Learning Forum

[loxliker]

Hey guys!

Quick question. I want to investigate the temperature fields in the boundary layer due to friction. I am using the k-omega-SST-model and an adiabatic wall (heat flux = 0). The flow is coming with a velocity of 18 m/s over an airfoil with an angle of attack of 8 degrees. The temperature of the flow is 278,15K (5 °C).

My calculated value (by hand) for dT should be something like 0,16-0,2 K but if I run the simulation i will get for the adiabatic wall a dT of 2,2 K or something.

Here are my settings:

- Energy is on

- viscous heating is on

- heat flux = 0 (adiabatic wall)

- wall thickness 3mm ( 0,003m)

Can someone tell my why my results are so different than the analytical ones? I reduced the criteria for the residuals and increased the number of itereations to increase the quality of the solution but this didnt help. Can you help me here?

Thanks in advance!

[Amine Ben Hadj Ali]

What is the basis for your "analytical" solution? Is you fluid compressible? Can you show the convergence residual?
Add more details to understand the question.

[Rob]

Additional info here /forum/discussion/34790/k-omega-sst-setup#latest

[loxliker]

Hey!
The Basis for my analytical solution was: dT = v^2 / 2 Cp = 18^2 / 2*1005 = 0,1611K
If you go for the definition Ma bigger than 0,3 then my fluid isnt compressible at all, because my Mach number with a velocity for the incoming flow of 18 is pretty small. Anyways the thermal effect in the boundary layer due to friction are still there, also in incompressible fluids. they might be small but they're there. Thats my basis. I know that my temperature fields would be small.
to answer your previous questions: My solution didn't converge. I decreased the residuals to increase the quality and after about 15000 iterations it didnt converge.
so i put all my settings into default and started again. i activated all limiters and for the inlet the intesity to 0.1 and the lenght scale to 1. I also changed the angle of attack in the geometry from 8 to 6. furthermore i changed it from 6 to 7 degrees by giving the flow an angle 1 degrees by using the direction components. sin(1) and cos(1).
I give you some numbers: Angle of attack: 7┬░
Surface integrals: min: 278,1505K , max = 278.5284 K
this gave me a dT of 0,37K but why is the temperature at the stagnation point (expressed by the minimum) not 278,17K ?
because in theoy it should be more heated than that because all the kinetic energy of the flow dissappears and results to a heating.
MANY thanks in advance. I hope you guys can help me out, I'm literally freaking out.
PS: is the k-omega-sst model the best one for that kind of simulation? i also tried the realizable k-epsilon model but this didnt change anything.

[Rob]

How well resolved is the mesh?

[loxliker]

Around the airfoil i used the inflation method. I used a calculator for it, and then i dercreased my first layer height to reach the desired y+ value of 1. i think the first layer thickness is something about 0.0000133 m and then with a growth ratio of 1.2 and 28 layers, to reach the y+ of 1. Around the inflation its also small to avoid strong volume transition but not that much becuase of the computional power and time.

[Rob]

And the streamwise resolution? A good inflation mesh means the gradients perpendicular to the wall should be well resolved, but won't help much if you have changes in the streamwise direction.

[loxliker]

how can i check on this?

[Rob]

Plot a contour on the wall & turn off node values. Temperature or pressure are probably best. You're looking to see if the gradients are resolved.

[loxliker]

This is my contour plot, I defined my fluid (air) as a ideal gas and this gave me a temperature of 278,26K in the stagnation point (surface minimum) and a maximum value of 278,31K which brings me a dT of 0.05K

[Rob]

The cell quality from the inflation to free stream isn't ideal, but otherwise is OK. For 2d do not pick any surfaces to display on and you'll see a continuous contour rather than the underlying edges. Also, plot the contour on the wall and with node values off and local range (global off) compute the temperature range.

[loxliker]

[loxliker]

this is my geometrie, i used this calculator for it: https://www.fluidmechanics101.com/pages/tools.html
velocity is 18 m/s, and desity and dynamic viscosity should be selected for 5 ┬░C (278,15K)
can you help me out, how to generate a good mesh to see the optimum temperature gradients?
first I started with the triangles method and then with edge sizing on the circle and further the sizing in the circle , after that i initialized an inflation around the airfoil.
I am saying this all because my result isnt 100% plausible. at some points yes, but its not uniform.
Thanks, Rob!

[Rob]

You'll need to repost the image in with the text as I'm not permitted to open attachments. Re the meshing, have a look at the tutorials in the help system.

[loxliker]

this is my airfoil, can you help me out with some values and methods?

[Rob]

Inflation, pave and watch the growth rates.

[loxliker]

Hey Rob this is my final Mesh, i did the whole meshing process again. I give you some numbers and let me know if this is okay, or if there is something you would change. You can see my geometry above.
The whole mesh has an element size of 0.1m
triangles method without supression
element size on the edge of the circle -> 0.05m with flexible behaviour
element size on the surface of the circle -> 0.005 without supression (to decrease the volume transition, from the inflation to the freestream)
inflation on the airfoil: first layer thickness: 3.98e-5, layers: 25, growth rate: 1,2 to make sure that the boundary thickness (99%) is inside the inflation.
I am using the k omega SST model. Is there something to watch explicit, for example the y+ value. Which value you would recommend ? I read in many forums, that this type of model doesnt need a strict specification of y+ because it works with a wide spectrum of values.

[Rob]

For an accurate solution the general guidance is a y+ of around 1, but it's good over most sensible y+ values. Regarding remainder of the settings, they look OK but given I know very little about your model I can't say for sure. The best approach will be to run the model, save, adapt the domain and run on. Compare the results. If there is a difference do it again: that's a mesh independence study.

[loxliker]

Okay thanks!
Another question:
I made a contour for the static temperature. As you can see in the second image, the temperature in the stagnation point is 278,3112 K as expected (because the whole kinetic engery will be transfered into thermal energy and so on) this is plausible.
BUT on the third image, if i click on the airfoil (global range) to see the interaction with the boundary layer (compare this to image 1) why is that all red ?
I set this up for an adiabatic wall (heat flux = 0, wall thickness 3mm), why is the airfoil all red?
the adiabatic temperature is a function of the recovery factor but i can't see different colors.
do you unterstood my question? iam not sure how to explain it better and btw greetings from germany :)

[Rob]

Look at the temperature range on the contour panel.

[loxliker]

okay, I see but why is it all red? It should be something like the other airfoil you know, more colorful, to see some differnent regions. I can see a dT of 0,5K but why is it only red, this is what i don't get.
At the end i want to investigate the convective heat flux and this depends on the heat transfer coefficient and the adiabatic wall temperature (recovery factor).
These two parameters depend very strong of the flow conditon (laminar, turbulent and seperated)

[Rob]

What is the temperature of the surface, which band of the contour plot does it fall into?

[loxliker]

The temperature of the airfoil surface should be 278,15 K the same as the outer flow temperature. So we assume that the airfoil has nearly the same temperature and we want to investigate the heating of the airfoil depending on the heat dissipation due to friction in the boundary layer. The aifoil is an adiabatic wall, so there is no interaction.
I used static temperature to investigate the temperatures. Local range to see the heating of the airfoil and global range to see the heating due to heat dissipation.

[Rob]

Not quite. You've got a contour range set, and the surface is "red". If the contour range is changed do you see banding?