Element Size investigation for stress analysis on square channel with pressure

Element Size investigation for stress analysis on square channel with pressure

Freek046
Contributor Contributor
793 Views
3 Replies
Message 1 of 4

Element Size investigation for stress analysis on square channel with pressure

Freek046
Contributor
Contributor

Dear all,

I am looking at the stress on a square channel with pressure. I therefore am investigating what Element Size I should choose. For simplicity I have cut open my channel to reduce elements.  This is compensated with two symmetrical constraints (constraint 2 and 3, see pictures). I hope the enclosed pictures make clear what I am doing.

 

The problem is that when I make the Elements smaller, the stress goes up and there seems to be no end to this. I did a trial with a channel with a 0.5 mm radius in the corner, which does not have this problem. Here the value of the stress stabilizes at some point. So, I think it has something to do with the sharp edge, which I really want.

 

Does anybody know what I should do?

0 Likes
Accepted solutions (1)
794 Views
3 Replies
Replies (3)
Message 2 of 4

plafortune
Enthusiast
Enthusiast
Accepted solution

Hi, 

 

If you keep refining the mesh in a sharp corner, like the one showed in the first image, the mesh will never converge. This is called a "stress singularity". Read more on this: http://www.acin.net/2015/06/02/stress-singularities-stress-concentrations-and-mesh-convergence/

 

To overcome this, you can add a fillet radius and/or a plastic material model. 

 

But seeing your images, it's not clear to me if you add a radius (an "arc"), or some other shape.

 

plafortune_0-1612780381362.png

[EDIT: ok! I think I get it. this shape is the "deformed" shape of the mesh. You didn't add any fillet...]

 

In any cases, to see if the mesh is "converging", plot on a graph mesh size versus stress on the concerned area, and see if the curve is becoming flat, like this one: 

 

plafortune_1-1612780459108.png

 

 

--
Pierre Lafortune
idrasimulation.com
0 Likes
Message 3 of 4

Freek046
Contributor
Contributor

Thanks for the explanation Pierre, that makes sense. The enclosed pictures are all without radius, so what you see is just deformation. I will have a look at your suggestions and come back later.

Message 4 of 4

Freek046
Contributor
Contributor

I decided I had to go for a model with a radius (fillet) because a 'plastic model' is no option for me due to the brittle nature of the material I am actually doing this for. The plot of the Stress vs Element Size approach works quite well. The big question though is for what size of radius I will set. This is still to be investigated.

 

For anyone who is also looking at this, I think the following video is very good.

 

https://youtu.be/9l9ZngKZRPs

0 Likes