Wierd fluid sim problem

For some reason, fluid simulations refuse to bake with any version of blender on my installation of windows 7. I have installed versions of blender from 2.5 to 2.7 ad even 2.49, but my fluid sims always go bad after about the 30th frame. I will upload a scene where you see the problem pretty clearly. after like 30 seconds or so of baking at around frame 25 the fluid starts to spread over the side of the domain. The whole side of the domain is covered by frame 40 or so. This also for some reason causes the baking to slow dramatically. eventually the fluid becomes chaotic enough that the bake just stops for hours and never reaches the end of baking. anyone know why this is happening?

[ATTACH]144241[/ATTACH]

Unless you need the domain to be 10 meters (blender units) large I would reduce the size to about 0.5 meter. In any case, you need to apply the scale (Control-A) to the domain and the sphere before you bake.

nope thats not my problem. Created a whole new sim and applied the scale to the doamin and fluid, and it did the same thing. Also, I can bake exactly the same scene in linux on the same computer, and it works fine at 200+ resolution

anyone have an answer???

I have similar problems and described it more in my thread “Fluid simulation unstability. Crashes, broken baking and system freezing.”

I tried your fluid.blend file using the latest svn version of blender and also the official 64 bit 2.57b blender version in Linux (Ubuntu 64), and it baked with no issues; took about 3 minutes to bake all 250 frames.

The only odd thing that I noticed was that the Z gravity value in the Physics > Domain World section was displaying 0 for the Blender svn version, while this wasn’t the case for the Blender 2.57b version.

http://users.xplornet.com/~gimble/randosity42_fluid_blend.png

Perhaps it’s an unrelated display issue, but perhaps it might be worthwhile to un-check the Gravity in the Scene context panel, and then set the Gravity Z setting in the Physics > Domain World section to -9.81 m/s^2 to see if it has any effect.

If that doesn’t work, you should probably submit a bug report.

I have the same problem, this happens on both Windows 7 and Ubuntu.

It started happening on 2.57, not earlier versions. The processor is an i7 w/ hyperthreading, so basically 8-threads, if it’s perhaps related to that.

sorry to bump this, but I am still having the same problem. Gimble’s solution did not work. I have a phenom II x6 at 2.8ghz if it helps…

My bake was also quitting in the middle of the animation.

I resolved it by baking at a real low resolution (like 80), and then gradually increasing the resolution.

Also check to make rue that you have assigned the “shell” and “volume” properly, that there is only one domain and that everything is in it, and that the fluid object isn’t intersecting any of the obstacle objects.

Sorry for posting on an old topic, but I’m running an intel i7 2600k based rig and also have the same issue: when I use an inflow object in a fluid sim, the fluid moves directly up one side of the domain before suddenly filling the whole domain and killing the bake.