Cascade problem

[Swihartc]

Hi, I am looking for a solution to what causes this issue? I have 2 simulations set up to cascade. These are the following things I have kept in mind:

1. Both have the same SPF
2. Both have the same resolution cell size
3. Both have the same grid export settings
4. I have even overlapped sim1 far enough into sim2 that a step is generated inside.

The problem I'm having is continuity. You can see a gap from sim1 and inside sim2. Also, the arrangement of particles in the grid seem to have become an array instead of random.
The resolution isn't huge as this is a test (1.5 million grid) but I'm doing this test because the same thing was happening in a project file that had grid sizes of 50-80 million.
I simulate sim1. Then simulate sim2 with cascade pointing to sim1. Any other ideas other than what is on the help document about cascade? I followed all the tips there to be sure I didn't have a grid channel not consistent or anything else.
* Simulated with Phoenix 4.40.00, Build ID: 20210407 for 3ds Max 2022 from Apr 7 2021

[georgi.zhekov]

Hey,

Could it be that your liquid is moving too fast and the simulation can't keep up? Does it help if you increase the steps per frame? How about if you increase the overlap of the two simulators?

Can you try with the latest Phoenix nightly build? Here is more info on how to get it - https://forums.chaos.com/forum/phoen...nightly-builds

[Swihartc]

georgi.zhekov Its possible. I'll try something really high like 100. I did these tests at 20spf and I thought it would be fine as its not moving that fast. It takes about 10 frames to reach the bottom of the simulator, so each frame filled about 10% of sim1 grid. Then when overlapping into sim2, i made sure to overlap it enough that a full frame was generated inside sim2 to give it enough info to transition. If 100 spf doesn't work but then trying a really high value like 200 spf does work; I'd need to then decide if cascade consumes more time than a single simulator with high res but lower spf.

[georgi.zhekov]

Ah, 20spf should be enough for most cases. If the latest nightly build doesn't help, can you send over the scene so that we can check what is going on?

You can either attach the scene here as a zip file or send it through the support form here (there is a "Submit a request" button at the top right) - https://support.chaos.com/hc/en-us

Thanks!

[Swihartc]

I downloaded a nightly and gave this test a shot again.
* Simulated with Phoenix 4.41.02 Nightly, Build ID: 2022031131297 for 3ds Max 2022 from Mar 11 2022

The results were still pretty much the same. It seems like the particles do not respect the previous simulation spf or inherited particle position.
Both simulators have only liquid and splash particles exported, both at the same spf, and both at the same resolution per cell size. The only difference is the cache save location, the grid dimensions of sim2 and that sim2 points to sim1.

I have attached the max 2022 test file. Hopefully you find my mistake to get a cleaner cascade transition.
Thank you!

[georgi.zhekov]

Thanks a lot for the scene. Looks like you have very fast moving liquid and unfortunately fast movement can't really be caught precisely because the second simulator reads the cache files files from the first one.
The cascade relies on the cached full frames, but there are no substeps involved, so If the liquid moved too fast from one frame to another the second simulator won't be able to keep up. I hope we can improve this in the future.

As a workaround for this maybe you can overlap the two grids even more so that the tall one comes closer to the bottom of the wide one. Then for the meshing I would recommend to check this video here - https://youtu.be/XzgoLgnXWag

Hope this helps.

[Swihartc]

georgi.zhekov Is there a way to make a stream of particles go slow enough to cascade properly, but to hit a collision surface with an altered velocity effect? Much like how you can set a splash to be more violent by adjusting the velocity setting on a collision object. (image)
I see the liquid source has a velocity spinner, but it does not add velocity pressure to a stream of water in motion when it hits a collision object. Anything like that or a trick to achieve that? It would not necessarily be physically accurate of course, but for still images of faucet streams at high pressure hitting a surface and splattering violently but using a cascade set up to have the original liquid stream moving slower. That is why my water speed is so high; to get a very high pressure collision of water particles on a surface. Or do I just need to use a very large single simulation grid?

[georgi.zhekov]

Ah, in this case sounds like it will be best to use a single simulator. You could possibly increase the velocity of the liquid when it gets close to a certain object with a Particle Tuner, but I guess this could be an overkill.

[Svetlin.Nikolov]

Just wanna drop in on this - the "Motion Velocity" you see on the Source and the Node Properties means velocity that is added by the motion of the geometry. If the emitter or obstacle geometry stands still, it will not do anything. Here is more info: https://docs.chaos.com/display/PHX4M...ode+Properties