T106c Low-Pressure Turbine Cascade

Cases
1

Hello, I’m replicating the T106c Low-Pressure Turbine Cascade case in the paper ‘Impact of Number Representation for High-Order Implicit Large-Eddy Simulations’. The mesh I used is in https://how4.cenaero.be/, the order of mesh is third. But the error comes out: NaNs. Then, I used the anti-alias, the error is AttributeError: ‘NavierStokesElements’ object has no attribute ‘_grad_upts’. How can I solve it? Thanks a lot.

The ini file and mesh are attached.
https://drive.google.com/drive/folders/1SKZb6hfqh0wLHkcVWptz2KshrginuwXN?usp=sharing

With best regards

2

This could be a bug. Let me get back to you in a couple of days.

Regards, Freddie.

3

Thanks for your reply. I simulated this case with anti-alias by pyfr-1.15, the error ‘NaNs’ also comes out. Could you help me check my ini file or mesh?

Thanks a lot.

4

A large part of running these cases is getting the mesh, start-up procedure, and boundary conditions correct. I would therefore suggest you work to get the simulation stable at p = 1 and go from there.

Regards, Freddie.

5

Thaks a lot. I have start at p = 1, but the simulation also breaks up. The span direction is periodic, and the up and bottom of the computational field are also periodic. The boundary condition and initial condition are presented in ini file, as the https://acdl.mit.edu/HOW5/WorkshopPresentations/CS2_T106_LPT_Cascades/CS2_0_T106_LPT_summary.pdf said. I don’t know which part is wrong.

6

You will first want to check that the boundary conditions are being enforced correctly. Next, you will want to look at where exactly in the domain it is starting to diverge. Then, in response to this you may want to either increase mesh resolution in this area or adapt your start-up procedure.

Regards, Freddie.

7

How to know where exactly in the domain it is starting to diverge?

By the way, what’s the aspect ratio in your mesh? Is there a specific numerical limit to the aspect ratio of the grid in the FR simulation?

8

You will need to look at the solution as it develops and see if there are any regions of low/high pressure and or velocity.

With regards to the aspect ratio this is something you can determine yourself based off of the information given in the paper. Generally, for explicit FR an aspect ratio < 20 is suggested.

Regards, Freddie.

9

The isentropic Mach number is along the chordwise distribution. How to get the isentropic Mach numberalong the chordwise distribution? I know how to get the isentropic Mach number along x direction, but in the chordwise distribution, could tell me how to get it?