Finding thermal resistance using average temperature

[afreeze]

Hey, I have a simple model set up right now that I'm using to verify my process is correct. This will later be extended to analyze finned heat exchangers. My model is a square block of a defined thickness. It has a convection applied to one side. The side of the block that is opposite has a heat flux source in an area 1/10 the size of the block face area. This heat source is centered. All other sides are insulated including the area surrounding the heat source.

The analysis runs fine and my issue is with postprocessing. I'm looking to do a process similar to COMSOL Multiphysics in which you can integrate over a boundary by selecting faces. This integration provides a temperature result in [m^2*K] and a normal heat flux in [W]. I use this integral and divide by the face area to get an average temperature. This is then used to back out the thermal resistance (R=deltaT/Q). The normal heat flux is used to check the energy balance between sides.

I've spoken with some regular users of ANSYS and they believe a similar process can be done in ANSYS but didn't know how. Can anyone help me with integration across a boundary to find an average temperature and back out the thermal resistance? Or is there another way built into ANSYS for finding thermal resistances?

 

[IRstuff]

I'm confused as to why you wouldn't be using the source-side temperature and area to determine the source-ambient thermal resistance, since a significant portion is the spreading resistance in your block.

TTFN

FAQ731-376

 

[afreeze]

The effects of spreading resistance is actually my focus. I am comparing the resistance that I find through this process to MATLAB code that I have written to verify its accuracy. The deltaT is the difference of the average of the source on the one side to the bulk temperature of the convective side. I am using the source side temperature and area for determining the thermal resistance. However, I'd like to use the average temperature of the source instead of the maximum temperature or guessing based on the scale in ANSYS. I would like to use a feature similar to the integral feature in COMSOL which would allow me to find the average temperature of the source. I just haven't been able to locate this feature in ANSYS nor can anyone else that I've talked to.

 

[IRstuff]

Why wouldn't you just compare the resultant temperature distributions? Seems to me that any integration would obscure whether the heat is spreading as expected.

TTFN

FAQ731-376

 

[afreeze]

I'm looking to verify my MATLAB program and not worried about the distribution. I'm looking for a numerical value for comparison. In the long run, the MATLAB code will be a small part of a much greater program that runs optimization routines. I need these to be quick so using ANSYS when optimizing fin geometries isn't favorable. At a later date, there will be experiments run in order to determine the heat spreading effects but right now, verification of the MATLAB code is of the most importance.

The equivalent procedure in COMSOL Multiphysics is
Postprocessing>>Boundary Integration>>Select Boundary(s)>>Temperature [m^2*K] or Normal Heat Flux [W]

I use this feature to take the integral of the source boundary and divide by the area of the heat source which provides me an average temperature. This average temperature is needed to determine the thermal resistance through the entire block.

Is there a similar feature in ANSYS? I've been told there is by colleagues but it is unknown to them where it is and I have been unable to locate it myself.

 

[afreeze]

The similar feature in ANSYS can be done by using Surface Operations in the Postprocessor. You must first create a surface using 'SUCR'. For my application, the surface was a flat surface created using the cutting plane and targeted at a specific volume which I selected separate from the total geometry. The working plane was changed to define the cutting plane needed. The results then need to be mapped to the selected surfaces using 'SUMAP'. After that, mapped results on the selected surfaces can be plotted using 'SUPL'. There are also different operations that can be performed on the mapped item using 'SUEVAL'. These include an integral, a summation, or an average across the selected surface. The result can be seen in the Command Prompt window or can be stored as a scalar parameter.

 

[gwolf2]

Dump the face results out to file, suck it into a spreadsheet and integrate if you can't get the package to do it.

 

[afreeze]

Thanks. The surface operations and integration in ANSYS worked fine. I was just posting how to use the commands in case anyone had similar questions in the future. Thanks again.