Mesh discontinuity

hippo407 · March 23, 2017, 06:49

Dear every cfx experts.
I have a question about different bodys to form an another part.
By my opinion, this is one way can make all mesh points merge together in the interface between each bodys. but my question is , in some complicate CAD case even a part contain many parts , it is still happen the nodes are not match together.

Is there any way to find and check which interface is not matching together or any way can make 100% the interface mesh be matched?

I know can set interface to use ggi match different mesh face by users . But I still curious about this situation whether have more convenient way to resolve .

ghorrocks · March 23, 2017, 17:23

Have a look at the overlap % in the output file and in CFD-Post in the results file.

hippo407 · March 23, 2017, 21:24

thank you for reply,Glenn

Can the overlap % you mentioned be used to check the interface inside a part which contain many bodys but mesh may not connect like hex with tetra and said bodys are full in the same domain?

ghorrocks · March 23, 2017, 21:38

The overlap % simply returns the % of the area of one side of the GGI which it has managed to find the matching face on the other side of the GGI. Whether the sides are in different domains or not, tet or hex mesh or anything like that does not matter.

hippo407 · March 23, 2017, 21:54

Actually, I may not describe very clear.

my case is a part contain many bodies (already form a new part in geometry) and only one body use subdomain to get energy source , and all bodies are in the same domain. When I finish the simulation , I find the temperature distribution is very strange and I guess the interface between the subdomain one and other nearby bodies many not correct connect .

So I operate an easier model to test , I find if the meshes of said all bodies inside are connected ,it is not necessary to set interface in one domain , default setting is fine ,but if mesh are discontinued since node not math or using differ mesh method on different body (hex with tet),the interface inside will not exchange the heat.

So my confusion is , Is there any way to check all mesh of a part contains many bodies really connected between every bodies inside a part ? or just to see by eyes and set interface even in the same doman and same part.

thanks!

ghorrocks · March 24, 2017, 00:45

Have a look at the non-overlap fraction in CFD-Post to see where your interface connects and where it does not.

hippo407 · March 24, 2017, 01:07

Yes , I got i . thank you Glenn for reply it , so in some complicate CAD, even bodies are in one part maybe need to set interface manually or just improving the mesh. I appriciate it!!

Otherwise, I see ANSYS HELP under the title "Non-overlap Boundary Conditions" mention that "Nonoverlap Fraction or (1-Nonoverlap Fraction) to ignore contributions from nodes on a GGI interface" I am confused about what is used for . like to set the heat flux turn to 0 ?

ghorrocks · March 24, 2017, 06:16

I do not know the reference you are talking about. What section of the documentation is that from?

hippo407 · March 26, 2017, 10:34

hey , my ref comes from here.
could you help me to have a look?

it located on the last sentence under 5.4.3.1

https://www.sharcnet.ca/Software/Ans.../CACBCJBF.html

ghorrocks · March 26, 2017, 20:44

If you read the full sentence it is quite clear to me:

Quote:

For example, you could multiply a quantity in an expression by Nonoverlap Fraction or (1-Nonoverlap Fraction) to ignore contributions from nodes on a GGI interface that are in the overlap portion, or non-overlap portion, respectively

hippo407 · March 26, 2017, 23:14

Is it used on the source term and make the flux crossing the interface can be zero or anything?

could you show me a simple example ? I am very appreciated

ghorrocks · March 27, 2017, 07:11

You simply put the Nonoverlap Fraction variable in the CEL expression to evaluate the source term. But before we talk about the details of this - what are you trying to?

hippo407 · March 27, 2017, 21:27

For my project , I am trying using auto generation for the interfaces, and I want to know whether this function is complete or not.By my test, I think it is powerful enough and only need to check the nonoverlap fraction as you said in the CFD post and find anywhere need to be set interface manually.

And the last question is just my curious problem which I saw at the help.

