Gravity in fluid simulation #73680

New Issue

David Whitaker · 2020-02-08T20:21:01+01:00

David Whitaker commented

2020-02-08 20:21:01 +01:00

Blender Version
Broken: version: 2.82 (sub 5), branch: fluid-mantaflow, commit date: 2019-12-16 14:05, hash: ed32db06c1

Short description of error
With scene gravity set to zero, if I change the fluid domain gravity, it ends up behaving as if it's about 10 times too strong. So when the gravity is set to -1 m/s squared it behaves like -10m/s squared

Exact steps for others to reproduce the error

Add cude,
add sphere.
scale cube (2x) and make it fluid domain (Liquid),
make sphere a fluid liquid flow with flow behavior set to geometry.
Turn off scene gravity
set gravity in fluid domain settings to -1m/s^2 (default).
Bake data and see result

fluid_gravity_bug.blend

**Blender Version** Broken: version: 2.82 (sub 5), branch: fluid-mantaflow, commit date: 2019-12-16 14:05, hash: `ed32db06c1` **Short description of error** With scene gravity set to zero, if I change the fluid domain gravity, it ends up behaving as if it's about 10 times too strong. So when the gravity is set to -1 m/s squared it behaves like -10m/s squared **Exact steps for others to reproduce the error** - Add cude, - add sphere. - scale cube (2x) and make it fluid domain (Liquid), - make sphere a fluid liquid flow with flow behavior set to geometry. - Turn off scene gravity - set gravity in fluid domain settings to -1m/s^2 (default). - Bake data and see result [fluid_gravity_bug.blend](https://archive.blender.org/developer/F8332475/fluid_gravity_bug.blend)

David Whitaker commented

2020-02-08 20:21:01 +01:00

Added subscriber: @davidwhitzim

Germano Cavalcante commented

2020-02-10 14:02:45 +01:00

Changed status from 'Needs Triage' to: 'Confirmed'

Germano Cavalcante commented

2020-02-10 14:05:01 +01:00

Added subscriber: @mano-wii

Germano Cavalcante commented

2020-02-10 14:05:01 +01:00

I can confirm.
The default values already seem to be 10 times faster.
This seems to be a regression with the new fluid system.

I can confirm. The default values already seem to be 10 times faster. This seems to be a regression with the new fluid system.

huajian qiu commented

2020-02-25 21:35:26 +01:00

Added subscriber: @wakin

Jacques Lucke commented

2020-03-04 11:23:00 +01:00

Added subscribers: @sebbas, @JacquesLucke

Jacques Lucke commented

2020-03-04 11:23:00 +01:00

The cause for this behavior is in manta_set_domain_gravity.

It looks like mantaflow uses normalized gravity values, for reasons I don't really understand. Also I noticed that the liquid seems to fall faster than a rigid body object. Not sure why that is.

In any case, I think the Scene Gravity and Gravity in the fluid settings should have the same scale. Two ways to achieve this:

Scale up FluidDomainSettings->gravity in versioning code and scale it down in manta_set_domain_gravity in any case.
Scale up FluidDomainSettings->gravity only in rna code using getter and setter functions and keep manta_set_domain_gravity as it is.

I prefer the first solution. What do you think @sebbas?

The cause for this behavior is in `manta_set_domain_gravity`. It looks like mantaflow uses normalized gravity values, for reasons I don't really understand. Also I noticed that the liquid seems to fall faster than a rigid body object. Not sure why that is. In any case, I think the Scene Gravity and Gravity in the fluid settings should have the same scale. Two ways to achieve this: * Scale up `FluidDomainSettings->gravity` in versioning code and scale it down in `manta_set_domain_gravity` in any case. * Scale up `FluidDomainSettings->gravity` only in rna code using getter and setter functions and keep `manta_set_domain_gravity` as it is. I prefer the first solution. What do you think @sebbas?

Sebastián Barschkis commented

2020-04-23 13:43:22 +02:00

Yes, having gravity on the same scale as in D7483 is definitely a good idea.

Will check now why the liquid still moves down quicker.

Yes, having gravity on the same scale as in [D7483](https://archive.blender.org/developer/D7483) is definitely a good idea. Will check now why the liquid still moves down quicker.

blender-admin commented

2020-04-23 15:10:20 +02:00

This issue was referenced by 694c0547c2

This issue was referenced by 694c0547c21aa3a5087c0dd4323624358135f4ef

Jacques Lucke commented

2020-04-23 15:11:44 +02:00

Changed status from 'Confirmed' to: 'Resolved'

Jacques Lucke closed this issue

2020-04-23 15:11:44 +02:00

Jacques Lucke self-assigned this 2020-04-23 15:11:44 +02:00

Thomas Dinges added this to the 2.83 LTS milestone 2023-02-08 16:39:36 +01:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Gravity in fluid simulation #73680