FLIP Fluids Addon: A liquid fluid simulation tool for Blender

lumpyoatmeal · March 8, 2018, 12:26pm

RLGUY:

The viscosity value in this simulator doesn’t relate to a real life physical quantity. So if you copy a real life viscosity value into the simulator, it may not match the visual thickness of the real life liquid. The simulator should be able to handle viscosity values as low as 0.001, so I would experiment with viscosity values lower then 0.1 if it is still too thick.

Tip: When experimenting with viscosity, try values in the range of 0 to 10 to get an idea of how the simulation will look. If you want to experiment with even thicker liquids, then you can try values up to around 100.

Another tip: The visual ‘thickness’ of the fluid is dependant on the size of the simulation domain. For example, a viscosity value may appear ‘thick’ in a small domain size, but if the same value is used in a larger domain, the fluid may appear visually ‘thin’.

Ok, thanks. I’m trying one now with the viscosity set at 0.001.

jensverwiebe · March 9, 2018, 6:07am

I would ask someone who knows howto :yes:

At least you can see how blender / fracturemodifier solves this non-problem.

Just support systems following the blender specs and you are on the safe side, supporting outdated versions is not worth it anyway.

Jens

anon65932258 · March 9, 2018, 12:20pm

What will be the difference between Flip Fluids Addon and mantaflow?

RLGUY · March 9, 2018, 3:22pm

I have spent much more time using this simulator than I have spent trying out Mantaflow, so I am not able to give an accurate/unbiased comparison for FLIP Fluids Addon vs Blender Mantaflow. Perhaps testers who have used both simulators would be able to make a fair comparison.

Both this addon and Mantaflow are based on the FLIP method. I believe we both even use the same method of simulation for foam/bubble/spray particles. What will be the difference is how we each implement the simulation methods, what features we have chosen to include in our respective simulation programs, how we decided to layout our user interfaces, etc.

Pavel_Blend · March 11, 2018, 5:22am

Question to beta testers:
Can you show the video with the fluid, simulated using Viscosity Solver?
I’m curious to know how this solver works.

dafassi · March 11, 2018, 10:03am

Like this?

Pavel_Blend · March 11, 2018, 10:14am

Yes. There are still examples?
Is it possible to create similar simulations using Viscosity Solver?:

Charlie · March 11, 2018, 4:38pm

Creepy!!!

RLGUY · March 13, 2018, 11:49am

Yes it is possible to create fluids with similar viscosities. Although, multiple fluids and mixing fluids is not possible as in the last animation.

Here are some more viscosity examples in order from thin to thick! Click images to view animation.

Created by Lawrence Jaeger

Razorblade · March 13, 2018, 5:08pm

RLGUY:

@draguu, thanks for your details! I have entered your info into the application pool.

At the moment, only the Windows operating system is supported. This is because we are currently developing on the Windows platform and it is easy for us to create and distribute builds of the addon for Windows users.

We do hope to support OS X and Linux operating systems in the future. The fluid engine code is written to be cross platform and should theoretically run on OS X and Linux, but will require testing to ensure that it does. We also need to find a reliable way of testing and creating builds for OS X and Linux.

OS X Support

At the moment we do not have access to a system with the OS X operating system installed. This will be required to create OS X builds and for testing

Linux Support

In order for us to support Linux, we will need to find a way to reliably test and create builds that will work on the many distributions of Linux.

The FLIP simulation method uses both voxels and particles to simulate the fluid. Both the viscosity solver and whitewater simulator is based on a particles/voxel combination also.

check https://docs.microsoft.com/en-us/dotnet/core/tools/dotnet-run?tabs=netcore2x

yvesbodson2 · March 14, 2018, 12:06am

Great! Sorry I would have engaged in your beta test but I am a Linux user… Let me know when you have a Linux version for testing…

anon65932258 · March 16, 2018, 3:29pm

please delete…

Pavel_Blend · March 18, 2018, 9:29am

Q: FLIP Fluid Addon bake simulation faster than Elbeem and MataFlow?

kumayuki · March 19, 2018, 1:47am

Sim Time：about 10 hours (i7-4930K)

FLIP Fluids Addon (Beta) is Wonderful

Other videos

anon54214979 · March 19, 2018, 4:16am

Very cool, I also made some beach waves tests, just no renders yet!

kumayuki · March 19, 2018, 4:57am

Thank you. I am looking forward to your wave.
I took 41 hours to render(500frames)

Bernardo · March 19, 2018, 5:44am

really good looking tests!

the beach one also looks great, but unfortunately I can see a bit of the same problem that I experienced with manta flow, even if less noticeable. On the shoreline, where the waves hit the beach and then retreat, you can see a “band” effect. Is this a matter of settings, or is a limit of the solver?

Other thing, in some simulation the foam on the surface looks a bit too tense. You have these narrow “lines” of foam, I think they should expand a bit more.

These wants to be constructive comments, the simulations I saw so fare looks great, keep up the great work!

kumayuki · March 19, 2018, 6:51am

Thank you for your comment.
Next time I would like to try a workaround for banding artifacts.

Banding artifacts against gradual sloping obstacles #1
https://github.com/rlguy/Blender-FLIP-Fluids-Beta/issues/1

I have not fully understood the bubbles yet. I want to improve.

Pavel_Blend · March 19, 2018, 9:11am

I would like to become a beta tester. I already sent the request to the mail, but I did not specify the processor speed. Therefore once again I write.

GitHub Username: PavelBlend
Operating System: Windows 7 64 bit
CPU Model: AMD A8-5600K @ 3.60GHz APU with Radeon™ HD Graphics
GPU/Chip Model: -
RAM Size (GB): 8 GB

RLGUY · March 19, 2018, 12:38pm

The banding effect is caused by a limit of the solver. Very gradual slopes such as in beach wave simulations result in very thin layers of fluid that are only 1 particle thick. This is something the FLIP method has trouble with. With these thin liquid layers, there just doesn’t seem to be enough data for the solver to resolve the physics in a clean way. I’ve been trying a few different things, but have not yet found a way to fix this issue. Hopefully I’ll come up with something!

There is a ‘Foam Advection Strength’ parameter in the Whitewater Settings panel that could help with this.

Advection Strength: The strength of the advection force that carries the foam with the fluid velocity. High values cause tighter streaks of foam that closely follow the fluid motion. Lower values will cause more diffuse and spread-out foam.

Thanks Pavel_Blend, I have your details entered into the application pool.