[attiquejavaid08]

Geometry consists of a hollow cylinder and two pipes normal to it.
Created blocking for all three cylinders independently (beginner level) by assigning them all different body points.
files are also attached for reference to what i did.

edit: please check google drive link for files in second post.

[attiquejavaid08]

I have managed to block hollow cylinder with it's connection to one pipe (marked in figure). i'm getting very low quality hexas at cylinder connection to pipe. i have tried a lot and still doing so to cope it up but still in vain. I want to create o-grid later for boundary layer capture.
blocking file is attached. if someone can help me out and suggest me what is needed to do then please. i'll be very thankful
https://drive.google.com/file/d/0B4Y...ew?usp=sharing

June 15, 2015
edit: updated geometry file with new blocking approach attached here and #9
https://drive.google.com/file/d/0B4Y...=sharing&pli=1

[shereez234]

Originally Posted by attiquejavaid08

I have managed to block hollow cylinder with it's connection to one pipe (marked in figure). i'm getting very low quality hexas at cylinder connection to pipe. i have tried a lot and still doing so to cope it up but still in vain. I want to create o-grid later for boundary layer capture.
blocking file is attached. if someone can help me out and suggest me what is needed to do then please. i'll be very thankful

I am not an expert in cylinder meshing However, I suggest you to try these things shown in the attached picture to this post.

prophot.jpg

Regards

Shereez

[shereez234]

Also, the last step I suggested is an O grid face adding which can improve the skewed surface of the connection between belt and inlet or outlet cylinder. However without this step ( O grid faces) you should still be able to achieve good quality without much trouble. If you add o grid faces there will be added vertices and edges thus you should manually move them to the right places ( even the internal edges since fluid is flowing through them in this project).

So yeah good luck and come back to discuss what else could be done.
Hope it helps you

[attiquejavaid08]

Originally Posted by shereez234

I am not an expert in cylinder meshing However, I suggest you to try these things shown in the attached picture to this post.

Attachment 40101

Regards

Shereez

hello shereez!
your tip of meging vertices improved quality. joining area is still a problem as it was, where highly skewed elements are.

idea of of adding ogrid-face for vertices like this, being a beginner, it was a good tip for me to consider in future.

for this geometry, joining block (middle block in figure 03) is already small enough as per my understanding that dividing it into smaller blocks and having vertices there, results in improvement in quality but not at joining section of large hollow cyl and small normal cyl.

check the images please. and revised blocking file attached as well.

[attiquejavaid08]

my objective and issue.

[shereez234]

Originally Posted by attiquejavaid08

my objective and issue.

Hi Attique. Sorry for the late reply. As I told you in PM i was away until today.

Anyway check this blocking I have revised for you just now? I have assosciated some curves and changed edge meshing parameters, moved some vertices around. Min quality is 0.2 and above at the connection now.

I improved only one connection and for the other one I did not do anything coz you know you can observe, figure out and re apply what I have done there for the first connection as geometry is similar.

Is this what you were looking for?

PICTUREREVISED2.jpg

and the blocking file from my drop box:
https://www.dropbox.com/s/4afzlxlicg...VISED.blk?dl=0

[attiquejavaid08]

Originally Posted by shereez234

Hi Attique. Sorry for the late reply. As I told you in PM i was away until today.

Anyway check this blocking I have revised for you just now? I have assosciated some curves and changed edge meshing parameters, moved some vertices around. Min quality is 0.2 and above at the connection now.

I improved only one connection and for the other one I did not do anything coz you know you can observe, figure out and re apply what I have done there for the first connection as geometry is similar.

Is this what you were looking for?

Attachment 40121

and the blocking file from my drop box:
https://www.dropbox.com/s/4afzlxlicg...VISED.blk?dl=0

really appreciate your effort dear. thank you so much. i'll come back as soon as i get my hands on personal computer to check it.

[attiquejavaid08]

Originally Posted by shereez234

Hi Attique. Sorry for the late reply. As I told you in PM i was away until today.

Anyway check this blocking I have revised for you just now? I have assosciated some curves and changed edge meshing parameters, moved some vertices around. Min quality is 0.2 and above at the connection now.

I improved only one connection and for the other one I did not do anything coz you know you can observe, figure out and re apply what I have done there for the first connection as geometry is similar.

Is this what you were looking for?

