Menu

Discussion Forum

New Discussion
Filter

How to model weld lines connecting two plates

Posted May 24, 2025, 8:45 a.m. EDT Geometry, Modeling Tools & Definitions, Structural Mechanics 4 Replies

Leander Hildebrandt
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi there,

I'm trying to simulate the stress that is exerted on weld lines that join two plates together. Now how do I model the weld lines? Do I create a geometry for them or is there a way to connect surfaces only in defined areas?

Thanks a lot in advance

Reply
4 Replies Last Post May 29, 2025, 11:01 a.m. EDT
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 3 days ago May 26, 2025, 4:54 a.m. EDT

The first question is if it is a solid model or a shell model.

In the Shell interface, you have the Edge to Edge feature, which has built-in options for weld stress evaluation. You also have the Edge to Boundary feature.

-------------------
Henrik Sönnerlind
COMSOL
The first question is if it is a solid model or a shell model. In the Shell interface, you have the *Edge to Edge* feature, which has built-in options for weld stress evaluation. You also have the *Edge to Boundary* feature.
Leander Hildebrandt
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 3 days ago May 26, 2025, 5:00 p.m. EDT

Dear Mr. Sönnerlind,

thank you for taking the time to respond to my question. In my case I used a solid model. Do I have to switch to shell models or can I use solid models to achieve reliable results as well?

Thanks a lot

Dear Mr. Sönnerlind, thank you for taking the time to respond to my question. In my case I used a solid model. Do I have to switch to shell models or can I use solid models to achieve reliable results as well? Thanks a lot
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 days ago May 27, 2025, 3:10 a.m. EDT

The problem with solid modeling of welds is that you get 'too good' results. The stresses will depend on the local weld geometry (which in practice is seldom known). So in solid models, the most common approach is to evaluate the stresses at a certain standardized distance away from the weld and then extrapolate.

The only time when it makes sense to do a detailed stress evaluation in a weld is when it has been carefully processed so that all irregularities are removed. This would, for example, be the case in nuclear engineering.

Evaluation of stresses is weld can almost be viewed as a separate field of engineering.

In this blog post, you can get some introduction to weld modeling:

https://www.comsol.com/blogs/how-to-predict-the-fatigue-life-of-welds

It is also mentioned briefly in

https://www.comsol.com/blogs/how-should-i-evaluate-singular-stress-fields

Which approach you should choose depends on several factors, including ease of modeling, number an type of welds, and the type of results you need. But if it is a standard check of fatigue life of welds between plates, then a shell model is usually easier to work with.

-------------------
Henrik Sönnerlind
COMSOL
The problem with solid modeling of welds is that you get 'too good' results. The stresses will depend on the local weld geometry (which in practice is seldom known). So in solid models, the most common approach is to evaluate the stresses at a certain standardized distance away from the weld and then extrapolate. The only time when it makes sense to do a detailed stress evaluation in a weld is when it has been carefully processed so that all irregularities are removed. This would, for example, be the case in nuclear engineering. Evaluation of stresses is weld can almost be viewed as a separate field of engineering. In this blog post, you can get some introduction to weld modeling: It is also mentioned briefly in Which approach you should choose depends on several factors, including ease of modeling, number an type of welds, and the type of results you need. But if it is a standard check of fatigue life of welds between plates, then a shell model is usually easier to work with.
Leander Hildebrandt
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 hours ago May 29, 2025, 11:01 a.m. EDT

Dear Mr. Sönnerlind,

thanks for you detailed response. However I still have two questions.

The thing about my simulation is that its about laser welding inside of glass with tooling that is used in the MEMS industry so there are almost no irregularities. In fact the geometry of the weld lines is well known. Therefore I think a detailed stress analysis would make sense.

The thing about my simulation is that the results don't match the reality.

First of all the simulation shows that the welds would break at far lower forces than they did in real world experiments. Second, when integrating the stress over the cross section of the weld, the resulting force is about 10% higher than the applied forces. For example a force of 10 N is applied on a boundary and the stress integral of the cross section results in 10.1 N. This means that there's somehow more stress than there should be. (I hope the explanation makes sense, let me know if it doesn't)

Therefore I wanted to ask: 1. What are common causes for simulations to calculate more stress than in reality? 2. How come that integrating the stress over the cross section results in more force than what was exerted on the model?

Dear Mr. Sönnerlind, thanks for you detailed response. However I still have two questions. The thing about my simulation is that its about laser welding inside of glass with tooling that is used in the MEMS industry so there are almost no irregularities. In fact the geometry of the weld lines is well known. Therefore I think a detailed stress analysis would make sense. The thing about my simulation is that the results don't match the reality. First of all the simulation shows that the welds would break at far lower forces than they did in real world experiments. Second, when integrating the stress over the cross section of the weld, the resulting force is about 10% higher than the applied forces. For example a force of 10 N is applied on a boundary and the stress integral of the cross section results in 10.1 N. This means that there's somehow more stress than there should be. (I hope the explanation makes sense, let me know if it doesn't) Therefore I wanted to ask: 1. What are common causes for simulations to calculate more stress than in reality? 2. How come that integrating the stress over the cross section results in more force than what was exerted on the model?

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content