[agoverno]

Hello everyone,


I am numerically investigating a type of radial turbine rotor where the hub is only partially attached to the blades. I managed to insert this hub gap in NUMECA Fine Turbo by means of a ZR effect as a FNMB (see figure), and it works fine for the fluid part, but for the FSI I believe I cant have this type of FNMB connections. I wanted to get deformations of these blades in the trailing edge part where it is not attached. My question was if I can impose an hub gap without using FNMB so I can do the FSI. Any other suggestions/advice is also very much appreciated.

Thank you in advance.

[alban1]

Hello,
Indeed, the FNMB connection next to the deformation is not recommended as indicated in the documentation. Instead, you can insert fixed Z cst line at the beginning and at the end of the cavity in AutoGrid5 as indicated in the user manual. This will ensure a matching connection between the cavity and the main channel. However, imposing a Z cst line on the blade may deteriorate locally the mesh if your blade is highly staggered.
Best regards,
Alban

[agoverno]

Hello alban1,

thank you for your reply! Indeed that is the case, I am having trouble generating a good mesh. For my case, I was using H&I topology for the B2B which was fine, but imposing a Z cst line on the blade only works with the default one. Currently I am having trouble on the expansion ratio (around 5) and in the spanwise angle deviation (around 75º). Do you have tips to improve my mesh in this case?

[colinda1]

Hello Agoverno,

Giving advise on how to improve the mesh quality would be difficult based only on a meridional view. An expansion ratio of 5 may not be show-stopping. Depends where it is located. It would also be good to see where this spanwise angle deviation is located.

Best regards,
Colinda

[agoverno]

Hello colinda1,


thank you again. I managed to get a not so bad mesh, which grid quality report I put in attachment. Anyways, I noticed a difference in the IGG part, for the case where the ZR effect has a FNMB patch and for the case I applied the connection via Z constant lines. However, it makes some kind of sense that this happens. These images are also attached below. When I try to run a flow simulation, the following error occurs:



!WARNING: IN CNMBEI, INTERPOLATION FAILED FOR THE NMB BOUNDARY CONDITION IN THE EDGE 3 BETWEEN BLOCKS 1 1 !THE SOLVER IS STOPPED. ! A MATCHING CONNECTION IS ADVISED IF THE FACE IS DEGENERATED (SINGULAR LINE).


Indeed I have some NMB but I dont know how to get rid of them. Currently Im using a single block for the meridional effect, with one Z cst line at the extremities.
Do you know why this is happening? If necessary I can provide more details. Thank you.

[alban1]

Hello Agoverno,
It is difficult to see from your pictures. The first check consists in going in the ZR effect and use "Detect Channel Matching Connection" under Topology Control/Detection Tools. If the tool detects a matching connection, the edge of the block at the connection should be highlighted in green. If it is not the case, you can use the tool "Add Matching Z Cst line" in order to add automatically the Z Cst line needed. If you have already inserted a Z cst line, it will detect it and will not insert a new Z cst line except if it is not properly positioned for ensuring the matching connection. In such a case, you will need to delete the Z cst line that you previously inserted.
If this does not solve your issue, could you use View/Patch Viewer in the 3D view and display the NMB patches? For this, you can filter the patches according to their boundary condition type, select them and use "show solid" and "show grid" in order to display a grid on a shaded surface. This would help us in understanding where the NMB patches are located. Thank you.
Best regards,
Alban

[agoverno]

Hello alban1,


yes, I did all the steps to having a match connection. I added the Z cst lines, and I do have a matching connection now. I was having problems to have a good mesh initially because I have a highly staggered blade, but I managed to obtain the values displayed in the previous post, which are reasonable. I was getting now that error from the previous post due to the NMB I think. I've attached all the NMB that I am getting. But I dont know how to get rid of them. Thank you for your help in advance

[alban1]

Dear Agoverno,
These NMBs come from the high staggered topology. This topology makes some grid lines coming from the inlet/outlet go on the suction/pressure side. This leads to a different number of points on each side of the blade.Since the number of points is different, it is not possible to ensure a matching connection in the tip gap. You can play with the number of points at the inlet/outlet in order to have the same number of points on each side of the blade or deactivate the high staggered topology in the blade-to-blade mesh control. However, this may decrease the mesh quality and, normally, the solver should cope with this NMB connection.
Another possible reason for this error message is that you have used the automatic partitioning for running the computation in parallel. The automatic partitioning may have split the NMB and automatically generated a FNMB connection which is not compatible with the FSI module. If so, you can still run the computation in parallel but you should use the methode "Parallel computation" instead of "Parallel Partitioned Computation". This will make the computation run in parallel by distributing the different blocks and preserving the NMB connection.
Best regards,
Alban