Using one sim as start state in another

[Syu_z]

Hi, apologies if this has been asked before. I did some cursory searching and couldn't find anything relevant except the waterfall cascading sim, although I am not sure how that applies to my case.

So I simulated two RGB smokes that collide in the middle and form a sort of cloud with a mix of the RGB color between the two. So far so good. Then on the next scene, this colored cloud needs to collapse to the ground, and created an explosion. My question is, can I take the end frame of sim1, and use it as a start frame in sim2? I tried adding sim1 that i have loaded into the scene as PHXsource, and using a very small volume inject value, but the results are not even similar to sim1 smoke.

Is there a method where I can just specify sim1 end frame as the start frame of sim2?

[georgi.zhekov]

Hey,

You can use the Load and Start option in the Simulation rollout of the Simulator node.

This way you can load a cache file to act as an initial state for the second simulation. You can find more here - https://docs.chaosgroup.com/display/...oke+Simulation

[Syu_z]

Thanks for the reply. That sort of works, but the sim grid becomes extremely large when it has to expand and it's almost impossible to tweak, so I have to think of another solution.

I have another question, what is the recommended way to do colored cloud collision to the ground? I initially tried to kind of mimic the initial two smoke colliding, but angled towards the ground so they are accelerating to the ground, but the result looks too uniform (it is lacking the blobs created by two smokes colliding in the first scene, probably because it doesn't have time to brew and grow) and it did not really expand nicely after collision. Does using gravity/smoke buoyancy is the way to go to create a natural smoke collision to the ground? One way that sort of works was to emit smoke from a displaced mesh, and then translating the mesh down (while using buoyancy and gravity to kind of match so I don't see trails of smoke behind).

I am open to another solution if there is any. Kind of stuck right now.

[Svetlin.Nikolov]

It is extremely subjective to define how a nice smoke behavior would look like in text, so if you have reference images and share images or video of the simulation you are getting currently, it would really help us recommend which options you need to tweak Thank you!

[Syu_z]

You are right, it's difficult to verbalize. But it's alright, I managed. Another question though, I need it the smoke to change color gradually halfway to another color. Based on the article here
https://docs.chaosgroup.com/display/...+RGB+Workflows
I need to create a second PHXSource for fire, add a non-solid box, Volume Brush, and only output RGB. I have double confirmed that the box is non-solid from phoenixfd properties, but the smoke is still blocked by the box. Any idea why?

Edit: Looks like I found it. Untick Motion Velocity from the Source as well. I cannot find that info inside the article, but it's unticked in 1 of the screenshot, maybe I missed it or it could be updated.

Edit2: Another issue. It works now, but only with 100% volume brush. When i tried any other value, like 50, 90, 99, 99.99, they either have no effect or negligible effect.

[georgi.zhekov]

Ah, the Motion Velocity was off by default in previous versions of Phoenix and now it is on by default. You can find more about what it does here - https://docs.chaosgroup.com/display/PHX4MAX/Fire+Source+%7C+PHXSource#FireSource|PHXSource-Emission
The motion velocity checkbox shouldn't block the smoke though - is this the only thing that made a difference?

As for the Volume brush not affecting the RGB when the brush effect is below 100% - this is a known issue and it is logged in our bug tracker system already.
In case you want to change the smoke color gradually then you can use either the Mapper approach as shown in Chapter 5 of the RGB Workflows tutorial or if you're using Phoenix 4 you can use the Voxel Tuner.
You can check this video here showing how the Voxel Tuner works - https://youtu.be/mGAyh3uR89U?t=851

[Syu_z]

Yes, it's the only thing from my memory. Once I uncheck it everything works.

I am still using older phoenix so I can't use the Voxel Tuner, I tried the Chapter 5 with VrayDistanceTex and PHXMapper, but it doesn't seem to be continuous? Let's say my initial smoke is green, and I change the build up time to 0.1, and my map is VrayColor Pure White (255). It doesn't affect the green smoke to gradually change to be full white, it just becomes a less green smoke, and it doesn't really do a gradual mix. I tried increasing the size of the box, thinking the effect only happens when it's inside the box, doesn't seem to make a difference. Plus there seems to be a very distinct seam (which can theoretically be fixed by using a more complex geometry rather than box). Did I do a wrong step somewhere?

[georgi.zhekov]

By any chance do you have the scene where the Motion velocity prevents the smoke to enter the box?

Here is a sample scene with the Mapper changing the RGB. Using the build up time you can control how fast the smoke will change its color.
Value of 0.1 (sec) means the color will change over 3 frames which is quite fast, so in my scene I have set the build time to 0.5(15 frames).

Hope this helps!

[Syu_z]

Hi georgi, apologies for the extremely late reply. I had it in the back of my mind and kind of forgot about it.

I don't have the scene anymore, but i think it's easily reproducible. It was just phoenix quick sim setup for a sphere, the quick sim for fire simulation, and then i tried turning off the temperature, and created a box, make it non-solid and all the works, it already doesn't work from there. PhoenixFD 3.14.

Thanks for the scene, it helps. Now I am stuck with another issue. Client basically wants something like this, but without tail.

My approach is, two cylinder as mesh source, with polygon id to determine the surface it's coming out from, and increase the surface force really high for 2 frames, and cease the output. But the issue is that, the speed is not nearly enough to traverse the distance while maintaining the shape. It's difficult to get a ring explosion in the middle if they don't collide perfectly. If i increase the surface force amount, the shape becomes much bigger because, understandably, the output is more now.

My question is, how do i get a small sized smoke ball that travels at high speed, if increasing the surface force is not doing it well?

I tried with default wind+decay, and it travels faster but the shape it created is ugly because i assume it's being pushed from the back.

[Svetlin.Nikolov]

Hey,

There are two factors to more realistic simulations - more conservation quality and more steps per frame. Both will make the simulation take longer, so watch out, especially with the steps per frame!

Cheers!

[Syu_z]

Hi Svetlin, so if i am understanding you correctly, the problem is that it needs to be more realistic? I somewhat felt that the issue is a wrong approach, so I was curious if there is a better way of achieving it. How do you achieve a small ball of smoke traveling at high speed? The more surface force i input, the bigger the ball is going to be.

[Svetlin.Nikolov]

When a fluid solver does less steps and uses less conservation quality, rolling vortices tend to grow large and dissipate more quickly - this is why I'm connecting realism with your setup. Indeed there is no "vortex scale" option that would directly influence the size of the plume, so you gotta control it indirectly. Please try these option and share you results, and if this does not solve it, at least we will know how these options affect your particular setup so we'd know how to move on from here and try other things

[Syu_z]

Understood, thanks for the tip. Seems like it made a difference, testing further.