View Full Version : maya fluids / how to "suck it in"?

03 March 2011, 05:13 PM
hello talented and generous folks around cgtalk!
maybe someone could help me:

is there a good way to first emit fluid
into a container, and then have it sucked
back in (and disappear) at another place
within the container?

something like a "hoover force field"? ;)

thank you for any help!

03 March 2011, 05:19 PM
if you didnt need to see the emission of the fluid you could just emit then run the cache backwards...

if you did need to see the fluid you could cache two simulations then do a cache blend between the two to get one emitting in the other (reversed) sucking it out.

id need to play with it but this approach should work.

03 March 2011, 05:43 PM
thank you mandark!

although...not quite sure whether this could really work?

what I would like to achieve is a smoke plume that is emitted for.... say 500 frames.
it collides with a transparent object, surrounds the object and partly occludes/fogs it this way.
then the object "wakes up", and sucks the smoke in....I would love to be able to see the smoke inside the transparent object.

hmm...not so easy I assume...very intrigued to hear a solution, even if it's only for parts of the problem :)

thank you again!

03 March 2011, 05:58 PM
now that you have explained what you need better....

if using 2011 try emitting fluids with a high dissipation from particles..(duncans example in whats new in 2011) and drive your sim with particles instead may be easier to control.

03 March 2011, 06:21 PM
will give it a try, thank you for the hint!! : )

03 March 2011, 06:38 PM
You could suck fluid to some degree with either a field or a fluid emitter with velocity. However think of it the way you would a real vacuum. If you just put a fan at the mouth of a container it would not suck anything in(due to the incompressibility of the air inside)... you need a hole for the old air to move out of the container. If you don't want smoke going out this hole you could put an emitter over it with the density method set to replace and the density at zero(or very near zero if you have no other emission on the node). This exit hole emitter could also be a fan by using speed emission (sort of like the placement of a fan on a typical vacuum.

The flow of the air is incompressible which makes it difficult to concentrate if only using a basic fluid simulation. It might help to use forward advection. You might even lower the solver quality attribute a bit which combined with forward advection can allow the density to concentrate and compress a bit.

The particle fluid emitter method mentioned might also work well, perhaps even combined with the above technique. It would allow you to trap the particles and compress the dust. The fluid could be used as a field to push the particles, much as air pushes dust particles. A filter at the vacuum exit allows air to pass but traps the dust.

03 March 2011, 06:46 PM
thank you Duncan, will try this!

03 March 2011, 07:45 PM
Here is a scene file that does the vacuum thing for a basic fluid (no particles).
Note that it can move the fluid to the inside, but the "fan" emitter also removes density as it exits the vacuum so that it never concentrates inside and eventually it gets fully eaten up(if only a real vacuum could work this way).

Concentrating dust would require additionally using particles, I think.


03 March 2011, 09:11 PM
cool setup Duncan!
was trying to figure this out when i did a smoke ring test -
but couldnt quite get it right... tried to fill an object (mouth made of a sphere with a hole) with fluids and then push it out by scaling/squashing the mouth but couldnt get the fluids to push correctly... (when are we going to get Advection in Maya Fluids?)

would be great if you could add a feature to do this same sort of suction or pushing without having to kill the density...

03 March 2011, 10:34 AM
really cool vacuum Duncan, merci beaucoup!

03 March 2011, 06:36 PM
would be great if you could add a feature to do this same sort of suction or pushing without having to kill the density...

Actually a way occurred to me. If one has a closed container one can make the pressure low inside the container using the temperature pressure attribute along with emitting temperature inside the container. Place the emitter inside the container where you wish to have suction and emit temperature(use the temperature replace method) with no density (density emission off). Make the temperature a dynamic grid on the fluid and then set temperature buoyancy, turbulence and diffusion to zero. Make the temperature Dissipation high so it is only at the emitter location and then set temperaturePressure to something like -.5 (the lower the more suction).

The density will tend to compress at the emitter location. In Maya2012 there is a density pressure and threshold one could use to limit how much the "dust" can compress. Also the collision boundaries when using forward advection are more airtight than in 2011, so it would tend to work better.

04 April 2011, 10:43 AM
will try this as well, thank you!

but will it be able to generate more "suction" than the setup you generously uploaded before?
sucking out a static fluid volume is one thing, but what I try to do is to suck in a billowing, uprising smoke plume (with some buoyancy) back downwards into an object it just passed by on the way up.
playing the sim backwards wouldn't work unfortunately.
it's not an actual hoover demo, rather an abstract scene.

will let you know if I had success :)

thank you again,

04 April 2011, 12:32 PM
The negative temperature pressure is generally better for lots of suction. The fan method must fight the incompressibily of the flow to generate suction(low pressure) so you need an outflow. The temperature pressure method instead directly sets the pressure.

Note that with both methods one generally wants a high solver quality setting to better propagate the effects of the suction.

04 April 2011, 02:38 PM
Hey Duncan,
sounds like a good idea. But where do I find the settings to adjust the temperature pressure?
Really want to try this ...
Thank you!

04 April 2011, 10:56 AM
Okay, I found them in Maya2012. And I was able to attach the 2012-caches to Maya2011-fluids, so no need to switch the pipeline :-)


CGTalk Moderation
04 April 2011, 10:56 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.