[tricha122]

Hi,

(note below concerns incompressible flows only)

I have been using a UDF to calculate the HTC at the wall for a steady solution fairly repeatably for a while now.

I now have a situation where i have a rotating mesh interfacing with a static mesh (both fluid) and so i now need to run unsteady. I noticed when i ran unsteady my HTCs seemed off by quite a bit from what i expected (about 4-5x off).

I then simplified the situation and ran just a simple cylinder pseudo-transient (ramping up flow rate, then holding until convergence) and unsteady (ramping up flow rate, and then holding until convergence).

The HTC i calculate on the pseudo-transient run is as expected. The HTC i calculate on the unsteady run is ~4-5x less than expected.

The velocity profiles (when i plot them on a line/rake) look the same.

the crux of the HTC calc comes down to the calculation of y*, which ultimately relies on the turbulent kinetic energy, kp = C_K(c0,t0)

Is there any reason why i would expect kp to be different in an unsteady solution vs. a pseudo-transient solution?

I have suspected that perhaps my solution wasnt fully converged so i ran it for many more iterations, ive switched to second order for momentum, k, epsilon, etc... no dice

Anyone have any idea on what to check to try and close this gap?

Alternatively i could run the unsteady solution for X iterations, then run a steady solution, and do this back and forth to get the information that i am looking for... is it possible to do this? Will the "steady" solution run in the instantaneous position of the rotating mesh after i iterations?

Any help would be greatly appreciated

Thanks!