Melt some ice cream

   3708   6   0
User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
Hi there,



this is my first flip simulation in Houdini, so I guess there are a lot of things which I might handled wrong, but there is still “some” space to improve the sim and my knowledge.

First of all, I want to melt some icecream. Right now I do have two different viscosity values applied to the sphere particles. This will be simulated later on, but for now it's fine with me for this test.

I managed to melt the sphere in some way, though it looks wrong over time. Sometimes the particles hardly move at all, then they seem to float away, though I can't recall to have changed any settings via keyframes. Also with tension applied the particles start to roll in a wheel like manner.

Since the outer particles “swim away”, these particles will cause meshing problems, since they are too far apart from each other.

In order to to force the flowing fluid to build drops and “real” shapes, I added surface tension, though this still seems to not solve the problems. Still particles are apart from the main mass and the strange behaviour (stop and go) continues.

Since I want to simulate everything in real life scale the diameter of the sphere is about 0.05. I changed all corresponding parameters to my best knowledge, but am not sure if I did everything in the right way.

The two videos demonstrate the current behaviour. The only difference is the enabled/disabled surface tension in the fluid emitter.

What I need to do (no idea how yet) is to apply the tension only to the low viscosity particles. Probably with an attribute VOP?

Any help is very much appreciated. Coming from SI ICE some things seem straight forward, but others impossible to solve.

Best wishes and thanks for any input

Tom

https://vimeo.com/134251136 [vimeo.com]

https://vimeo.com/134251135 [vimeo.com]

Attachments:
iceCream_melting_001_a_0095.hiplc (1.8 MB)

User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
So, here are the next tests. I thought that strange stop and go behaviour might be related to the friction and/or dynamic friction scale.

In both tests it seems that this is not the cause for this problem. It looks like some values are changing, but I couldn't find any.

Some help is more than welcome!!! Got stuck.

https://vimeo.com/134296167 [vimeo.com]

https://vimeo.com/134296168 [vimeo.com]
User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
Mmmmh. My iceCream_partciles consists out of approx. 64.000 particles. The DOP import in iceCream_meshed tells me something about 360.000 particles (frame 5). At frame 10 it's 400.000. Where do all these extra points come from?

Can't figure out how to get the amount of particles in the AutoDopNetwork. Middle klicking doesn't give me any particle statistics, neither on the network itself, nor on any of the nodes inside.

I added $SF == 1 for creation frame in the flipfluidobject1. Didn't make a diverence.
User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
O.k., in another attempt I set the friction of the ground plane to .1 (before 1). The sim still looks the same. Initially the partciles which flow come to a sudden stop (on the ground plane) and approx. 100 frames later they start flowing away from the ice cream scoop.

It somehow feels like something related to friction or stickyness, but no idea where to change anything to alter this behaviour.
User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
Surface extrapolation! Who would have thought it!
User Avatar
Member
15 posts
Joined: Jan. 2015
Offline
So, I got further. The simulation seems to get more “stable”. At least it looks more and more like something I would expect.

One point needs to be solved though. I'm using the gas tension in my flip solver. Seems that I'm close to a value that feels alright. One major issue: The surface tension seems to “override” the viscosity. The frozen parts start to form the same liquid/drop patterns like the areas of the ice which already have melted. So I need to combine the surface tension with the viscosity. How would I do this?

The viscosity was applied in my iceCream_particle object in a attribute vop. Since the visosity will be changing later on (via temperature) how do I apply the surface tension to the particles in the AutoDopNewtwork during the simulation? Practically?
User Avatar
Member
68 posts
Joined: Sept. 2014
Offline
hi

I'm new to Houdini too
and watching some Tutorials

maybe the LAVA Tutorals from Ben helps you
a little different Topic but same Behavior

found it on CG-Press
http://cgpress.org/archives/cgtutorials/creating-lava-in-houdini [cgpress.org]

direkt Link
https://vimeo.com/126886471 [vimeo.com]
  • Quick Links