Torchfire script for feedback

Eddy for Nuke - TorchFire on Vimeo .

Hi all,
I’m working on this torchfire combustion script for tutorials, and would be great to get some feedback or tips to make it better before I do some training and tutorials with it. I’m not using initial state at this time, but I’m considering it so the initial ignition isn’t as large.

Here is the nuke script and data for anyone who wants to take a look.

EDIT: updated pathing on camera node



This is a great example, thanks. What’s the reason for the plane collider under the torch?

Hi frank,

The collider ground plane it to affect the way the combustion happens/moves. Without the plane the heat will run straight upwards, like more of a stiff straight boring candle. With the ground plane the air comes in from the sides and then up, creating more interesting licks and curls to the combustion.

Christoph will be able to explain it more accurately i’m sure.

1 Like

Also there will be an updated one of these scipts soon once the UV stuff goes in fully. Painted some eddy emission maps and improved the textures in substance.

interesting, thanks. would have never thought about that. So you are effectively creating a bit of suction from the sides by blocking off the air flow from the bottom?!

Yeah that’s my understanding, pretty neat. Eddy’s calculating the pressure and ambient temperature around the combustion to some extent.

Unsure how this relates to sparse grid bounds/padding and if you need the collider within the bounds though. You may have to increase the padding aggressiveness/size on the element for this to actually do anything significant.


Cool, thanks. I’m just dicking around with setting a basic sphere on fire in several ways to understand the basics. Can’t wipe the grin off my face while I’m doing it :grin: :

It is heaps of fun :slight_smile:
Here’s a basic combustion i was tinkering with this week, think i got the settings sitting fairly nice, especially when you turn up the voxel resolution.

thanks, just rendering it. why did you set the fuel and temperature emitters to be guides in this one?
Also, how would you create fire without any smoke?

fuel carbon amount seems to be the ticket with controlling the amount of smoke

I’m learning guides are more useful than just for velocity. They guide the values to what you input, as opposed to just adding and building up values in the simulation. Using guide weak on a smoke stack for example keeps it a steady drift, as opposed to building up a billowing smoke.

Yep, its density that creates the smoke, fuel carbon amount turns fuel into extra density when it is spent.
Also for tweaks you can play with extinction coefficient in the fire shader.

cool. One more: I set the fuel expansion in your fire sim to 0 but the fire still spreads.
Is that to e expected?

yeah that’s normal, fuel expansion makes it grow as it ignites, you can use it to tweak fireballs/explosions overall size… the fuel will still ignite as normal without it though.

so even though the fuel itself doesn’t cover more than small spherical volume in your density the fire still seems to spread. In my test it doesn’t. Are there other parameters that influence this behaviour?

Temperature and density alone is all that’s needed for a combustion. Think of fuel as a multiplier to the temp.
The ignition temperature on E_Element may help you there, i usually leave that to default of 50 though.

You may not have enough temperature in there, or shader setting may be off.

Happy to do a quick google hangout if you like?, have a look at your script.

Ok, thanks. I will play a bit more first (just poking it during render breaks at the moment).
Once my current deadline is over in about a week I may take you up on the Google Hangout offer.

Hi Frank,

just a couple of guides:
Fuel gets advected as well, so when a combustion event takes place the uncombusted fuel amount gets carried along with the density+temperature and more can combust until it hits 0.0. Fuel combusts when the temperature reaches the fuel ignition point, generating a combustion which adds the combustion temperature amount/second to the temperature channel. values are in kelvin to match the planck blackbody radiation values of the fire shader ( which expects kelvins ). the fuel expansion is a model that emulates the expansion of the gas during the combustion event. usually I keep it below 0.5 as it can go nuclear pretty quickly, when you have too much fuel around and your burn rates aren’t high enough to get rid of it. The fuel carbon amount is as you’ve noticed responsible for adding soot/density into the system. also worth playing with is the fuel/air threshold, which limits to the combustion event to region that have access to oxygen. sorry not sitting in front of it atm, so I hope I recalled the parameter names correctly.


1 Like

Ok, thanks guys, I’m starting to get the picture. I just removed all fuel from my sim and cranked up the temperature fields which creates nice localised fires which seems great for spot fires that are already burning (as opposed to the actual ignition/combustion event).
Will keep playing for sure, this is too much fun…

1 Like