Liquid: a few questions about flow and shape

[Ivaylo Katev]

no, you changed the liquid channel, the source is that with the barrel icon. when you use the liquid as temperature, it loses its usual meaning and the simulation is temperatureless, but when you use the fuel as liquid, then the simulation have temperature and the result is different

[simmsimaging]

So it actually does matter which one you use? If the temperature creates a different result then when using fuel as the source (set to 1) what should the temperature be set to?

[Ivaylo Katev]

just uncheck the temperature in the source. anyway, just for the information - 300.

[simmsimaging]

ok - thanks.

What does the gizmo do (rendering section)? How do you employ that to smooth out liquid/container intersections? I tried using the container as the gizmo but that didn't work at all.

b

[simmsimaging]

The result seems more consistent with all the other channels turned off - which clarifies a few things. Thanks.

However two things: I still think the temperature channel looked best for dynamics and flow of liquid when it *was not* the liquid source. It just didn't render well, but the fluid itself looked like a better starting point to me anyway.

The other is still problems getting decent flow. See the screen captures. First one shows the result I got in lower res (175K cells) to set things up. The second shows what happened as soon as I increased the res t0 350K. No other changes.

What exactly do I need to adjust to get it flowing more smoothly like it was with the lower res grid?

thanks /b

[Ivaylo Katev]

if the simulation changes its behavior when the resolution is increased, this may refer to reached spf limit. if you simulate in fixed spf mode, you have to change the spf by hand with the same ratio as the resolution (linear) was changed.

[simmsimaging]

Thanks, that is good to know. I'll try it.

What about the gizmo question?

[simmsimaging]

if the simulation changes its behavior when the resolution is increased, this may refer to reached spf limit. if you simulate in fixed spf mode, you have to change the spf by hand with the same ratio as the resolution (linear) was changed.

I had basically doubled the res of the grid so I doubled the SPF (from 5 to 10). It's perhaps a bit closer, but I'm still seeing an odd splitting of the liquid into two bands that doesn't really make sense to me. You can see the liquid and simulation parameters in the screen cap - what needs to be changed to get a more natural flow?

Thanks
/b

[simmsimaging]

no, you changed the liquid channel, the source is that with the barrel icon. when you use the liquid as temperature, it loses its usual meaning and the simulation is temperatureless, but when you use the fuel as liquid, then the simulation have temperature and the result is different

I have tested this again more carefully. My experience is that PH definitely generates different flows of liquid depending on which source you choose for the liquid, and this is true if other channels (beside the fuel source) are active or inactive in the source helper. The difference is just much greater when multiple channels are active. From your earlier comments it sounds like this should not be true - so perhaps there is something else going on?

[Ivaylo Katev]

you didn't double the resolution actually your factor is 1.25
but even with correct spf the result will be not exactly the same, just the behavior will be the closer one. about the channel differences - well, if you think so then you are lucky, just select this one that produces the best result

[simmsimaging]

OK, I give up.

I think I'll just wait and check back in on Phoenix for version 2 or something.

Thanks/b

[Paul Oblomov]

It is exactly THE SAME. Temperature and fuel used as source.

[simmsimaging]

Fuel:


Temperature:


These are not THE SAME. I asked before if there was something else that could cause this and got no answer, but I am quite sure that simply changing fuel for temperature in the liquid source, and switching one or the other on in the source helper, leaving all else the same, can produce different results.


/b

[Ivaylo Katev]

in general, the multithreading can cause some result differences, because the numerical computations give not the same result for a+b+c and c+b+a, i.e. the order of adding does matter. however, this produces differences that are smaller than in your image, so i still have no explanation.

[Paul Oblomov]

I think we need a scene.