Follow up of cascade simulation project
 

Ive been very bussy lately doing my thesis...

Ive posted on this topic http://forums.procooling.com/vbb/showthread.php?t=8507 before

As a summary it goes something like this

My thesis is based on CFD (computational fluid dynamics).

In this project i will be simulating how a waterblock reacts with different variables..

In my case ive chosen a waterblock simmilar to Stew Foster's Cascade(jet impingement)...

Other than that i will be also doing exprimental research on this wb....

Thats why ive made a lot of this blocks to be tested.....

here are some pics of the simulation results using CFX a new CFD software

pic of the fluid's streamline

from a different perspective

The temps of the proc...

see how hot it gets when usng a proc from the future.....

250 watter proc .........Im feeling hot in here

lets see how hot the the water in the dimples get......

Very cool.

Nice work.

Cheers

Incoherent

Looks very nice.

What's the water velocity coming out of a jet?

Have you tried it with the jets submersed into the cups slightly?

I do like how the temperature maps show that the temperatures are very consistent across the heated surfaces (except at the very edges as expected), even if the color selection does make it look a lot more varied than it is.

In the pic, velocity is set at 1m/s..... ill be tring a few other velocities

As for the protruding jets,

havent tried it yet.......but obviously this has some effects on the performance of the block...


The jets from the centrplate is effected by the crossflow of the spent jet which you can see from the 2nd pic i posted.

The jets at the sides of the block does not impinge directly into the middle of the baseplate dimples

Other than that..... the velocity of the impinging jets degrades as it reaches the bottom of the dimples

Ill try the protruding jets but the complexity makes it hard for me to analyze the reaction for each change in design variable.

Very interesting cfd results, with that software you could change around the jets depth in the cups. Or you could just by some trial and error adjust the depth of the cups and see the results quickly. As seeing the results of a shallower cup depth around the edges and how the heat acts when you change around things on the block.

As a thought, try jet velocities up around 3m/s which are easily achievable with most pumps. It should change the temperature-map picture somewhat.

yeah....
have tried a whole lot of variables.........

at the moment ....
the best size and depth for the baseplate dimples is 4 mm depth and 3 mm diameter for 1mm jet holes with 2 mm height...

the analysis is still in progress.....

will post when i get the optimum jet size and height

A simple question:
Am I understanding the contours correctly?
For example:
( 1) In the Velocity profile is a red contour of 1.087e+000 equivalent to 1.087E+00 which equals 1.087?
(2) In the Temperature profile is a red contour of 3.450e+002 equivalent to 3.450E+02 which equals 345.0 ?

Anyone
1.087e+000 =1.087 ?
3.450e+002 =345.0 ?
The pictures are pretty, but with my interpretation of the values for Velocity and Temperature, the Convection Coefficients used are far from reality.

Yes... Easy, no?

345... thats Kelvin, for the benefit of slow people like me ;)

Units are shown on the pics... No difficulties.

Am also slow.
Being "K" helps if ambient is 298.2K.
Would still be informative to see the Convection Coeffs.

In fact/fantasy the DeltaT(metal -water) of ~ 50c(K) is close to my guess for a 250w load, using* h~ 10kw/m*m*K over all the dashpots' surface areas.

* v~ 0.25m/s, d~ 1mm using Flomerics rather than Sieder-Tate(ish) calculations.

Edit
h ~ 23kw/m*m*k for the nozzle velocity (v) of 1m/s.
This would reduce my guess Delta T from 67.5c to 39.75c

I like the simulation. It makes me want to try out Cathar's idea of blocking the Cascade's jets that are outside of the Barton core. I can see how the block should work more efficiently that way, given that I am using a very powerful pump.

Thats problably some expencive pice of software..
Anyone knows if there is a "lite" version for us mortals?

Reeealy wanna try this...