[Soc-2017-dev] Weekly Report #13 - Mantaflow Particles and Fluid Guiding

[Sebastián Barschkis]

Hi Nils,

Thanks! Yes, I recently tried out a guided "liquid tornado" and that seemed to work just fine:

https://twitter.com/sebbas/status/901590925008400384

Once the "velocity bug" is out of the way we should be able to test even more interesting setups, e.g. guiding a constant liquid inflow. Then we'll see if it really works.

I think your way to thicken obstacles is a good addition to my recent fix for thin obstacles. There I did the following:

1. Compute SDF from mesh (SDF only contains positive distances if no closed mesh is present, e.g. plane as obstacle)
2. Find and mark all cells near mesh surface ("near" determined by "surface distance" value -> already exists for smoke).
3. All cells from 2. that were "near" the surface are considered to be part of the obstacle.

After this, the setup becomes inaccurate (I am currently just forcing all cells from 3. to be negative -> SDF not preserved -> FLIP parts at obs border in danger).  
But to me it looks like your idea with the double extrapolation could fix exactly this problem. Will try that out now :)

Best wishes,
Sebastián

> On 30. Aug 2017, at 15:48, Nils Thuerey <nils.thuerey at tum.de> wrote:
> 
> Hi Sebastian,
> 
> interesting, congrats on wrapping up the project. Looks very good. Btw., have you tried the guiding for liquids? I'm just curious whether it works at all..
> 
> Also, could you re-cap how you're currently handling the thin-obstacles? I would recommend re-extrapolating the SDF there. E.g., you could add a "thicken" option (1 float value), and then do:
> - compute obstacle SDF (ideally for a single object)
> - extrapolate
> - subtract thicken-value 
> - then extrapolate again (to get thickened version)
> That should close the gaps. We can talk about it again during our next meeting.
> 
> Cheers,
> -> Nils 
> 
> 
> === Technical University of Munich , http://ge.in.tum.de/ ===
> 
> 
> 
> 
> 
>> On 25. Aug 2017, at 23:28, Sebastián Barschkis <sebbas at sebbas.org> wrote:
>> 
>> Hello everyone,
>> 
>> my report #13 for the last GSoC week is now online:
>> 
>> https://wiki.blender.org/index.php/User:Sebbas/GSoC_2017/Reports/Week_13
>> 
>> = Week 13 =
>> 
>> This was the last GSoC week. I worked on fluid guiding and the documentation.
>> 
>> === What's been done ===
>> 
>> * I added the missing control parameters for fluid guiding to the UI. With them you can control guiding lag and the size of the vortices.
>> 
>> * Fluid guiding functions now kick in automatically. All it takes is that there is least one guiding object present.
>> 
>> * Part of the documentation is now online at: https://wiki.blender.org/index.php/User:Sebbas/GSoC_2017/Documentation
>> 
>> === Next up ===
>> 
>> * Over the weekend I'll add the missing sections to the documentation (on Monday I'll send the final update)
>> 
>> === Questions ===
>> 
>> * No
>> 
>> Best wishes,
>> Sebastián
>> _______________________________________________
>> Soc-2017-dev mailing list
>> Soc-2017-dev at blender.org
>> https://lists.blender.org/mailman/listinfo/soc-2017-dev
> 
> _______________________________________________
> Soc-2017-dev mailing list
> Soc-2017-dev at blender.org
> https://lists.blender.org/mailman/listinfo/soc-2017-dev