3
$\begingroup$

Original Post

This is the problem that led me to post this.

Whenever I scale the domain of a smoke simulation, the inflow object's smoke also scales up/down. Applying the scale of the domain does not fix the behavior.

Here's an image:

enter image description here

Notice how large the smoke boundary is.

Here's a video of what's happening:

https://streamable.com/3akto7

More issues I've encountered with Mantaflow:

  1. When you set the cache method to "Replay", the simulation won't bake when pressing play, like it should. Sometimes, if you manually drag the current time marker to frame 1 and then 0, and then press play, it will bake. This was a nightmare so I almost always use a Modular Cache as it seems more reliable.
  2. If you scale the domain arbitrarily, it can mess up the entire simulation. Smoke or liquid will flow influenced by the new size of the domain, rather than the exact size of the inflow object. Applying the scale on your inflow object and domain does not fix the issue. The only solution I've found is to recreate the simulation entirely.
  3. Collision objects don't work properly despite being set up correctly. For some reason I have yet to identify, sometimes collision objects will just get completely ignored by the simulation. No collision happens at all, and you have to recreate them, hoping it will work. Sometimes, it will work, but liquid will "leak" through solid collision objects.
$\endgroup$
5
  • $\begingroup$ Perhaps you should leave the "My Question" section out since it is still very broad. Although, I sort of have an answer for it: For doing smoke sims reliably, I would recommend watching tuts from this guy: youtube.com/c/Iridesium He has some of the best Mantaflow tutorials that I've seen and really explains things well. Also, issue 2 sounds very similar to the problem described in the first section. $\endgroup$ – HISEROD Dec 22 '20 at 5:59
  • $\begingroup$ Have you rebaked your simulation? I had some issues when changing various parameters and having the fluid particles scale wierdly. I remember changing from replay to cahe and then back again to clear the "old bake". $\endgroup$ – DrDress Dec 23 '20 at 9:42
  • $\begingroup$ I have no idea what mantaflow is, but EEVEE seems to encounter similar non-refreshing problems. $\endgroup$ – ZargulTheWizard Dec 26 '20 at 23:55
  • $\begingroup$ @ZargulTheWizard Mantaflow is the built in fluid/gas/smoke solver for Blender. $\endgroup$ – Jay Dec 27 '20 at 21:20
  • $\begingroup$ :-P I meant FLIP. $\endgroup$ – ZargulTheWizard Dec 27 '20 at 21:47
1
+100
$\begingroup$

These are mostly cache refreshing problems, but for now there are workarounds:

  1. What you said. Bake the simulation manually. I usually use "all" since I still don't understand the types.
  2. In the case of an outdated cache, change something like the "resolution" factor then change it back to force the cache to reload.
  3. This seems to be limitations of the engine. don't make fluid move as fast, increase substeps, etc.
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.