Announcement

Collapse
No announcement yet.

Liquid: a few questions about flow and shape

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Liquid: a few questions about flow and shape

    How do you actually control the volume of fluid flowing from the source? I'm using temperature as the liquid source and set it to 1, and have been playing with the discharge value. It seems like you can increase it to some degree but it's not predictable, and I can't seem to get it lower than what I get with 1.

    Even with discharge at .1 it seems same as at 1 or 10

    See the vid clip below: the volume coming from the small emitter is pretty crazy. There are also some strange behaviours (seem weird to me anyway). THe stream of fluid breaks off into 2-3 streams at some point and doesn't seem natural. The parameters should be visible on the left of the frame. I based these on experimenting with the values shown in that set of parameter variable videos on youtube.


    Questions:
    1) how to control fluid volume up and down?
    2) how to get the fluid in this example to flow more like a normal (water/beer) kind of fluid?

    Thanks /b

    http://screencast.com/t/za44y9lQK9db
    Brett Simms

    www.heavyartillery.com
    e: brett@heavyartillery.com

  • #2
    I'm testing the fixed SPF approach indicated in another thread (and from the paint sample file). I can't get a handle on how to employ the maximal step in combination with fixed SPF. Some specific (dumb user oriented) direction on that possible?

    What exactly does SPF actually mean by the way? (I know it means 'steps per frame', but what does it mean in practical terms?)

    Thanks /b
    Brett Simms

    www.heavyartillery.com
    e: brett@heavyartillery.com

    Comment


    • #3
      I'm finding another odd thing: the liquid in my test is flowing down and then curving off to one side for no obvious reason. I have vorticity set to 0 for this one.

      See the screen cap to see the motion.

      How do I get it to flow more straight/with gravity?

      /b

      http://screencast.com/t/zNEm5OdGXd
      Brett Simms

      www.heavyartillery.com
      e: brett@heavyartillery.com

      Comment


      • #4
        - the liquid quantity is fully controlable with the discharge, that you said is fully unexplanable for me. if the discharge is very low the surface continues to release some small amounts of liquid (because of this stopping a source is a bit complicated) but with normal discharges the amount control is out of question.
        - the weird behavior looks like preview problem, if you have very low value of the preview threshold, you will see the liquid vapours in the air and the simulation will look like this. but i will check your scene to see for other problems.
        - if you force the spf value via the limits, the advection step (the maximal step) does not matter. each simulator has a value like SPF, PF has it and if i'm not wrong it is called "integration step" or something like this. in the most cases this value us an integer. the only difference between phoenix and the other simulators is that here SPF can be non integer and can be auto calculated from the velocity of the fluid.
        - i suppose you have turned on the surface tension . honestly the surface tension has serios issues, i recomend to be not used yet.
        ______________________________________________
        VRScans developer

        Comment


        • #5
          Originally posted by Ivaylo Katev View Post
          - the liquid quantity is fully controlable with the discharge, that you said is fully unexplanable for me. if the discharge is very low the surface continues to release some small amounts of liquid (because of this stopping a source is a bit complicated) but with normal discharges the amount control is out of question.
          I will send you my current file - I can increase the amount but I need to go a long way to see much difference (like 1 to 100 or even 1000) but going below 1 to reduce flow seems to have no effect.


          - the weird behavior looks like preview problem, if you have very low value of the preview threshold, you will see the liquid vapours in the air and the simulation will look like this. but i will check your scene to see for other problems.
          I unfortunately didn't save that file separately as I meant to (continued playing and saved over it without thinking). At the moment that weird curving thing is not happening. If I get it to do it again I will save and send you the file. I don't think it was just the preview though ,as it rendered that way too. Regarding the preview: how do you determine the "correct" range to show? The max is set by the source value (1 in most liquid cases I guess) but what about the min? A ton of my fluid disappears unless I set it to something like .001 or even .0001 Where should it be?

          - if you force the spf value via the limits, the advection step (the maximal step) does not matter. each simulator has a value like SPF, PF has it and if i'm not wrong it is called "integration step" or something like this. in the most cases this value us an integer. the only difference between phoenix and the other simulators is that here SPF can be non integer and can be auto calculated from the velocity of the fluid.
          - i suppose you have turned on the surface tension . honestly the surface tension has serios issues, i recomend to be not used yet.
          Surface tension was certainly on, as was Viscosity. I'm really struggling to figure out the meaningful values there. ST seemed to need to be in the range of 1-4 but Viscosity was having a marked effect even at .01-.02 I think it would be very helpful to normalize the value range (behind the scenes) so that as much as possible the meaningful range of all parameters are within the same world (at least for ones that can be done for).

