Advice for fluid simulation

Become a member of the CGSociety

Connect, Share, and Learn with our Large Growing CG Art Community. It's Free!

THREAD CLOSED
 
Thread Tools Search this Thread Display Modes
Old 01 January 2013   #16
Hi Aikiman,

I've actually decided to go with maya fluids, not because I wouldn't much rather use real flow, - the client wouldn't be willing to pay an extra for it.

I also didn't use particles because I couldn't visualise how that would work, although I'm sure it would.

I managed to find an intro to maya fluids on simplymaya.com which helped a bit. There is a also a section in the maya help which tells you what properties to use to simulate water, these helped a lot.

I'm getting some acceptable results which I will hopefully post on here later. As you say it's not ideal but I think it'll do for the level of budget they are willing to pay.

Now I just need to work out how to create a decent water shader for it. I'm hoping that reflections on the water will disguise the form a bit when it starts to look questionable towards the end of the simulation.

I'm not sure if you can 'bank' a section of the sim and then continue the rest of the sim with different parameters like you can with cloth.

Thanks again for everyones advice.

Spencer
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #17
Okay, look forward to some shots/examples of your progress. I do have one idea that may or may not work and that is to use the ocean texture to displace the final polymesh. I've used SOuP in the past to extract ocean alpha and pipe into a deformer so that you can see it update in retime, it might just add that extra detail if used correctly. Maybe thats something you can think about after you've done sim'ing.
__________________
www.jeremyraven.co.nz
Vimeo
 
Old 01 January 2013   #18
Little progress animation

Okay thanks for the tip. I'm not very fluent with maya and haven't used it for a couple of years but I'll try to use your suggestion when I look at the shading tomorrow.

I've only had time to sim the first 500 frames but here it is:

http://youtu.be/dhzqeeCyoUs

Not perfect by any stretch but fortunately (hopefully) I should be able to view this from different angles during the sim so i can avoid those globules that seem to have a life of their own.
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #19
Thats looking pretty good so far, Ive noticed in some of the example files from visor have motion blur on Im sure this helps hide some of the artifacts and be more convincing. Also as mentioned before you can use a SOuP peak deformer to 'suck' the mesh in a little and get rid of those single blobs, and make the water look more like water although what you have there is pretty sharp, nice job.
__________________
www.jeremyraven.co.nz
Vimeo
 
Old 01 January 2013   #20
Progress update

Here I made some more changes, I had to half the base res from 600 to 300 so it would sim over night. I also changed the Fluid emitter > density method to 'replace'.

I upscaled the scene 10 fold.

http://youtu.be/oZekL5fBosc

Still shooting in the dark here a bit as the tests take so long.

This sim took nearly 8 hours.

Thanks again for your advice Aikiman although it doesn't seem to need it anymore, it almost needs the opposite. Maybe if I upres it back to 600 the water will separate a bit more....



Thanks

Spencer
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #21
I think it needs more volume and the turbulence in the middle seems a bit odd to me but maybe thats the effect you are after.

I believe to increase the volume, increase the density pressure, you might need to lower the pressure threshold a fraction then re-sim at the 300 res. When you are happy chuck it back onto 600 for another overnighter.
__________________
www.jeremyraven.co.nz
Vimeo
 
Old 01 January 2013   #22
Ok great I'll try that thanks.
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #23
If you want an emitter that is underwater to emit fluid then instead of density replace you may wish to use add. Normally it can be useful to have the density emission at 1.0 with method = replace and rate high with no dropoff. This makes it so the density is always 1 in the emission region( always water that is not pressurized). However if the density value is higher than 1.0 then it might suddenly explode outward... your simulation looks like you have values where the density is suddenly much greater than 1(or greater than the density pressure threshold) in the emission region and thus explodes outward. Also for a very high emission rate from a small region you may need to increase the substeps. Note that with an omni emitter the radius of emission is determined by the max distance attribute. Volume emitter types are useful if you want to control the exact emission region.

To emit at a fixed rate, like from a hose the add method may be better.

Also the fluid container boundaries work a bit better than mesh boundaries, so something that can help is to enable the -y, or bottom boundary on the fluid and turn off resizeClosedBoundaries, then make the bottom of the fluid a hair higher than the bottom of the collision mesh.

Last edited by Duncan : 01 January 2013 at 11:15 PM.
 
