Learning Forum

[akndnzmn]

Hello

I'm working on 3D gear pump analysis. In the original model, two gears rotate by contacting each other, but since it will be easir to understand dynamic mesh logic, I reduce the model to a single gear and add it here.

 

 

I wrote a simple udf for the analysis model and named the inner surface of the gear. Then I turned on smoothing and remeshing for dynamic mesh, and defined rotational motion for the inner surface of the gear. I see the movement of the walls via Zone Motion without getting any problem, but when I run the analysis I get negative cell volume error. Also I get this error when trying to see mesh motion. I tried reducing the time step size and add dynamic mesh deformation to the side surface of gear, but no matter how much I reduced it or add deformation to the walls, the problem was not solved. In addition, after each error, the meshes appear to be corrupted as follows.

 

Can someone show me where I make mistake or help me understand how I can perform this 3D gear pump analysis with UDF?

[Federico]

Hello, 

Remeshing is currently not compatible with polyhedral type cells. I suggest your remesh your model with tetrahedral cells.

[akndnzmn]

 

 

Hello Again,

 

I changed mesh type as you suggest. Also I add deforming dynamic mesh zone to the region that include gear, but I have a small problem. When I start analysis it solve like 100 iteration, but after that it gives those warnings:

Then analysis were crashed due to negative cell volume. I guess I need to increase mesh quality, and change some dynamic mesh settings, but neither I know how to increase mesh quality (because clearence between gear and walls are nearly 10 micron), nor what properties should I change for dynamic mesh settings. So what should I do?

 

 

 

[Federico]

What is your initial mesh quality? Do you get these warnings from the start or after some iterations? Perhaps your initial mesh is Ok but the remeshing settings are not.

What method are using for Remeshing? We recommend using Unified Remeshing method.

Can you show a screenshot of your Dynamic Mesh settings? Identifying each zone?

[akndnzmn]

Hi Federico,

 

I couldn't get back because I didn't have access to the data over the weekend. Sorry for the late reply. I add every data below;

 

Mesh Statics

- I didn't pay much attention when meshing this model, because in the original model with 2 gears, the clearence dropped to 10 micron, and I will have to focus mostly on the mesh while working on the original model. So in this model I just focused on understanding how dynamic mesh works in 3-D.

 

 

 

Model

 I divided the model into 3 different bodies: inlet, outlet and udf zone. After that, I named the faces of the gear with named selection. I don't know why, but I defined the side surfaces of the gear as symmetry, as in the gear analysis video I found in Ansys Resource Center. Also, I defined the surfaces between the UDF body and the inlets and outlets as interfaces.

Interfaces

 

 

Walls

 

Apart from this, I defined all the surfaces shown below as walls.

Dynamic Mesh Settings

I tried both spring method and diffusion as smoothing in dynamic mesh settings. As for remeshing, I preferred methods-based remeshing and reduced the remeshing interval to 1 with default sizing settings. After reading your comment, I changed the remeshing to unified remeshing and ran the analysis again. Analysis made more iterations this time, but again gave the negative cell volume error. Below I added the mesh image for before the analysis and the after the error.

- Before Solution

-After Negative Cell Error

 

 

 

Rigid Body Type Dynamic Zone

Deforming Type Dynamic Zone

  

Some Results

-          Since I can’t add video here I shown some results as a images. These images are taken from flow animation.

- Although the mesh quality is poor at the beginning of the analysis, the analysis solves for a while and then gives an error. The reason for this confuses me.

Although the mesh quality is poor at the beginning of the analysis, the analysis solves for a while and then gives an error. The reason for this confuses me.