Dam break flow 'fly' after approaching the slope & mDBC setting for the slope
Hi, everyone
Instead of an obstacle, I changed the downstream of dam break model as a beach. However, when the dam break wave propagated on the beach, like hitting on the beach, it 'fly', as shown in the images below. But, it is only a 5-degree slope, not an obstacle, the dam break wave should flow on the slope surface. Did anyone have this situation before? I saw this phenomenon in V4.4 and V5.0. And what should I adjust to make the flow similar to the real one? Thank you!
To build the slope in the right part, I used the <drawextrude>, so I can use the <layers> command to generate boundary particles for applying mDBC. And for the left part and the right wall, I used <drawbox>. Here in the image, it seemed there is 'steps' in the boundary particles in the right slope part. Will this be the reason that the flow 'fly'? Thank you!
Here I attached the XML file in text format for your view. Thank you all !
Comments
You need to apply some changes to reproduce this with mDBC and to compute more accurate normals:
1) You need to create more layers, 3 at least, however this depends on your smoothing length "h"
note that 2h should be smaller than the layers*dp
2) In the extrude option you can apply the option "setfrdrawmode". You can find more information in page 57 "XMLGUIDEv5.0.pdf"
other good example you can follow using beach with the FreeDraw mode is examples\mdbc\06_WaveTank
Regards
Hi Alex,
Thank you very much for your immediate reply!
Sure, I will check the normals again. add 2 more layers, and also try the "setfrdrawmode".
I read the file of examples\mdbc\06_WaveTank which is good, but it is 2-D, so no front and back wall was built, becasue the domain is limited at y=0. In my case, it is 3-D, so I have to build the front and back wall. I will refer to this case and leave feedback later. Thank you again!
Hi Alex, based on your advice, I solved the flow "fly" problem. But a new warning happened: There are 2648 of 402650 fixed or moving boundary particles without normal data. Is this because I set too many layers for the boundary particles? As when I decrease the layers, this number also decreased. Thank you again for your kind help!!!!!!
Exactly, normals are only computed for the particles at a distance of 2h from geometry limit!!!! since those are the ones included in the different computations of mDBC particles. Meanwhile the rest of extra layers will not interact with fluid.
In addition, you can also compute normals only for some objects, so those boundaries will use mDBC while the rest of boundary objects of your system will use DBC
Regards
Hi Alex, got it! Thank you again!!!