[fligh-high]

Hello community,

I am an absolute beginner in CFD and want to ask for your help urgently, as I'm beginning to despair . I want to investigate the flow of different engine inlets at different angles of attack and different velocities. The aim is to make a statement about the total pressure drop, efficiency, performance and mass flow. The simulation is to be carried out via ANSYS CFX. So I have created a simulation volume with different resolution regions in which the engine is located. This consists of a velocity inlet and a pressure outlet. The rest of the simulation volume consist of a free slip wall and a symmetry plane. At the aero engine there is a boundary condition before the compressor (velocity outlet) and one at the nozzle (mass flow outlet). All othere faces here are declared as wall. The values for the boundary conditions I determine by one-dimensional calculations.The inside of the engine, i.e. the compressor, combustion chamber and turbine are ignored for simplification.


The problem now is that the total pressure drop is much too small compared to the expectation. Mostly it converges towards zero or sometimes it gets even negative. I tried a lot with changing the boundary conditions. I achieved the best results with the settings you can see in the Report file in the appendix. There you can also see all other settings. In the Meshing i used an Inflation layer with 10 layers and a first layer thickness of 0,5 mm to get a y+ of ~200. But I also tried some other variations without success.


What am I doing wrong or what could i do better?

I am very grateful for your help.


Best Regards.

Sebastian

[ghorrocks]

FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

If you want accurate results you need to do a sensitivity analysis on mesh size, time step size (if transient), convergence criteria and boundary proximity.

Also consider whether your turbulence conditions are correct, and whether you need a turbulent transition model. That would have quite and effect here I suspect.

Also you mesh quality is not too good. The key problem is the jump in mesh element size from your last inflation layer to the first tet volume mesh is too big. I would extend your inflation layers further until they can transition nicely to the volume mesh. (But note you volume mesh is almost certainly too coarse anyway).

[fligh-high]

Many thanks for your quick reply ghorrocks. I will try to fix those things.

[fligh-high]

I think the mesh qualitiy should be better now. I am using the student software, so I am limited to 512k nodes. But I think this should be enough to get a first qualitativ statement about the impact of diffrent angles of attack on the total pressure drop.



As far as I understand it, the SST model is the best turbulence model for my purposes. But honestly, I don't really understand the theory behind all the turbulence models.



For weeks I have tried several simulations with different settings, for example I changend the roughtness of the walls, I tried variations in the mesh/inflation Layer/y+value, in the Boundary conditions/Parameters, I tried the k-epsilon and the SST model and now I have tried to use your tipps. I am also monitoring the total pressure in the inlet and in the intake-outlet.

But unfortunately the best results I am getting look like in the picture in the attachment. The totalpressure drop always converges against zero. But in the theory the total pressure loss should be ~2% (in this case >1000 Pa)

[ghorrocks]

Your mesh looks significantly better quality now.

Have you done a sensitivity analysis on mesh size, time step size (if transient), convergence criteria and boundary proximity? In addition, if you are limited on the size simulation you can do the sensitivity analysis can also show whether the accuracy you want is possible in the size resource you have. With your node limit it is quite possible that reasonable results are not going to be possible, and a sensitivity analysis will prove it.