If there is any misunderstand concept ,plz tell me. thanks

March 23, 2017, 06:49	Mesh discontinuity	#1
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	Dear every cfx experts. I have a question about different bodys to form an another part. By my opinion, this is one way can make all mesh points merge together in the interface between each bodys. but my question is , in some complicate CAD case even a part contain many parts , it is still happen the nodes are not match together. Is there any way to find and check which interface is not matching together or any way can make 100% the interface mesh be matched? I know can set interface to use ggi match different mesh face by users . But I still curious about this situation whether have more convenient way to resolve .

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
sliding mesh problem in CFX	Saima	CFX	46	September 11, 2021 07:38
decomposePar problem: Cell 0contains face labels out of range	vaina74	OpenFOAM Pre-Processing	37	July 20, 2020 05:38
[snappyHexMesh] Multi Region Mesh of a car filter	Zephiro88	OpenFOAM Meshing & Mesh Conversion	3	September 11, 2019 19:34
[ICEM] Problem making structural mesh on a surface	froztbear	ANSYS Meshing & Geometry	1	November 10, 2011 08:52
engrid -> save as .stl with boundarie codes	Zymon	enGrid	31	August 29, 2011 13:40

March 23, 2017, 17:23		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	Have a look at the overlap % in the output file and in CFD-Post in the results file.

March 23, 2017, 21:24		#3
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	thank you for reply,Glenn Can the overlap % you mentioned be used to check the interface inside a part which contain many bodys but mesh may not connect like hex with tetra and said bodys are full in the same domain?

March 23, 2017, 21:38		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	The overlap % simply returns the % of the area of one side of the GGI which it has managed to find the matching face on the other side of the GGI. Whether the sides are in different domains or not, tet or hex mesh or anything like that does not matter.

March 23, 2017, 21:54		#5
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	Actually, I may not describe very clear. my case is a part contain many bodies (already form a new part in geometry) and only one body use subdomain to get energy source , and all bodies are in the same domain. When I finish the simulation , I find the temperature distribution is very strange and I guess the interface between the subdomain one and other nearby bodies many not correct connect . So I operate an easier model to test , I find if the meshes of said all bodies inside are connected ,it is not necessary to set interface in one domain , default setting is fine ,but if mesh are discontinued since node not math or using differ mesh method on different body (hex with tet),the interface inside will not exchange the heat. So my confusion is , Is there any way to check all mesh of a part contains many bodies really connected between every bodies inside a part ? or just to see by eyes and set interface even in the same doman and same part. thanks!

March 24, 2017, 00:45		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	Have a look at the non-overlap fraction in CFD-Post to see where your interface connects and where it does not.

March 24, 2017, 01:07		#7
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	Yes , I got i . thank you Glenn for reply it , so in some complicate CAD, even bodies are in one part maybe need to set interface manually or just improving the mesh. I appriciate it!! Otherwise, I see ANSYS HELP under the title "Non-overlap Boundary Conditions" mention that "Nonoverlap Fraction or (1-Nonoverlap Fraction) to ignore contributions from nodes on a GGI interface" I am confused about what is used for . like to set the heat flux turn to 0 ?

March 24, 2017, 06:16		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	I do not know the reference you are talking about. What section of the documentation is that from?

March 26, 2017, 10:34		#9
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	hey , my ref comes from here. could you help me to have a look? it located on the last sentence under 5.4.3.1 https://www.sharcnet.ca/Software/Ans.../CACBCJBF.html

March 26, 2017, 23:14		#11
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	Is it used on the source term and make the flux crossing the interface can be zero or anything? could you show me a simple example ? I am very appreciated

March 27, 2017, 07:11		#12
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	You simply put the Nonoverlap Fraction variable in the CEL expression to evaluate the source term. But before we talk about the details of this - what are you trying to?

March 27, 2017, 21:27		#13
hippo407 New Member John,Lu Join Date: Jul 2015 Posts: 10 Rep Power: 10	For my project , I am trying using auto generation for the interfaces, and I want to know whether this function is complete or not.By my test, I think it is powerful enough and only need to check the nonoverlap fraction as you said in the CFD post and find anywhere need to be set interface manually. And the last question is just my curious problem which I saw at the help. If there is any misunderstand concept ,plz tell me. thanks