I have a problem with fluid simulation.

The project is pretty simple, I have just a domain object and an inflow object.

The inflow object is supposed to start drop oilish fluid at frame #1100 and ends after 8 seconds.

It was working almost correctly until I moved the inflow object a little (inside the domain), after that it started to act in a weird way.

The inflow seems to behave if it was a long tube (instead of a little sphere) and the fluid seems dense and is not following the initial velocity settings.

As you can see in this picture that strange long melting tube is supposed to be a flow of oil sprinking in the direction (x, -y, -z).

enter image description here

I've already tried to delete the bake cache, restart Blender, using different OS (Windows, Linux and macOS), but with no luck.

You can download the project file here.

  • $\begingroup$ Have you Applied Scale on each of your mesh objects (inflow and domain)? Odd scaling can cause odd behaviour. $\endgroup$ Aug 17, 2018 at 6:16
  • $\begingroup$ Yes, I've scaled the both objects after I've assign it the fluid type, just to fit better in the scene. Is it a problem? I rebaked the simulation (removing the cache). It would be a bad/odd limitation requiring to not scale object of simulation. What if baking the simulation you realize that the inflow or domain object has a wrong size? $\endgroup$
    – Kill KRT
    Aug 17, 2018 at 9:15
  • $\begingroup$ Changing the size before the bake isn’t a problem providing you then CTRL+A and Apply Scale. This ensures the scale is 1.0 in each direction - otherwise the skewed scale can cause problems with the simulation. If you need to change the elements of your simulation you should always re-bake anyway - otherwise the simulation won’t be accurate to the new situation. $\endgroup$ Aug 17, 2018 at 9:23
  • $\begingroup$ Sorry, maybe I didn't get your comment. That is what I have done: I changed the size of both objects and then I rebaked the simulation (removing the cache folder and pushing again the [Bake] button), but anyway it seems to acting in a weird way. $\endgroup$
    – Kill KRT
    Aug 18, 2018 at 1:50
  • $\begingroup$ Problem appears to be with a negatively scaled Inflow object - I've added an answer that should hopefully resolve your issue. $\endgroup$ Aug 18, 2018 at 6:43

1 Answer 1


The problem appears to be due to the geometry of the Inflow object - it has a negative scale (-0.025). This causes havoc with the Volume Initialization of the inflow in the domain.

The simple solution is to select the Inflow object and simply press S-1Enter. This will 'flip' the scale of the mesh so that it is enclosing the correct volume.

Re-baking produces the following :


When using Volumes it's imperative to ensure the Normals are facing in the correct direction - in your case the negative scale resulted in them pointing 'inwards' rather than 'outwards', confusing the fluid sim.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .