strange behavior of particles in a glass or cup

Describe your realflow bugs
duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

strange behavior of particles in a glass or cup

Postby duke3d » Wed Aug 31, 2016 12:40 am

I got problem with standard and Dyverso (both SPH and PBD). When I pure water into a glass, particles stick to the it's walls creating a thing film with 1-particle thickness. Here you can see that after water stops moving and stays calm that particles are still on the walls (I generated a mesh to be able to see the problem clearly)
Realflow problem.jpg
Realflow problem.jpg (164.91 KiB) Viewed 1337 times


First I thought that the problem is in the mesh properties. I changed friction as well as sticky to 0.0, didn't help. I increased particles quantity, didn't help. I changed substeps quantity, didn't help. However I got strange behavior of Dyverso PBD particles after increasing MIN iterations. The more! MIN iterations I had the less stable the simulation was (exploding (emitter and particles in a glass) and particles flying through the walls of a glass)! It makes no sense at all. At MIN iterations set to 1 I got the most stable simulation and I got less pronounced . But, after I changed a scale I got the same problem again.
In general, changing any parameters of the particles and objects didn't help me to rid of the issue I have :(

Any ideas what is the reason of the problem?

Realflow 2015

Thanks,
Andrew


User avatar
ChristianZ
Posts: 3276
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: strange behavior of particles in a glass or cup

Postby ChristianZ » Wed Aug 31, 2016 9:47 am

Hi Andrew,

may I ask about the other properties of the liquid particles such as internal-external pressure and viscosity?

I have not seen this behavior so far, but I have more experience with the legacy solvers.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: strange behavior of particles in a glass or cup

Postby duke3d » Thu Sep 01, 2016 3:02 am

ChristianZ
Thanks for reply :) The liquid properties are:

Resolution 10
density 1000
Int Pressure 1
Ext pressure 1
Viscosity 3
Surface Tension 0

Just default values.

Daemon force scale for fluids (objects as well) is 2

I've made another scene. The liquid properties are the same except resolution. I was able to get decent result with standard particles, but Dyerso particles still have the same problem :(

Right now I launched simulation with much higher resolution. I will show the result in few hours.

Thanks

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: strange behavior of particles in a glass or cup

Postby duke3d » Thu Sep 01, 2016 4:27 am

I found that I'm not the only one who has such a problem.
For example in this tutorial https://www.youtube.com/watch?v=SU-DGSfXEpg
right at the end on the video you can see the same issue.

Here is the result of my recent simulation implemented on another computer:
RF_issue_01.jpg
RF_issue_01.jpg (202.09 KiB) Viewed 1316 times

RF_issue_03.jpg
RF_issue_03.jpg (193.27 KiB) Viewed 1316 times

RF_issue_04.jpg
RF_issue_04.jpg (193.42 KiB) Viewed 1316 times

RF_issue_05.jpg
RF_issue_05.jpg (193.54 KiB) Viewed 1316 times

RF_issue_06.jpg
RF_issue_06.jpg (193.11 KiB) Viewed 1316 times


The blue and orange liquids are standard. The green one is Dyverso. Resolution is 30, all other settings are default.

The geometry is a standard Vase.
The first Vase has "Thin face test activated", other settings are default.
The second Vase has default settings.
The third one has "Cell size" set to 0.015 (I played with this one, no changes). Other settings are default.
Scale options are set to default, i.e. 1.

Standard particles look better, but there is still some strange edging.

As for stability of Dyverso particles, still the same, if I increase MIN (even 1 or 2 steps up) iterations particles explode and pass through the geometry.

