please? Could the TH be applied for example in a water level boundary?
Is there any other method to free the boundaries for salinity?

Thank you,
Kostas
Erik de Goede, modified 2 Years ago.

RE: Thatcher - Harleman time lag

Youngling Posts: 13 Join Date: 1/7/11 Recent Posts
Dear Kostas,

From your figure it is clear that the problem occurs at outflow. However, at outflow the TH boundary condition does not play a role. (NB. This only has some effect at inflow.) I guess that the problem is due to horizontal diffusion at theopen boundaries. Do you apply a sigma-model? (NB. In a Z-model the horizontal diffusion at open boundaries is neglected. However, in a sigma model this isn't the case.)
Assuming that you are using a sigma model, you can reduce the horizontal diffusion near open boundaries by applying spatially varying horizontal diffusion values. As a simple test you might do a simulation with the horizontal diffusion set to zero.

Wondering whether the explanation above is valid or not and with kind regards,

Erik de Goede, Deltares
KM
Konstantinos Matsoukis, modified 2 Years ago.

RE: Thatcher - Harleman time lag

Youngling Posts: 3 Join Date: 1/19/17 Recent Posts
Dear Erik,

Thank you for your help and suggestions. Indeed I am using a sigma model.
I have managed to surpass the issue by setting the horizontal diffusion equal to zero at the last two grid lines at the open boundaries.
In this way, salinity is not constrained anymore by the boundaries.

Thanks,
Kostas