null

Message Boards

Delwaq2: Scientific reference about numerical methods in Delwaq

GS
Gholamreza Shiravani, modified 1 Year ago.

Delwaq2: Scientific reference about numerical methods in Delwaq

Padawan Posts: 62 Join Date: 6/25/16 Recent Posts

Hi,

due to high diffrences between explicit and implicit numerical methods in Delwaq computations, could you mention a scientific Report/manual about their theoritical applications in Delwaq. I read and know the concise introduction in Delwaq-manual. In Manual it is referenced to :" D-WAQ TRM, 2013. D-Water Quality Technical Reference Manual. Deltares, 4.00 ed", but it is not published and when I search in Internet I found only :

https://oss.deltares.nl/documents/183920/185723/D-Water_Quality_Processes_Technical_Reference_Manual.pdf/9356affd-4453-4fef-9289-48d0a8d043ab?version=1.0

, which has nothing about implemented numeical methods.

 

AM
Arjen Markus, modified 1 Year ago.

RE: Delwaq2: Scientific reference about numerical methods in Delwaq

Jedi Knight Posts: 223 Join Date: 1/26/11 Recent Posts

The manual at https://content.oss.deltares.nl/delft3d/manuals/D-Water_Quality_User_Manual.pdf does describe the numerical integration methods in some technical detail and contains quite a few references to the underlying publications. Is this the manual that you already know? If so, in what way are these references not sufficient - no criticism, just trying to understand your question?

GS
Gholamreza Shiravani, modified 1 Year ago.

RE: Delwaq2: Scientific reference about numerical methods in Delwaq

Padawan Posts: 62 Join Date: 6/25/16 Recent Posts

Thanks Arjen for your answer. I read them but unfortunately I did not find information about how do these numerical schemes deal with the tidal areas (like islands). I have different sand islands that their boundaries are in some times wet and some times dry. I included these in D-Flow  mesh, without excluding them as land. My D-Flow works well. In contrast, Delwaq gives NaN values when the cells are dry and NaN-values are Extended in all parts of domain. When I try an implicit method like 16, it gives the GMR.. error and in .mon-file it recommends to reduce time-step, increase the Iteration number , etc .... but doing these will not solve the problem, at least the time step of 1 sec! . When I use explicit method like 11, the model runs without errors but the results becomes NaN after first time-step! When I see the corresponding map-file, I realize that  NaN-values are started from the islands in my open boundary and Extended to the whole part of domain. I have no substance and only continuity as input.

AM
Arjen Markus, modified 1 Year ago.

RE: Delwaq2: Scientific reference about numerical methods in Delwaq

Jedi Knight Posts: 223 Join Date: 1/26/11 Recent Posts

Ah, that makes it a bit clearer. Treatment of the drying and flooding grid cells is not part of the numerical method perse, but rather of how the flows are applied to set up the transport (before actually solving the equations). The drying/flooding is determined via a constant called DRY_THRESH. By default it is set to 1 mm (0.001 m), but in your case it may be beneficial to set to a somewhat higher value.