## Fluids

#### Nusselt number calculation and contour colors in the channel

• turanyank
Subscriber

Hi,

I am looking for corect expression for Nusselt Number between ribs as below.Black squares are ribs and flow comes from left side.

Then I need normalized Nusselt number contours, how can I create like below? I have constant heat flux at the wall. As we know Nu is dimensionless so how can I define it to get colors. Also still confused about local and global values at CFD-post.

Nu0=0.023 Re^0.8 Pr^0.4

Thank you

Turan

• Rob
Ansys Employee

Have a look at Custom Field Functions.  Expressions are replacing CFFs but I can't remember if they're available for post processing in the latest version(s).

• turanyank
Subscriber

Thank you Rob,

I finally find a same tendency like fig but errors are between 10-70%,

How I calculated is:

Defined point on heat flux surface, which is bottom of the channel then I wrote maxVal(Temperature)@pointt

for Tbulk, I made a section at that point perpendicular to flow and wrote areaAve(Temperature)@areaa

I take thermal conductivity with this Tbulk, used 500W/m^2 which is not mentioned in the article.(But I tried for different Reynolds number for this heat flux and got 10% error as compared to experimental study, so it looks 500W/m^2 is fine for entire surfaces between ribs)

Then I calculated

Nu=(500*hydraulic dia)/(k*(Tw-Tbulk))

Normalized it Nu/Nu0

Error becomes huge but tendencey look similar, it looks I have some formulation errror while writing the expression for Tw and Tb

Here is my graph and I dont know what am I missing.

Thank you,

Turan

• Rob
Ansys Employee

As a guess try adjusting Tbulk with position. With any comparison you need to be very careful, and many papers are deliberately confusing or omit something to make it harder to reverse engineer the work.

• turanyank
Subscriber

Thank you Rob, I rearranged Tb using interpolation through channel and max error reduced to 50% but this is still not enough.

Yes, you are right and it is really hard to find proper paper.

• Rob
Ansys Employee

If you've got the trend, and are fairly confident that your model is set up correctly review the correlations to see what could be responsible for the difference. Material properties are one possibility.

Plotting the two lines together may help.