[CellZone]

Hi,

I have a box where air enters through a blower. Inside the box, I have a plate, where I want to analyze the convective heat flux.

It is a transient RANS simulation with a constant mass flow + temperature. I calculate 50 seconds with inner iteration = 30 ; time step size = 0.1s and I use the k-w-sst model.

Now what I get when I do a surface average report on this plate for the convective heat flux is a strong oscillating signal. So the heat flux varies from -2 W/m^2 to -14 W/m^2 (see picture).

The residuals look fine.

Can you help where: where do those fluctuations come from: is it just natural that I have some fluctuations?

So in real life, I would also see some fluctuations on the plate regarding heat flux. But does RANS capture those fluctuations?

Or are here some time step, convergence problems?

What do you think? Thank you for your help!!

Cellzone

[FMDenaro]

From a purely theoretical point of view, URANS adopts the ensemble averaging so that an external time-dependent force allows for a time-dependent solution.
On the other hand, one can argue that if the flow has no external driving force (that is statistically steady), the URANS solution can have a (numerical) transient but it has to converge towards the steady RANS.
But as the type of averaging is never really applied and only the turbulence model define the type of formulation, in practice there is a lot of confusion between the LES-like solution obtained by URANS.



In practice, your question has many reasons to be asked and, indeed, a couple of months ago I personally opened a topic on RG exactly about this topic and there are many interesting opinions:
https://www.researchgate.net/post/UR...ompared_to_LES