[sushil1151]

I am simulating flow over blunt body with aerodisk(Compressible flow: Mach -2). I've selected k-w sst turbulence model .I am getting area average weighted wall y+ as 27. Is this value acceptable? I've read from different sources that the y+ value should be less than or equal to 1.

[LoGaL]

If you read the sources why ask? You need to add a few more layers near the surface of your body

[xisluke]

Originally Posted by sushil1151

I am simulating flow over blunt body with aerodisk(Compressible flow: Mach -2). I've selected k-w sst turbulence model .I am getting area average weighted wall y+ as 27. Is this value acceptable? I've read from different sources that the y+ value should be less than or equal to 1.

It depends on the level of accuracy you are searching for. I mean, original k-omega SST has been developed for Low Re modeling approaches, subsequently proper wall functions especially for omega has been added to make it usable also for higher y+ values; I am not sure this happens in all CFD packages, Fluent adopts this approach. If you have enough computational power you can try to increase the number of inflation layers by refining down to low y+ values (remember that in order to solve correctly the boundary layer you need A LOT of cells, having only the first cell adjacent to the wall at y+ = 1 means nothing). If not, you can increase a little bit the dimensions going for y+ > 30 trying to check differences between k-omega SST and another model like k-epsilon Realizable, made generally for High Re depending on the wall functions adopted.

[shereez234]

Originally Posted by LoGaL

If you read the sources why ask? You need to add a few more layers near the surface of your body

Of course he is asking because he is in doubt . . . I am sure you have had some doubts along your career as well.. He read some stuff and now he needs to verify that information

[shereez234]

Originally Posted by xisluke

It depends on the level of accuracy you are searching for. I mean, original k-omega SST has been developed for Low Re modeling approaches, subsequently proper wall functions especially for omega has been added to make it usable also for higher y+ values; I am not sure this happens in all CFD packages, Fluent adopts this approach. If you have enough computational power you can try to increase the number of inflation layers by refining down to low y+ values (remember that in order to solve correctly the boundary layer you need A LOT of cells, having only the first cell adjacent to the wall at y+ = 1 means nothing). If not, you can increase a little bit the dimensions going for y+ > 30 trying to check differences between k-omega SST and another model like k-epsilon Realizable, made generally for High Re depending on the wall functions adopted.

Just FYI a more through approach is to have the Y+ equal to 2 or 3 for coarser grids with 10-20 layers, then Y+ 1 for more refined grids with 20-30 layers and anything lower than Y+1 for even more refined grids with subsequently more layers..

[xisluke]

Originally Posted by shereez234

Just FYI a more through approach is to have the Y+ equal to 2 or 3 for coarser grids with 10-20 layers, then Y+ 1 for more refined grids with 20-30 layers and anything lower than Y+1 for even more refined grids with subsequently more layers..

Yes, your answer is for sure more complete. I limitated mine on the basis of a first general comment on the topic