          Realflow does have the integration step value, but it's not something one has to mess with much in terms of shaping liquid - it's more like an optmization thing in my experience, but I am not a huge expert by any stretch. More or less you played with the viscosity, surface tension, particle density/mass, stickiness (for interactions) and friction. The rest came down to forces and emitter shapes/animations.
          Brett Simms

          www.heavyartillery.com
          e: brett@heavyartillery.com

          Comment


          • #6
            Regarding the preview: looks like I don't get that either I have some fluid that looked to be disappearing partway through the sim. The liquid is derived from temperature, and that is set to 1. Discharge is at 10 but that doesn't seem to matter.

            In order to see the fluid I have to actually set the preview at max 5 and min .01 - then I get roughly what shows in the render. Why would the preview need to show temps as high as 5?
            Brett Simms

            www.heavyartillery.com
            e: brett@heavyartillery.com

            Comment


            • #7
              usually the liquid preview is 0.5 - 2. values above 2 are unusual for the liquids, the values below 0.5 are invisible in the rendering (if the surface level is 0.5).
              ______________________________________________
              VRScans developer

              Comment


              • #8
                Those values are producing virtually nothing in my scene. At .5 there is amost nothing at all and at current I had to set it to .002 to get the full liquid to show up in the render. Surface level of .05 only seemed to work with a very lo-res sim, not sure why.
                Brett Simms

                www.heavyartillery.com
                e: brett@heavyartillery.com

                Comment


                • #9
                  Just posted another screen cap. I have been playing around quite a bit with fixed SPF but the fluid behaviour just never seems quite right for the pour. I went back to Maximal Step of 5, and max/min of 5/2 (these values are just guesses really, I cannot really predict what will happen when I change them independently. When in fixed mode the result is far easier for me to anticipate).

                  Now the fluid behaviours seems kind of erratic. At the beginning it almost explodes out, then settles down into a smooth pour, then goes wobbly again. I have no idea at all how to control it anymore. Any pointers on what might cause this behavour? Vorticity is totally off now, as is viscosity, and surface tension is also off.

                  b

                  http://screencast.com/t/RV0azKwm4DVG
                  Brett Simms

                  www.heavyartillery.com
                  e: brett@heavyartillery.com

                  Comment


                  • #10
                    use classic advection, you have liquid in a vessel, no chances with FT
                    ______________________________________________
                    VRScans developer

                    Comment


                    • #11
                      That is classic advection
                      Brett Simms

                      www.heavyartillery.com
                      e: brett@heavyartillery.com

                      Comment


                      • #12
                        actually... that might not be classic. I'm having trouble keeping the input/output straight once the $implicit saving configuration gets broken. I think that may have been loading a an FT version...

                        So I redid it with Classic for sure, but I'm really getting some weird stuff. I was trying to experiment to see if using other channels than temperature would help the liquid. I tried smoke and fuel but was not getting anything at all.

                        However, if I set the Liquid to use Smoke it totally changes the way the *temperature* simulates. I get a *way* better result, but in order to see it I need to raise the preview max temperature up to something like 200 (even though it's only at 1 in the Source object output).

                        The video shows the difference.

                        now the problem is that this nice smooth fluid won't render - again, see the video.

                        What am I doing wrong here?

                        Thanks /b

                        http://screencast.com/t/tuIr4vTYQ
                        Brett Simms

                        www.heavyartillery.com
                        e: brett@heavyartillery.com

                        Comment


                        • #13
                          - in my haste I made mistake in that video: when trying to show that the surface level wasn't fixing the render problem I was adjusting the displacement amount by accident. It still doesn't work with the surface level though.

                          To get anything but a solid block I have to put that level up around 200, which gives me a weird fluid inside the block. Higher than that and it just started getting more and more artifacted.

                          http://www.evernote.com/shard/s2/sh/...ScreenClip.png
                          Brett Simms

                          www.heavyartillery.com
                          e: brett@heavyartillery.com

                          Comment


                          • #14
                            Any direction/input for me here? I'm kind of stuck with this one.

                            Thanks /b
                            Brett Simms

                            www.heavyartillery.com
                            e: brett@heavyartillery.com

                            Comment


                            • #15
                              are you sure that the liquid channel and the effects channel are the same?
                              ______________________________________________
                              VRScans developer

                              Comment

                              Working...
                              X