[SrSaito501]

I'm trying to do a parametric study of a nozzle geometry, in which I keep everything the same except the curvature of the throat. From a separate matlab code in which I use the Bartz equation, theres a visible tendency in which, the more the Rc (curvature radius), the less the peak temperature on the nozzle inner wall.
The thing is, when I simulate this in ansys, the plots of the wall temperature look fairly good, but the tendency of more Rc = less Twmax doesn't apply. Here is comparisson between Bartz and the results Im getting in Ansys:

-----------------Tmax Bartz ---Tmax Ansys
Rc = 1.25mm----1418K---------1299K
Rc = 6.25mm----1391K---------1317K
Rc = 15mm------1375K---------1303K
Here is some info of my setup:

Mesh
I've put inflation layers, first layer height being 3e-7m, with 90 layers and a growth rate of 1.06, getting a Y+ of 0.7.

I've checked skewness<0.85, orthognal quality>0.2 and element quality>0.4 (except from the inflation layers, in which I read you have to have a high aspect ratio to capture the gradient normal to the wall), and everything seems to be fine.

Im using edge sizing, and the setting are the same for every Rc.

Setup
Im using RKE EWT with compressibilty effects and production limiter on (I've also tried k-w SST and the tendency still resides, so I just stick with RKE as the computional time is not as costly).
As for solution methods, Im using second order upwind in everything and have put HOTR of 0.25 on all flow varibles, activating High Speed Numerics as well.
Im using Hybrid Initialization + FMG + patch to the solid body of the nozzle so it starts at ambient temperature.

Im using a time step size of 0.0005s, but have also made a time step study trying different values and the tendency seems to persist.

The problem seems to be on the Rc=1.25mm, where the temperature is just way off compared to the Bartz model. Does anyone know what is going on?