[averis007]

So I'm simulating flow over a simple lambda wing. Initially I'm running a base case with a spherical farfield (Actually hemispherical, I'm simulating a symmetrical half-body). The conditions are Mach 0.5 and Angle of Attack between 8-14 degrees. My BCs for this problem are as follows:

1. Wing - HeatFlux (No slip wall)
2. Farfiled - Farfield
3. Symmetry Plane - Symmetry Wall

These cases are running as expected, producing coefficients (cl, cd, cm) which are close to the expected values and on post-processing in Tecplot, showing vortex cores as expected. Looking at the volume data by taking spanwise slices, we can clearly see the wake of the wing behind the trailing edge clearly.
https://ibb.co/7tS2DkH (LE Vortices developed on upper surface)
https://ibb.co/cbbqXyR (Wake visible behind the wing)


So far so good.


After establishing this baseline calculations, next what I wanted to do is to have an additional boundary condition over the wing. A small patch on the wing was saved with a separate marker and given the mass flow inlet boundary condition. If it helps to understand the problem better, imagine a wing flying through air at 0.5 Mach, and someone is blowing a jet from the wing with an air-blower. This is when the problem started.
First, the coefficients went for a toss. Drag became almost 12 times of what was expected!!!!
Next, analyzing the volume data showed there are absolutely no upper surface vortices (or any vortices for that matter). This is not in agreement with any kind of published experimental results.
Finally, when I analyzed the volume data by taking spanwise streamlines, the mach number in the entire domain was same, my input mach number of 0.5. No wake of aircraft, no suction on upper surface, no stagnation below the leading edge (depending on AoA), nothing.


https://ibb.co/2ZpQt7n (LE Vortices not developed on upper surface)
https://ibb.co/dPkdxHy (No Wake visible behind the wing)

My question is, what seems to be the problem in the second case? Are we not allowed to use farfield and mass flow inlet BCs together? Because even after 50000 iterations, there seems to be absolutely no flow developed around the wing. It is as if there's no wing inside the domain and everywhere the flow properties are the same as my farfield.


All problems are run on RANS solver.

[atelcikti1]

As far as I know and experienced in 6.2.0 version, both farfield and mass-flow inlet applicable together, I solved the simple nozzle case without problem.



It seems your case is not converging, in other words solution is not proceeding. If your initialization by mass flow then it is expected, you can run at higher CFL to see some process. Hope you can solve your problem.

[averis007]

Originally Posted by atelcikti1

As far as I know and experienced in 6.2.0 version, both farfield and mass-flow inlet applicable together, I solved the simple nozzle case without problem.



It seems your case is not converging, in other words solution is not proceeding. If your initialization by mass flow then it is expected, you can run at higher CFL to see some process. Hope you can solve your problem.

Thanks for the reply. If I'm not wrong, my case was run for more than 20k iterations. Ofcourse I had to give a low CFL value as the case was blowing up otherwise, but I'll try again.


I even tried changing from spherical farfield to a cuboidal domain, with the left face being an inlet and right face being an outlet, the other four faces being euler wall. The problem is in this case, the flow is highly dominated by freestream values, the inlet doesn't seem to work at all. I'm unable to understand what I'm doing wrong.


If it's not an issue, may I look at your cfg file for the nozzle case you mentioned. I only need to look at the Freestream and BC Definition blocks.