YOU. Tell me everything you know about building a physics engine for fluid simulation.

Name
Email
Subject
Comment *
File	Select/drop/paste files here
Password	(Randomized for file and post deletion; you may also set your own.)
* = required field	[▶ Show post options & limits] Confused? See the FAQ.

Flag
Oekaki	Show oekaki applet (replaces files and can be used instead)
Options	Do not bump (you can also write sage in the email field) Spoiler images (this replaces the thumbnails of your images with question marks)
Allowed file types:jpg, jpeg, gif, png, webm, mp4, pdf Max filesize is 16 MB. Max image dimensions are 15000 x 15000. You may upload 3 per post.

Anonymous 06/18/18 (Mon) 03:26:22 No.931653

YOU.

Tell me everything you know about building a physics engine for fluid simulation.

Anonymous 06/18/18 (Mon) 03:32:23 No.931655

quaternions

Anonymous 06/18/18 (Mon) 04:07:39 No.931662

Y'all niggers don't even Navier–Stokes

>>931656

>nigga

>>>/reddit//

Anonymous 06/18/18 (Mon) 04:10:31 No.931663

>>931653

It's just solving PDEs, particularly Navier-Stokes >>931662, but since that doesn't have analytical solutions, it requires numerical methods to find approximations, and according wikipedia, finite volume methods are commonly used

https://en.wikipedia.org/wiki/Computational_fluid_dynamics

https://en.wikipedia.org/wiki/Finite_volume_method

Anonymous 06/18/18 (Mon) 04:20:26 No.931665

I know it's an unsolved problem. We can't even get realistic simulations for all fluids yet (we're actually getting kind of close now).

Anonymous 06/18/18 (Mon) 04:24:47 No.931667

>>931665

They still can't get hypersonic flight working, so I'm guessing their simulations aren't very accurate if they think their x-planes will work and they don't.

Anonymous 06/18/18 (Mon) 04:25:56 No.931668

>>931667

Although that probably involves issues simulating combustion, so those simulations would require fluid dynamics and molecular dynamics combined.

Anonymous 06/18/18 (Mon) 04:44:35 No.931673

yall niggas dont even know about Texture Distortion

https://catlikecoding.com/unity/tutorials/flow/texture-distortion/

Anonymous 06/18/18 (Mon) 04:59:34 No.931677

File: d22028fef815e85⋯.jpg (39.16 KB, 604x604, 1:1, 5djfRQSoTtQ.jpg)

>>931676

I should have filed this question under Tech Questions General #25 - Subsection QTDDTOT #5402 - Subsection Simulations #3

Anonymous 06/18/18 (Mon) 06:19:50 No.931699

>>931653

Sure but first I'll need to see some cuts. Make them deep.

[k00l]shamoanjac!!v.CX2tveU. 06/18/18 (Mon) 07:00:26 No.931723

I've read you can use cellular automaton-based approaches

Anonymous 06/18/18 (Mon) 07:39:04 No.931734

If it's something that doesn't need actual flow and volume simulation (eg. just going for visuals), don't do it.

Just pre-render extremely high quality animated textures instead.

For example (pre-rendered waves):

https://www.youtube.com/watch?v=L36orF8rC-M

Which is from:

https://forum.unity.com/threads/water-system-based-on-baked-fluid-simulation.504291/

Anonymous 06/18/18 (Mon) 07:42:52 No.931735

>>931673

That doesn't look good though.

Can get better results with some hand-made textures and subtle blending/displacement layers.

Anonymous 06/19/18 (Tue) 18:36:06 No.932288

Something something voxels something.

Fuck I donno. What do you know?

Anonymous 06/21/18 (Thu) 17:01:24 No.933261

>>931653

Particle Mesh Ewald

Anonymous 06/24/18 (Sun) 11:57:35 No.934489

File: 35ab0dc1f4eb151⋯.jpg (34.21 KB, 500x493, 500:493, 1478053285913.jpg)

Bump.

What are some new features that one could fundamentally implement in a physics engine in order to increase efficiency where other physics engines are too far in development to correct now?

For example, is there a particular way of representing objects or storing them that has superseded efficiency of prior implementations?

Anonymous 06/24/18 (Sun) 13:12:36 No.934496

>>931653

Simulate n balls with radius m where n approaches infinity and m approaches 0.

Anonymous 06/24/18 (Sun) 13:17:57 No.934497

File: 7cbacf5ea2c51fb⋯.webm (7.84 MB, 640x360, 16:9, 1468880685660.webm)