Hi, I am studying a case of internal air flow through a nozzle, and my solver doesn't predict correctly the recuperation temperature (temperature at the wall). I also made tests with Pr = 1, so I should obtain Tr=Ts (Stagnation temperature) but the prediction is very poor (Tr is 10 degrees higher Ts). Can someone suggests why everyone says "It is a difficult task to to obtain accurate temperature at the wall" and why CFD solvers find this difficulty ? thank you, Om.

(1). Near the wall, you have a boundary layer. (2). At high Reynolds number, the boundary layer will be relatively thin. Thus there is a very high flow gradient region next to the wall. (3). This requires very fine mesh next to the wall to resolve the high gradient and obtain accurate solution. (4). In addition, this can be complicated by the turbulence modeling when the boundary layer is turbulent. (5). Without knowing the flow conditions and the algorithm used in the solver, including your mesh size and arrangement, my suggestion is: refine the mesh near the wall until you obtain the mesh independent solution. Beyond that point, you will have to do your own research.

Dear Sir, my specific test-case is a low Reynolds number case (40000, based on the throat). So, I consider the laminar transonic shock-free (low downstream pressure) flow through the nozzle. My mesh is very fine (I already made grid independency investigation) and can clearly show the boundary layer distribution ... However, the recuperation temperature is overpredicted at the throat ! I wonder if this is not due to the artificial dissipation in my case. It is true that the boundary layer thickness is slightly larger than expected. Anyway, I've heard many times that solvers in general have difficulties to predict the recuperation temperature ... A reason could be the meshing, that cannot "capture" the high gradients. But what are the other possible reasons ?

(1). I would start from the governing equations first, and check the equation to see whether it has everything in it for the compressible laminar flow.

What about the Nozzle geometry at the throat? Are you sure you are reproducing it exactly in your model? At wall boundaries what boundary conditions are you using for the energy equation?

I see from the posts below that you've seemed to have eliminated the mesh and turbulence factors. The fact that your BL is overly thick (what are you comparing this to?) indicates that you may have excess dissipation (physical viscosity etc). Also what are you using for your wall thermal bc? Probably you are using adiabatic? Is that correct?