Sebastian Gürke M. Sc.

Hey there,

has anyone of you succesfully used the free outflow boundary condition while simulating precipitation in MIKE 21 FM? I think it was first introduced in MIKE 2019.
When I use this boundary condition my models usually shows blow ups, hundreds of CFL-Errors etc. while the model runs ok when using a land boundary.

Quote from the manual:
[quote]The free outflow condition is typically used at downstream boundaries where the water is flowing out of the model. In case of water flowing into the model domain the simulation may become unstable.[/quote]

What means "(...) in case of water flowing into the model (...)"?

Thanks

[quote]

I actually do have a copy of an academic version of MATLAB, would this be sufficient enough to run the DHI MATLAB TOOLBOX?
[/quote]

I don't really know, sorry. Give it a try :-).
[quote]
Also tried googling the said toolbox and found a Github page for DHI, is this where I should be able to get all the necessary info?
[/quote]
Yes, exactly: https://github.com/DHI/DHI-MATLAB-Toolbox
Note, that there is a short UserGuide: https://github.com/DHI/DHI-MATLAB-Toolbox/blob/master/Documentation/DHI_MATLAB_Toolbox_User_Guide.pdf
On page 15 it says, that there is an approach to interpolate element center values to the node positions.

Greetings
Sebastian

Hey,

[quote author=bakjee link=topic=21986.msg29063#msg29063 date=1592366027]

I was hoping to display the velocity at each time step for all the nodes in my domain, something that might look like the image in the attatchment. I have a feeling that this could be the with the result viewer tool but have no clue how this can be achieved?

(...)

Edit:

So, im not sure if this is the correct way however but I converted dfsu to a reasonably sized grid files where I can kinda achieve what I wanted to do in the tabular view but i'm guessing there is an easier and more appropriate way. Also converting into grid files takes very long and a lot of computing powers...
[/quote]

I think your solution is one way to go, but as you mentioned it doesn't show the velocity at the nodes. As far as i know, there is no build-in function to achieve this, since MIKE21FM uses an Finite Volume approach, which means that the values are calculated at the center of the mesh elements. In case you have Matlab you can check the DHI-Matlab-Toolbox.

Another way is MIKE2SHP (Mike Zero Toolbox), which converts the dfsu-file to a shapefile. But due to the the Finite Volume Approach and the way MIKE2SHP works, the result only shows the velocity for each element and not at the nodes.

Greetings,
Sebastian

Hey,

[quote author=jkurzweil link=topic=21985.msg29062#msg29062 date=1592321794]
For what ever reason when I create the .dfs2 file using the grid2mike tool box, it does not really give the file all that it needs, specifically it does not give it an item. Maybe I am just doing it wrong, so if someone knows something I don't please let me know. To fix this I opened the created .dfs2 file directly from my explorer. This opens the file in a MIKE zero editor. I hit edit->items, then under type I selected elevation for type and meter for unit.
[/quote]

As far as i know, thats the "normal" way to do this. The reason is, that MIKE simply doesn't know what kind of data the values of the asc-file represent. Im using asc-Files and Grid2Mike to import spatially distributed rain patterns, so the item should be precipitation height rather than elevation or anything else.

Greetings,
Sebastian

Hey Simon,

just a guess:
If you select the "Geographic View" Tab in the Output Menu in MIKE 21, do you see some sort of "fishnet" on top of your model domain? If not, i guess that the coordinates of your area output do not fit your model domain. This might happen, if you "reuse" an old *.m21/*.m21fm-File and you change the model domain (e.g. the extent). The coordinates in the "Output specification" Tab do not change automatically and need to be changed to fit the new model domain.

Hope this helps!

Hey,

check this:
https://forum.mikepoweredbydhi.com/index.php/topic,1592.0.html
and
https://faq.dhigroup.com/default.asp?module=MIKE+21&ID=57#answer

Hope it helps!

Hi,

I haven't really worked with MIKE11 so i just know some sort of workaround. Do you have MIKE HYDRO? In this case you can add Shapefiles as Background Layers and afterwards open the menu Tools > Auto generate Cross Sections and choose "Survey points from shape file". Now you can select the shapefile you added as Background Layer in the step before. The cross section shapefile should be a point shapefile with a field that defines which survey point belongs to which cross section number.

The river network can be imported in MIKE HYDRO by File > Import from shape file and then choose "Branches".

Hope this helps!

Any hints?  :)

Hi Christian & Hasan,

[quote]would be interesting, if you got any solution for that.
[/quote]

No yet. To be honest I didn't have the time to figure it out and hoped someone knew a solution, sorry. :D

[quote]We for ourselfs have solved the mentioned problem with a GIS - Tool inside our toolbox.[/quote]
Interesting, could you explain how you do that in GIS (I'm quite familiar with ArcGIS).
[quote]I understand that you only have point data of your cross-sections not any data between successive cross sections right? not dense data or point cloud?[/quote]
Thats absolutely right, thanks!

Do you have any hints how you would do that in GIS?

Creating the quadrangular-mesh (without Z-Information) in MIKE21 ain't perfect, but it works. I'm just afraid that the Z-Interpolation based on only those cross-section height points is going to create a weird z-interpolated-mesh between those single-cross-sections.

Thanks in advance!
Sebastian

Hej,

I'm planning to model a river including its floodplain and I'm not quite sure, what the best way is to do so. The floodplain will be based on a digital elevation model and will be modeled as a flexible mesh; easy so far.
But how can I model the river based on river cross sections (point-shapefile) as an quadrangular mesh? Would you simply digitalize the left and right river bank (f. ex. based on aerial images as polyline; without Z-Information) and import them as arcs, connect the alternate ends by arcs as well, redistribute the vertices of the two river bank arcs, generate the mesh AND afterwards apply the Z-information based on the river cross sections by importing the points from the point-shapefile as .xyz-file?

Coming from other software I'm used to a different approach: Import the cross section points (with Z-Information) and connect the single points by arcs and afterwards generate the mesh to ensure that each measured cross section point is part of the generated mesh.

Feedback & ideas are much appreciated, cheers.

Sebastian