Floating particles stick together
Hi developers and community,
Recently I have tried to simulate shear suspention flows in 2D shear cell. Because it needs periodic condition and so far chrono haven't supported it. So I chose DCDEM mothod to simulate floating particles. Unfortunately when several DEM partiles met each other in fluid, rather than bounce off, they tended to stick together and then swim in the fluid like chains.
To find out the reason of unphysical phenomenon above, I tested several cases.
- I set pretty large Young's modulus E as 10E10 and set restition coeffient as 1 (which means no damping in DEM). But it still tuck togther.
- I setted Young modulus E as 10E10 and restitution coeffient as 1 (no damping in DEM). But particles still tended to stick together.
- I enlarged coefficient of stiffness function from 4/3 to 40000/3. Particles still stuck together.
- I simulated the Drafting-Kissing-Tumblicases with DCDEM and chrono method seperately. And two particles were more likely to kiss and then never seperate in DCDEM cases.
Is there anyone can give me some advice or tell me what point did I miss in my simulation?
Comments
Hi,
I ran a similar test some time ago. At that time there was only DBC (not mDBC), my interpretation is that the gap around boundary particule was the cause of the effect you describe. When two floatings (or floating and wall) get close, at one point, there is no particule in between (double the gap) and so they get closer (no force or pressure to conteract) and finally stick together since there is no way for a fluid particule to get in between.
That's only my humble opinion though...
Thanks for your explanation @TPouzol. That makes sense. Have you found the way to solve it? Because now In release version. the mDBC technique still didn't work for floating bodies.
No I didn't. And I'm not sure that mdbc is applicable on floatings, some post in the forum should enlighten you !
PS: This is the photo of certain part. There is some overlap but the particles don't bounce off. So I'm not sure it's problem of setting of DCDEM or the DBC or both of them. I also tried setting ViscoBoundFactor as 0 and the result wasn't satisfactory.
Have you tried using chrono ?
The error here is DBC. In next version of DSPH when you can use mDBC, this should be much better.
For now I would try to play with the "interaction gap", i.e. when particles register each other to see if you can promote the feature you want.
I would also slightly reduce the diameter of my circles, so that when using DBC, the radius would be = actual radius + DBC error.
Using Chrono for the interaction is probably smarter, since here you can play with the interaction gap directly.
Kind regards
Thanks@TPouzol. Yes I have tried using chrono in Drafting-Kissing-Tumblicases cases. But sadly it don't have "periodic condition" function.
And thanks@Asalih3d for your advice. I will give it a try.