[Jonathan]

Hi everyone,

i hope all are well.

i was wondering whether anybody could assist me with how to make the following blocking periodic (i dont want to use non-conformal periodics).

for a whole lot of reasons, i cannot use the usual blade-in-the-middle type turbine flow volume, and so i need to use one which which has the suction side and pressure side of the blade as the outer flow volume edge.

the blades are (obviously ) not symmetric, so the portion of the flow volume which uses the blade surfaces is NOT periodic, however, the surfaces before and after the blade ARE perfectly periodic, as is the geometry in the tip-gap (periodic).

however, in ICEM, when i specify the periodic vertices, the premesh tries to make the mesh on the blades periodic as well, which is cannot be. If i remove the periodicity between the two vertices on the leading edge of the blade, i get the correct mesh on the blades, but then, i do not get periodicity in the first two periodic surfaces in front of the blade.

i have attached a picture, which shows the 2nd pair of vertices with periodicity removed.

i have seen some other posts where this type of flow volume is use (admittedly only in 2D), but i am sure that there must be a way to do this. it is, after all, not too uncommon to model turbine flow like this.

anyone have any idea? any help would be much appreciated.

best regards
cheers
jonathan

PS I have added some other pictures showing some perspective of the overall model, plus the periodic faces shaded in red, as well as a shot of the full 3D blocking i have used. many thanks, jon

[Jonathan]

Hi - anyone have any ideas? surely someone has seen this before

I found this blocking strategy from Psymn which seems to suggest this is would work, but i cant seem to replicate the o-grid's suggested ...

anyone / any icem experts have any idea?

may thanks and regards
jonathan

[kad]

Block and Edge selection for boundary layer grid:

[Jonathan]

oh my word, thanks kad, you are a legend. might sound ridiculous but i just couldn't figure this out, i feel a bit dumb. thanks a lot. i cant believe it is so simple ... !

thanks a lot, i really appreciate that.

btw, do you have have suggestions / idea as to whether this can be made periodic as per the post above?? i.e. if you look at Psymn's blocking i attached, he left out the LE vertex. if you set that pair of vertices to periodic, ICEM tries to make the blade faces periodic as well!

i am just about to try this so i will know in a few mins, but just incase you had any more experience / suggestions.

again thanks a lot,

best regards
jonathan

Originally Posted by kad

Block and Edge selection for boundary layer grid:

[kad]

Maybe you can try to maintain periodicity (in the highlighted area attachment) manually. Try to copy/edit mesh params, so that node distribution is identical on these two edges.

Maybe even linked bunching (under edge params) works for this. But originally it is meant to link the bunching of a series of smaller edges to one single edge. I do not know if it works for linking edges 1:1.

[Jonathan]

Hi kad, (+ anyone else who encouters this problem )

thanks again for your help / suggestions - i think i have worked it out though ...

its not explicit in the (ICEM) manual, however, if you read the section on periodic vertices, it says (effectively):

periodic faces are whose 4 corner vertices are periodic transforms (i.e. set as periodic vertices) of one another ...

therefore, the way to get ICEM to do it is to trick it by making an extra split in the blocks between the blades, and therefore, the 'blade' faces are not seen by ICEM as periodic anymore, and therefore, you can have different face geometries for them ...

i.e. i needed make the LE vert's periodic (and the TE) but then add the extra split shown on the picture (which i didnt have before) - neat little trick

all best regards
jonathan

Originally Posted by kad

Maybe you can try to maintain periodicity (in the highlighted area attachment) manually. Try to copy/edit mesh params, so that node distribution is identical on these two edges.

Maybe even linked bunching (under edge params) works for this. But originally it is meant to link the bunching of a series of smaller edges to one single edge. I do not know if it works for linking edges 1:1.

[Far]

Nice way of dealing with periodic vertices.

[Jonathan]

thanks,

hopefully it helps someone out, was a bit of a pain to figure out

cheers
jonathan

Originally Posted by Far

Nice way of dealing with periodic vertices.