Old 01 January 2013   #24
Here is a scene that roughly matches your example that has more stable settings. For a fast filling arena at normal time scale the water will be very turbulent. If one instead wanted to simulate a slow fill, but in sped up motion(like time-lapse) then one could increase the substeps and simulationRateScale(also lower emission rate). It will be slower to compute, but it requires more computation. A time-lapse style simulation like that would fill more evenly with a flatter surface, but would also have very fast moving waves in much the same way as a real time lapse.
Attached Files
File Type: zip arenaFill.zip (9.8 KB, 40 views)

Last edited by Duncan : 01 January 2013 at 06:02 PM.
 
Old 01 January 2013   #25
Originally Posted by Duncan: If you want an emitter that is underwater to emit fluid then instead of density replace you may wish to use add. Normally it can be useful to have the density emission at 1.0 with method = replace and rate high with no dropoff. This makes it so the density is always 1 in the emission region( always water that is not pressurized). However if the density value is higher than 1.0 then it might suddenly explode outward... your simulation looks like you have values where the density is suddenly much greater than 1(or greater than the density pressure threshold) in the emission region and thus explodes outward. Also for a very high emission rate from a small region you may need to increase the substeps. Note that with an omni emitter the radius of emission is determined by the max distance attribute. Volume emitter types are useful if you want to control the exact emission region.


Well spotted, I thought it looked like a pressure bubble but had no idea how to remedy it.
__________________
www.jeremyraven.co.nz
Vimeo
 
Old 01 January 2013   #26
Hi Duncan,

Thank you so much for your help and I really appreciate you setting up a sample scene for me.

Sorry I've been slow in responding, not meaning to be rude, I have been working 16hour days for the past week and haven't had any spare time.

Just to fill you in on my progress when I turned up on monday the client said they were happy with the first sim that I posted, the turbulent one. As I had only sim'd 500 frames and could not work out how to import the existing sim to another workstation I had to resim it. It's still going now! - after a failure to pause and restart it on tuesday.

I'm at a point where I really need to get into the file and start doing test renders with the current sim.

Can anybody tell me if I should be able to open the file that is siming from another workstation and playblast the current sim that has been created without disturbing the workstation and file that is mid-sim?

I keep getting all sorts of errors about the cache when I try to play about with it.

Is it possible to open the file from another workstation and delete the cache and then re-attach it, or will this actually delete the sim frames from the sim folder?

I will definitely be loading up your file this afternoon too Duncan, thanks again for that.

Spencer
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #27
I have to say Duncan that on applying your fluid shape to my deflection objects (the stadium) it looks absolutely great and is sim'ing much faster. I've up'ed the mesh res from 6 to 12 (with no clear understanding of how this compares to my 600 resolution version!) and also animated the rate almost exponentially increasing from 4 to 400 over 1000 frames.

I am looking for a turbulent look as opposed to a slow fill that you mentioned.

I think it will take about 8 hours to simulate and will post the results up.

If you have any other advice considering my need to up the rate dramatically as time goes on - and if I need to adjust anything else in tandem then I'd appreciate it.

Thanks

Spencer
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #28
Simulation update

Here's shows the simulation using Colins settings (including in an earlier post). Thanks again Colin. I'm happy with this now and I believe the client will be too.

http://youtu.be/o3GKQTFfK54

Obviously this will look different with a water shader applied.

Does anybody think that if I drop a few thousand small particles on top of the water surface (using the water surface as a deflection object) and allow them to get thrown around by it that they might mimic small splashes and water particles?

Does anybody know I might be able to include foam into the shader as I would imagine there would be lots in this situation...I have no experience with the ocean shader but I'm hoping this might be able to produce the foam...

Thanks

Spencer
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Old 01 January 2013   #29
Yep, you could add a toon Outline to create a polymesh that will be positioned at the intersection of your 2 mesh objects - water and arena. You'd need to change some of the toon outline settings to achieve this but thats not hard, then convert to polymesh. From here you can emit from surface and hide the newly created polymesh. Now you have particle emission when the water intersects your arena mesh, make a collision between particles and water mesh.
__________________
www.jeremyraven.co.nz
Vimeo
 
Old 01 January 2013   #30
Thanks! I'll try that
__________________
Spencer Carpenter
Freelance 3D Artist
www.spencercarpenter.co.uk
spencer@spencercarpenter.co.uk
 
Thread Closed share thread



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
CGSociety
Society of Digital Artists
www.cgsociety.org

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump
Miscellaneous

All times are GMT. The time now is 03:16 PM.


Powered by vBulletin
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.