gas solver exploration

[keltuzar]

I have been playing with houdini's gas solver, and here is one of my first simple test.

http://www.serrao3d.com/TestMovs/gasTest1.mov

So this is a new sim, and this is what I found what trumps maya's fluid solver.

This sim has temp (red stuff) advecting differently with custom vorticle fields from density which isn't moving with a voritcle force.

http://www.serrao3d.com/TestMovs/gasTest2.mov

Edited November 16, 2008 by keltuzar

[anakin78z]

Interesting idea. So the vorticles are sessentially messing up your temperatures, which then gives you an interesting look in your density?

That get's me thinking... I'm sure one could set it up so that vorticle strength is controlled by temperature, so that you get the most turbulence in hot areas... or should it occur only in the areas of higher temperature change? I guess there's probably a way to get the gradient of the temperature at each sample... well, maybe anyway.

Anyhoo, neat stuff .

-z

[SpencerL]

Interesting idea. So the vorticles are sessentially messing up your temperatures, which then gives you an interesting look in your density?

That get's me thinking... I'm sure one could set it up so that vorticle strength is controlled by temperature, so that you get the most turbulence in hot areas... or should it occur only in the areas of higher temperature change? I guess there's probably a way to get the gradient of the temperature at each sample... well, maybe anyway.

Anyhoo, neat stuff .

-z

I could be wrong since Ive been learning this by trial and error, but the behavior of increasing the Vorticle Confinement parameter on the Smoke Solver seems to increase turbulence in already turbulent areas. But actually I think that deals with the velocity vector field, but Im sure like you said, it could be modified so that it looks at the temperature instead.

[keltuzar]

Interesting idea. So the vorticles are sessentially messing up your temperatures, which then gives you an interesting look in your density?

That get's me thinking... I'm sure one could set it up so that vorticle strength is controlled by temperature, so that you get the most turbulence in hot areas... or should it occur only in the areas of higher temperature change? I guess there's probably a way to get the gradient of the temperature at each sample... well, maybe anyway.

Anyhoo, neat stuff .

-z

so there are various fields in the smoke object. These fields are either scalar, vector or geometry.

In houdini there are predefined data names like Vorticles which is geometry. So basically if u lay down say Gas Vorticle Geometry and turn on points, you can see that it is actually a geometry datatype. To make vorticles work the solver is looking for a "mag" variable... and you can see that in the details view. Also in the node itself in the data name you can see data name vorticles.

In the smoke solver, if you dig around you can see that they use a gas vorticle forces node and lo and behold! it is looking for the vorticles data and a vector field to dump the output to. In the smoke solver, the output is going to the general vel which is advecting all the grids that are active.

so here is wht I did, I created my custom vector field, called it temperatureVel, created my own tempVorticles data node, and dumped the voritcles output off the gas vorticles forces into temperatureVel. Then all I ended up doing is using gas advect temperature with temperatureVel.

You can do interesting stuff like this, where you affect only a certain grid to move the way you want keeping the rest of the grids to advect using the default vel grid...

Hopefully the fluid gurus can drop by and give their 2 cents on how the solver is working... again this is me trying to decode what is going in the complex setup they have provided.

cheers

ps did I mention I heart houdini gas solver?

Edited November 18, 2008 by keltuzar

[keltuzar]

so here is a new exploration, where I use a low res sim, and used a low res vel to advect the high res sim...

Here is a playblast

http://www.serrao3d.com/TestMovs/upscale1.mov

Here is a side by side of the high to low res

http://www.serrao3d.com/TestMovs/upscale2.mov

Cheers

Edited November 24, 2008 by keltuzar

[keltuzar]

so I spent a good amount of time trying to understanding what is going on under the hood for the whole flame front/ gas dsd solver.... and I still haven't completely grasped the whole understanding...

But anyways here is a quick 100 frames of flame sim using fuel that adds to the flamefront which emits temperature.

What I love about the gas dsd is the flame front/surface(fuel) relationship. If someone here can better explain it to me it would be really helpful.

http://www.serrao3d.com/TestMovs/simpleFire1.mov

[melazoma]

so I spent a good amount of time trying to understanding what is going on under the hood for the whole flame front/ gas dsd solver.... and I still haven't completely grasped the whole understanding...

But anyways here is a quick 100 frames of flame sim using fuel that adds to the flamefront which emits temperature.

What I love about the gas dsd is the flame front/surface(fuel) relationship. If someone here can better explain it to me it would be really helpful.

http://www.serrao3d.com/TestMovs/simpleFire1.mov

Check out the 2008 Master Class on Fire. It goes into the gory details of flamefronts and DSD...

[Aliencorpse]

I would really like to see these files, I'm trying to get this to work in houdini 12 and I'm haveing a lot of trouble getting the flame fronts to work properly

Any help would be great.

I want to get this dsd fire working in houdini

http://physbam.stanford.edu/~fedkiw/papers/stanford2007-02.pdf