Attachment 40121

and the blocking file from my drop box:
https://www.dropbox.com/s/4afzlxlicg...VISED.blk?dl=0

hi Shereez! thank you so much for your inputs. i've just removed pipes to make geometry more simple. and at weekend, i did some blocking. both blocking approaches, mine and your suggested, i've incorporated in a single file now in following link: https://drive.google.com/file/d/0B4Y...ew?usp=sharing
kindly do check it please.
and pictorial description:

[shereez234]

Originally Posted by attiquejavaid08

hi Shereez! thank you so much for your inputs. i've just removed pipes to make geometry more simple. and at weekend, i did some blocking. both blocking approaches, mine and your suggested, i've incorporated in a single file now in following link: https://drive.google.com/file/d/0B4Y...ew?usp=sharing
kindly do check it please.
and pictorial description:

I will get back to you on this some time today or tomorrow. And good you are working hard. without effort nothing is possible and without mistakes learning is impossible too

[shereez234]

Seems to me like where you have created O grid inside the elements are of low quality due to very fine wall spacing there? 1e-06? Is this the cell initial height you are going for?

And you can try to do this :

edit edge - spline or automatic linear
bend the edges so they represent a circle. Even the O grid internal edges.

Furthermore some associations also seem to be wrong. One of the internal edges was projected to a surface. Review again.

[attiquejavaid08]

Originally Posted by shereez234

Seems to me like where you have created O grid inside the elements are of low quality due to very fine wall spacing there? 1e-06? Is this the cell initial height you are going for?

And you can try to do this :

edit edge - spline or automatic linear
bend the edges so they represent a circle. Even the O grid internal edges.

Furthermore some associations also seem to be wrong. One of the internal edges was projected to a surface. Review again.

yes shereez. i'm trying to achieve y+ ~ 5 for Re of 10^8 in my case. for y+ 30, it will be 8e-6 m.
i am going to review what you said. will check it shortly and later i'll post thre results.

[attiquejavaid08]

i want to ask one more question. should i consider "build topology" when i import my geometry from designmodeler? is it necessary?

[attiquejavaid08]

Hello shereez and everyone who is reading this post!

I’m very thankful for timely response of shereez which made me able to keep looking for issues, errors. (to be honest, at one point I was so frustrated and I just wanted to back off last night. Just took a little break from it.)

A little change in blocking and to make better associations as per my understanding, i added curves and points on horizontal plane as separate parts so that at one time using index control of blocking I can concentrate on a single curve and points on it. This geometry breaking really helped me a lot in figuring out the issues which were probably due to some wrong associations I made (pointed out by shereez as well).

updated geometry and blocking files before and after ogrid are attached as following:
https://drive.google.com/file/d/0B4Y...ew?usp=sharing

I’m happy I have not expertise in ICEM but having a helping hand through this forum and going through discussions, other’s work and doing practice hard this meshing tool is making sense to me now. There are still a lot of things to do.

AJ

[attiquejavaid08]

Originally Posted by shereez234

Seems to me like where you have created O grid inside the elements are of low quality due to very fine wall spacing there? 1e-06? Is this the cell initial height you are going for?

And you can try to do this :

edit edge - spline or automatic linear
bend the edges so they represent a circle. Even the O grid internal edges.

Furthermore some associations also seem to be wrong. One of the internal edges was projected to a surface. Review again.

Hello shereez and everyone who is reading this post!

I’m very thankful for timely response of shereez which made me able to keep looking for issues, errors. (to be honest, at one point I was so frustrated and I just wanted to back off last night. Just took a little break from it.)

A little change in blocking and to make better associations as per my understanding, i added curves and points on horizontal plane as separate parts so that at one time using index control of blocking I can concentrate on a single curve and points on it. This geometry breaking really helped me a lot in figuring out the issues which were probably due to some wrong associations I made (pointed out by shereez as well).

updated geometry and blocking files before and after ogrid are attached as following:
https://drive.google.com/file/d/0B4Y...ew?usp=sharing

I’m happy I have not expertise in ICEM but having a helping hand through this forum and going through discussions, other’s work and doing practice hard this meshing tool is making sense to me now. There are still a lot of things to do.

shereez, i'm not good at editing edges yet, but next i'm gonna give it try.

AJ

[shereez234]

Originally Posted by attiquejavaid08

Hello shereez and everyone who is reading this post!

shereez, i'm not good at editing edges yet, but next i'm gonna give it try.

AJ

AJ;

Hi. I looked at your updated project this morning. And the one with O grid blocking seems to be well done with only a few improvements I brought to it. I did not play with any wall spacing. However I increased the amount of nodes on the small curve of the big belt where problematic elements were. And I moved some nodes. Have a look at these pics. This is quality interms of unstructured hexa - not premesh quality remember. (EDIT: I have just seen that I was not able to capture the quality histogram into this photo - but it was 0.25 and above for ICEM quality)

qualiTY_MATRIX.jpg

Now I have also checked this improved mesh via fluent - and here is a screenshot of mesh quality and statistics. Min orthogonal quality is 0.17 I think.

fluent_PRO_PIC.jpg

pp.jpg

Updated flies can be downloaded from my dropbox - I will share the folder with you in PM.

Regards

Shereez