Hello,

I post an new animation with simulation settings very close to the one used
for my previous post.

I worked on the mass loss. I changed a lot of things but the pressure
calculation on the free surface and at the water / object contact were the
main issues.

The rendering is more simple than in my previous post (no photon and
radiosity).

The simulation time was around 12 hours and the rendering around 12 hours. I
use a P4 3.2 GHz with 1 Go of RAM.

The video is MPEG1 encoded.

Any comments are wellcome.

Regards,
Fidos.

Post a reply to this message

Attachments:
Download 'coffeecup2.mpg' (671 KB)

fidos wrote:
> Hello,
> 
> I post an new animation with simulation settings very close to the one used
> for my previous post.
> 
> I worked on the mass loss. I changed a lot of things but the pressure
> calculation on the free surface and at the water / object contact were the
> main issues.
> 
> The rendering is more simple than in my previous post (no photon and
> radiosity).
> 
> The simulation time was around 12 hours and the rendering around 12 hours. I
> use a P4 3.2 GHz with 1 Go of RAM.
> 
> The video is MPEG1 encoded.
> 
> Any comments are wellcome.

I have to say that your liquid/water simulations have been spectacular 
to watch, great work!! My only complaint is that the liquid seems to act 
like a much more viscous fluid than watter, something more akin to very 
very soft gelatin. Like Jello that has just started to set in the 
fridge. The water surface seems to act too much in unison instead of 
there being a lot of inter fluid interaction, like ripples and waves.

Post a reply to this message

very nice. The water volume loss seems to be fixed to :-)

Post a reply to this message

> fridge. The water surface seems to act too much in unison instead of
> there being a lot of inter fluid interaction, like ripples and waves.

Of course the surface acts in unison. For the surface to act in ripples as
you suggest, the grid would have to be refined at least two or three times
over. To define a ripple the number of cells per ripple (wavelength) must
be at least 5 to 10 cells, higher for proper definition. I spent my thesis
on data analysis for improving the performance of a Go Kart and the same
principle applies there too.

I assume there are many more iterations in the code than the animation
suggests? I did a basic anim of heat flows across a plate and had several
subiterations in the code.

Fidos, I don't know if you've seen an Airzooka, its a novelty toy found in
shops in the uk. You might like to model the flow of air when the gun is
fired. A plastic diaphram connected to the base and powered by elastic
provides the pulse.
http://www.thinkgeek.com/geektoys/warfare/60b6/

James

Post a reply to this message