Multi-phase (air/water) + moving object simulation
- This topic has 1 reply, 2 voices, and was last updated 2 weeks, 2 days ago by .
-
Posts
-
Hello everybody.
New user here (of both LBM and openLB).
My question has some similarity with the one in the “Two-phase LBM-DEM” thread, but I found it different enough to create a separate conversation. We can move it there if you prefer.
Since a couple weeks I’m investigating the possibility of simulating in OpenLB the immersion of an object into water and the computational cost of this.The domain is a cuboid partially filled with water, and air on top (immiscible, no phase change).
The object will be defined using an STL file, and I want to control its movement over time according to specific key positions, which will be interpolated. There is no need to account for inertial or hydrodynamic forces acting on the object. I’m not concerned with the object’s interior; it can simply be “empty.”Surface tension, gravity and (air) buoyancy are important, since I want to see possible air bubbles trapped in the object surface cavities. Surface wettability (contact angle) may have a role here, but secondary at this stage. Modelling of air compressibility is also not essential at this stage. Laminar flow is also enough for starting.
To my understanding, The two main ingredients needed should be: 1) some implementation of multi-phase/component, and 2) a framework for the moving object.
Correct me if I’m wrong, but I haven’t found an example with such combination.From the user guide (and the posts on the other thread) it seems that the preferred framework for the object movement is the particle ones (is it always HLBM framework?), reference example: settlingCube3d.
Another option may be the immersed boundary method (IBM), which I only see documented in the deformableParticleInShearFlow3d example.About the multi-phase/component models available, for what I see, the single-component multiphase pseudo-potential approach seems the easiest and lighter suitable one (dropletSplashing3d example). However, I have some doubts on the best model choice for my case.
Clearly, there is also the whole big question about its compatibility with the chosen moving object framework.I have indeed many questions on the topic, but the first ones are:
Can you confirm that this type of simulation is (easily enough) doable in the current implementation of OpenLB?
Can you confirm that there is no existing example combining these functionalities that I could use as a starting point?
Do you have any suggestion on the single models/framework selection?
And key insight on how to merge them?Thank you very much for any help on this.
Hello,
in theory, it should be doable by coupling the multiphase code with the mentioned HLBM code. This would mean creating new dynamics as well as figuring out a way to do the contact angle computation on the HLBM interface. There is no example, that does both models yet, that is correct. In terms of the multiphase approach I would suggest conservative phase field (i.e. bubbleChannel2d) as this phase field model allows most stable conditions for high density ratio and surface tension.
Hope this helps.
Kind regards
Tim
- You must be logged in to reply to this topic.