increase the mesh and change in results?

heydari · July 30, 2015, 16:40

Hello:
Why increase the mesh, the results are far from the reality?
Of course, after every picture you can see, the results
Especially in right side of semi-cylinder on the the experimental results distances!

Opaque · July 30, 2015, 16:46

What do you mean by increasing the mesh ?

What are you increasing ? Your images from left to right are increasing the mesh element size, not the mesh element count.

heydari · July 30, 2015, 17:22

What is the difference between the element size and element count there? And what to do with my question?

ghorrocks · July 30, 2015, 19:29

I do not understand your question either. Can you write it again more clearly?

heydari · July 31, 2015, 02:09

hi;
I have some exprimental data. After each of these conditions mesh changed .Data modeling has been compared with the results. In the next section of the half cylinder is not good results. And every time fine mesh results more Different. Why? What should I do?

ghorrocks · July 31, 2015, 02:15

That means your mesh is not fine enough to achieve adequate mesh insensitivity. You need a finer mesh still. It is normal for the results to jump around when the mesh is way too coarse.

This assumes that the remainder of your simulation is correct, of course.

heydari · July 31, 2015, 02:37

Dear ghorrocks;
With the increase in the fine mesh model can be meshed together in the third dimension. While I'm a two-dimensional model. (1 mm thick two-dimensional model. The maximum length of 5 meters is my model).

ghorrocks · July 31, 2015, 02:43

I do not understand your last comment.

Also flow over a cylinder has regimes where it is three dimensional. If you model a three dimensional flow with a two dimensional model you will get unpredictable results. Are you sure a two dimensional model is appropriate for the flow regime you are in?

heydari · July 31, 2015, 02:58

High time software solution, I decided to do a two-dimensional model. Is this wrong?

ghorrocks · July 31, 2015, 03:32

If you are saying you chose to do a two dimensional simulation of a three dimensional flow because the simulation time was too long - if you want an accurate answer then yes, that is wrong.

heydari · July 31, 2015, 03:48

Why wrong? More Info
Your suggestion to solve this problem?

heydari · July 31, 2015, 03:50

Quote:

Originally Posted by ghorrocks

If you are saying you chose to do a two dimensional simulation of a three dimensional flow because the simulation time was too long - if you want an accurate answer then yes, that is wrong.

Why wrong? More Info
Your suggestion to solve this problem?

ghorrocks · July 31, 2015, 03:51

Desmond is right. If you change your flow conditions to one where it is a two dimensional flow (possibly by reducing the flow velocity) then you can do a two dimensional model. There are good results to compare against for this flow, so it is a good benchmark simulation to make sure you can do an accurate simulation before taking on the more difficult three dimensional simulation.

ghorrocks · July 31, 2015, 04:01

Quote:

Why wrong? More Info

If you model a 3D flow with a 2D model you have unphysically constrained the flow. Your simulation will have an error, and it will probably be large.

Quote:

Your suggestion to solve this problem?

That is pretty obvious, isn't it? Model it in 3D.

heydari · July 31, 2015, 04:02

thanks alot

July 31, 2015, 03:51		#13
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	Desmond is right. If you change your flow conditions to one where it is a two dimensional flow (possibly by reducing the flow velocity) then you can do a two dimensional model. There are good results to compare against for this flow, so it is a good benchmark simulation to make sure you can do an accurate simulation before taking on the more difficult three dimensional simulation. anon_h and heydari like this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
sliding mesh problem in CFX	Saima	CFX	46	September 11, 2021 08:38
Significant change in results with changing either mesh size or time step size.	nettpanic	FLUENT	5	October 24, 2014 03:37
[snappyHexMesh] Layers:problem with curvature	giulio.topazio	OpenFOAM Meshing & Mesh Conversion	10	August 22, 2012 10:03
Mesh sensitivity when using phase change models	BKaus	STAR-CCM+	0	February 28, 2012 21:51
[Gmsh] 2D Mesh Generation Tutorial for GMSH	aeroslacker	OpenFOAM Meshing & Mesh Conversion	12	January 19, 2012 04:52

July 30, 2015, 16:46		#2
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	What do you mean by increasing the mesh ? What are you increasing ? Your images from left to right are increasing the mesh element size, not the mesh element count.

July 30, 2015, 17:22		#3
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	What is the difference between the element size and element count there? And what to do with my question?

July 30, 2015, 19:29		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	I do not understand your question either. Can you write it again more clearly?

July 31, 2015, 02:09		#5
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	hi; I have some exprimental data. After each of these conditions mesh changed .Data modeling has been compared with the results. In the next section of the half cylinder is not good results. And every time fine mesh results more Different. Why? What should I do?

July 31, 2015, 02:15		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	That means your mesh is not fine enough to achieve adequate mesh insensitivity. You need a finer mesh still. It is normal for the results to jump around when the mesh is way too coarse. This assumes that the remainder of your simulation is correct, of course.

July 31, 2015, 02:37		#7
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	Dear ghorrocks; With the increase in the fine mesh model can be meshed together in the third dimension. While I'm a two-dimensional model. (1 mm thick two-dimensional model. The maximum length of 5 meters is my model).

July 31, 2015, 02:43		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	I do not understand your last comment. Also flow over a cylinder has regimes where it is three dimensional. If you model a three dimensional flow with a two dimensional model you will get unpredictable results. Are you sure a two dimensional model is appropriate for the flow regime you are in?

July 31, 2015, 02:58		#9
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	High time software solution, I decided to do a two-dimensional model. Is this wrong?

July 31, 2015, 03:32		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	If you are saying you chose to do a two dimensional simulation of a three dimensional flow because the simulation time was too long - if you want an accurate answer then yes, that is wrong.

July 31, 2015, 03:48		#11
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	Why wrong? More Info Your suggestion to solve this problem?

July 31, 2015, 04:02		#15
heydari Member A.heydari Join Date: May 2015 Location: IRAN Posts: 31 Rep Power: 11	thanks alot