Thanks,
Andrew

User avatar
ChristianZ
Posts: 3276
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: strange behavior of particles in a glass or cup

Postby ChristianZ » Fri Sep 02, 2016 7:23 am

Ok, I see the problem. First, I thought it might be the type of vessel because the vase with its single-sided wall (shell mode) is an unusual and not very useful type of container. So I made a simple amphora, with a thick wall so that the volume-inside mode can be used, but this effect that you described with the 1-particle thin sheet was very obvious, even after many frames of settling the liquid and friction=0 of the wall.
Image

I will do some more testing over the weekend.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

User avatar
ChristianZ
Posts: 3276
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: strange behavior of particles in a glass or cup

Postby ChristianZ » Fri Sep 02, 2016 10:45 pm

Ok, so I changed some parameter like adding friction again on the amphora and making the object smaller. This time, the outcome was much more normal. Can you confirm this on your system? Simulation parameters are as follows:
FPS=30, liquid PBD solver, Resolution=1, min substeps=2, max=4, iterations min=3, max20, accuracy 0.5
for RF2015 standard
https://www.dropbox.com/s/hv9f1b4xwjw1o ... a.zip?dl=0
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: strange behavior of particles in a glass or cup

Postby duke3d » Sat Sep 03, 2016 1:38 am

Yeah It looks better. However, on frames 60-68 it's obvious that liquid is divided into 2 layers near the walls of a glass.
RF_problem 07.jpg
RF_problem 07.jpg (192.12 KiB) Viewed 1281 times

Around frames 127 and 145 you can still see that weird effect clearly :( Actually, it reminds me Marangoni Effect :lol: , which occurs due to the differing surface tensions between two liquids. Needles to say that this effect should not be presented in our simulations.

In general, it's still pretty much the same, the difference is that the thin film on glass's surface doesn't stick to the walls and goes down faster.

By the way increasing "Deamon force scale for fluids" 10 times, and decreasing Drag Force 10 or 100 times (to make act like a water in small glass), leads to more pronounced effect :(

Today I will play with it a bit more

Thanks for your concern ;)
Andrew

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: strange behavior of particles in a glass or cup

Postby duke3d » Sat Sep 03, 2016 2:16 am

I've just noticed that "DY Roughness" helps to reduce a strange effect a bit.

User avatar
ChristianZ
Posts: 3276
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: strange behavior of particles in a glass or cup

Postby ChristianZ » Sat Sep 03, 2016 4:12 pm

By the way, this effect also occurs in "Realflow for C4D", the plugin for C4D that also uses the Dyverso system.

Although this sticking of particles to the wall is mostly pronounced during the activity of the emitter, when the particles are pushed up, there remains a band of sticking particles after the settling of the liquid.
Image
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: strange behavior of particles in a glass or cup

Postby duke3d » Mon Sep 05, 2016 1:53 am

Ok, here is what I found. :D

I made a new simulation with 2 Vases. I changed "DY Roughness" for the first(left) Vase to 1.0, and for the second (right) 0.0.
The left liquid looked good, right one was not good :))
RF_issue_07.jpg
RF_issue_07.jpg (193.39 KiB) Viewed 1211 times


Then I jumped to my previous simulation with 3 Vases and changed "DY Roughness" of the left Vase to 1. It looked pretty good, but still there were sticky particles on the wall, not much, but noticeable.
So I thought what is the difference between 2 projects. The size of the Vases. I jumped to my new project and played with different scales. After I changed "Daemon force scale for fluids" to 10, liquid in right Vase looked good.
RF_issue_08.jpg
RF_issue_08.jpg (185.38 KiB) Viewed 1211 times



Then I changed "Daemon force scale for fluids" back to 1.0, "DY Roughness" of both vases to 0.0 to eliminate the influence of this parameter at all.
Also I changed "DY Interact distance" of to 1.0 for the left Vase and 0.0 for the right Vase.
RF_issue_09.jpg
RF_issue_09.jpg (181.34 KiB) Viewed 1211 times


So, left Vase looks great :) Moreover, the simulation was more stable :)

To confirm that it works I opened my previous project with 3 Vases again and changed nothing but "DY Interact distance" of right green Vase to 1.0. It worked! :D
RF_issue_10.jpg
RF_issue_10.jpg (165.11 KiB) Viewed 1211 times


Later I will try few more simulations just in case :)


Andrew


Return to “Realflow bug report”

Who is online

Users browsing this forum: No registered users and 1 guest