[urosgrivc]

hello

I have a question regarding interface with faces which are not the same size or shape.

https://drive.google.com/file/d/0Bw9...theU1GSjg/view

I set up a solid-solid interface than i include a boundary source (flux or total source) let say 10e9 [W], I set the nonoverlap conditions to adiabatic.
The question is on which side should i set up the source?
And why doesent it work on any of the sides, I also tried 1/2 on each as the materials are curently the same?

Im runing a transient rotor stator interface with pitch change set to (none) and results are there, just wrong.
The nonoverlap area results are right though.
I set up two monitors which monitor how much work got into both of domains. =>(volume*density*heat cappacity*delta-t)=[J]
And monitors should be the same as Power[W]*time[s]=[J]
The disc domain is rotating and pad is stationary.
Geometry of the disc has slots that is why i canot use (solid motion) with fixed power input area.
figure in the link shows what my problem is.

-I woul like to set the heat input into the disk to be set by the location of the "brake pad" but geometry makes it hard for me.

https://drive.google.com/file/d/0Bw9...ZIT1JtX2c/view (temperatures)

Unphisical results:https://drive.google.com/file/d/0Bw9...lHMTRBdEU/view
Wherever pad is touching the disk temperatures arent shown on the disc side but are on pad side.

Simulation is taking place in ANSYS CFX 16.2
I want to implement this onto much more dificult problems (including HTC contour from CFD res.) but i need to learn how to walk before i can run.
Thank you for the help

[ghorrocks]

You will need a much finer mesh than that before the results start looking realistic. Also inflation layers are probably a good idea, as there is likely to be a thin region of high thermal gradient at the boundary.

[urosgrivc]

Thank you

Well I have refined the mesh. https://drive.google.com/file/d/0Bw9...NmS0FZU3c/view

And results are looking much nicer now, but I have a problem with seting it up correctly.

one of the problems.
https://drive.google.com/file/d/0Bw9...dDYTRlVUE/view

let's say I would like to determine how much power goes in to one domain and how much in to the ohther two diferent materials in contact (distribution of power inputs)?

[ghorrocks]

Are you trying to model heat due to the pad friction on the disc? Is that why you have a heat source on the interface?

If this is the case then I can see why the temperature will go crazy on the non-overlap regions. I can't see a CEL variable to tell you whether you are in the GGI overlap region or not - but that is what you need.

[urosgrivc]

That is exactly what i want to model.

At first I wanted to model distibution of power betwen two bodies of different marerials in contact (that will be a project for the futrure)
to determine how much heat goes to the pad and how much to the disc as materials in contact are wery diferent, and than include aditional slots in the surfaces ,
Now I could simplify the interface, with calculating the total power input into the disc by hand, and include only the disc side, but the interface is stil problematic.

Overlaping area seems to be correct
https://drive.google.com/file/d/0Bw9...dLTU1TMkU/view

i think that (total source) is not taking SLOTS (area) into account so the Work delivered to the disc is smaller than it should be acording to monitors.
Wall heat flux does not get bigger when slot goes threw the interface and reduces its contact area.

And I cant find any detailed information about Sources in interfaces. help does not offer enough.
It is hard to uderstand (source) on one side of the interface and how does this difer from the 2nd side source (as you can put sources on both sides of the interface )
, results arent showing useful data and (total source) doesent seem to take the (changing) area into acount.

[urosgrivc]

I have made some progres:
I we tried to set the material of PAD to be extremly conductive and had no mass and heat capacity.
Now i could observe what was going on.

Graph shows (average_wall heat flux_PAD side of the interface)
https://drive.google.com/file/d/0Bw9...BDSXV3RDQ/view

Something is going on that I can not understand.
I think that when contact area decreses (becouse of slot) heat flux should increse but it does exactly INVERSE of that.

And the power input to the disc is NOT constant, it drops when slot goes by. How can I fix that
And total source is acting the same as flux which is also droping.

[ghorrocks]

Try specifying the heat flux as a W/m2 rather than a total heat flux.

[monkey1]

Just a remark..might be wrong, but:
When I try to imagine what happens in reality:
I have a defined (constant force) Pressing the two brake parts together.
So I will get a resulting total heat flux and power input correlated with the contact surface.
Now i decrease the contact surface (because of your slots) and without increasing the force pressing the parts together, I will result in lower total heat fluxes and power inputs.
Meaning: due to the force you're pressing both parts together you will generate W/m^2 and when the m^2 is smaller the resulting value should also be smaller.
Your Slots are like "cooling channels" reducing the heat and transmissible power, so that your results seem totally plausible to me.

Or am I totally wrong with what I'm imagining?

[urosgrivc]

Well thank you for a reply

But i think that isn't correct
When slot goes by, the area decreses but the force is constant so Surface pressure must increese. And if you are imagining a disk spinning around with constatnt pad FORCE and constant coefficient of friction TORQUE should be a constant and if this is hapining on a downhill slope at constant velocity, POWER input should also be a constatn as Power=Torque*angular velocity.

I think that power is a constant and if it is, than FLUX should wary with area.
When slot goes by the contact area decreses and flux shoul increse acordingly.
But i am still confused about this problem. i dont think that i set up the interface correctly. i think it is an interesting problem

Thank you for your thoughts

[urosgrivc]

Is there an expresion that would tell me the area of nonowerlap at the interface?
or the owerlap area of interface

As I can not point an expression to the interface area just to each side independently (not usefull in this case), and there is no (owerlaping area) in the wariables list