[peatmac]

Hello,

I am currently trying to simulate freely oscillating 2d cylinders as part of my University Thesis. At the moment I forcing the cylinders to oscillate, after which I plan to apply a UDF to make the oscillate freely with 1 DOF then 2 DOF.
At the moment I keep having problems with my dynamic mesh, I constantly get negative cell volumes. I can get the problem to work using a very simple tri-mesh. I was advised the I would be better off using a quad-mesh. After creating a simple mesh (see attached), I can make fluent collapse and re create new cells in the mesh.
Ideally I want all the re-meshing to be done as far away from the cylinder as possible, so I think creating a zone of stationary fluid/mesh near the cylinder, and allow the mesh to adapt further away. This should give me the most accurate flow results.
Attached are some pictures of me mesh and all of my dynamic mesh parameter. Can someone kindly give me some pointers on how to solve my problem?

Thanks

Peter

[peatmac]

If anyone would like I can forward them a .cas file with the mesh, UDF and so on. Any help will be greatly appreciated.

[trantoan2008]

Originally Posted by peatmac

If anyone would like I can forward them a .cas file with the mesh, UDF and so on. Any help will be greatly appreciated.

Hi,

For quad element, you have to type additional commands like below:

define/models/dynamic-mesh-controls/smoothing-parameter/spring-on-all-shapes
yes

Then, all elements in computational domain will be remeshed.
I hope it is usefull for you.

[peatmac]

Many thanks for the reply trantoan2008.

This has allowed the cylinder to displace further in the Y direction; however I am still running into negative cell volumes. No new elements are being created on the bottom edge of the cylinder and all the elements at the top are being squashed until they have negative cell volume.

My Dynamic Mesh Parameters are -

Smoothing-
Spring constant factor = 0.5
Boundary node relaxation = 1
Convergence tolerance = 0.001
Number of iterations = 20

Remeshing -
Min length scale = 0.03705723
Max length scale = 0.1523991
Max cell skewness = 0.6
Size remesh interval = 10

(remeshing values taken from Mesh scale info)

Dynamic Mesh Zones -
I have applies a simple sinusoidal motion UDF to the cylinder, the adjacent fluid is fluid 1, with a cell height of 0.1m.
The upper and lower walls of the mesh are defined as stationary; the adjacent fluid is fluid 1, both with a cell height of 0.1m

I have also turned on the spring-on-all-shapes parameter.

Do you have any further advice on how I can get new cells forming at the bottom of the cylinder and make the cells at the top of the cylinder collapse?

Thanks for your help so far.

Peter

[trantoan2008]

Originally Posted by peatmac

Many thanks for the reply trantoan2008.

This has allowed the cylinder to displace further in the Y direction; however I am still running into negative cell volumes. No new elements are being created on the bottom edge of the cylinder and all the elements at the top are being squashed until they have negative cell volume.

My Dynamic Mesh Parameters are -

Smoothing-
Spring constant factor = 0.5
Boundary node relaxation = 1
Convergence tolerance = 0.001
Number of iterations = 20

Remeshing -
Min length scale = 0.03705723
Max length scale = 0.1523991
Max cell skewness = 0.6
Size remesh interval = 10

(remeshing values taken from Mesh scale info)

Dynamic Mesh Zones -
I have applies a simple sinusoidal motion UDF to the cylinder, the adjacent fluid is fluid 1, with a cell height of 0.1m.
The upper and lower walls of the mesh are defined as stationary; the adjacent fluid is fluid 1, both with a cell height of 0.1m

I have also turned on the spring-on-all-shapes parameter.

Do you have any further advice on how I can get new cells forming at the bottom of the cylinder and make the cells at the top of the cylinder collapse?

Thanks for your help so far.

Peter

Hi,

Send fluent and UDFs file to me so that I can check it.

Best

[peatmac]

Files are attached.

[Spring.s]

Hi
I proffer unstractural mesh for your case because i had same problem former.

[grayback87]

Hi i had the same problem, Then i had to make tri-mesh, and then, spring constant factor=0.0

then it works... but, i have other problems.

When It works, even without moving my CL it isn´t simetrical!! (I think the cause is i'm using triangular meshing near the cylinder).
I encourage you, once you get the correct mesh try to get some Cd with different Reynolds number and compare with a text book.

[grayback87]

Originally Posted by Spring.s

Hi
I proffer unstractural mesh for your case because i had same problem former.

Did you use Tri-mesh (pave)?
Did you check your CL ? It was symmetrical?

I have that problem, and I made the mesh many times and I can`t get it .



I need help

[Spring.s]

Yes,
Course i have one Airfoil instead your cylinder and my modeling include the entire of geometry in Gambit no symmetrical. For validation, i compare CL with experimental data.