Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Define simple-support boundary condition of a beam (Solid Mechanics)

Please login with a confirmed email address before reporting spam

Hello everyone,

I am a new Comsol user and through the past few months I have made simple models comparing the results with some theoretical models.

I will bother you with plenty of questions the next days(haha), and here is the first one:

I know till now that if we want to have a beam clamped or free we simply choose the edge and make it fixed constraint or free respectively.

Now if we need a simply-supported BC, how could we make that?
(I chose fixed constraint point and prescribed displacement=0 at a point and the results are pretty much the same but they do not match the theoretical ones.

hope to have nice conversations here,
Tsagarakis Nicolas

6 Replies Last Post Nov 11, 2015, 9:51 a.m. EST
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 7, 2013, 4:51 a.m. EDT
Hi

Do you mean that you want a beam = simple rectangle in 2D, to be fixed at 1 point/vertex ?, but rigid or to allow it to rotate ?

If you look into the theory of beams, ofthen the beam physics is easier to handle than SOLID, as the beams solve for displacement and rotation angles, so you have all 6 DoF at hand. I solid you need to rethink correctly your physics.

In SOLID if you fix a point, you ceate a singularity (except in 1D solid bt that is a beam without rotation approximation) any FEM programme will fail, as you cannot esitmate any derivative from a vertex point at the limit of two different materials. So always apply your BC to "boundaries" and use lower space elements only as exceptions, once you really know what you want to do.

If you want to make a hinge, today the easiest is to use a rigid connector and enforce the point of rotation on the "point" / vertex of your beam. This will "make the small edge fully rigid" but it enforces that forces and moments are evenly distributed over a (often small) boundary. In similarity, you do it without doubts and without thinking over it, by sing Terminal or ports in ACDC, you say you give a voltage on a port, means you set the full boundary to an average value = desired port voltage (or current) without studying in thedetails the propagaion of the voltage/current inside this boundary

One suggestion: use the latest version, it's far more stable than te early v4.

--
Good luck
Ivar
Hi Do you mean that you want a beam = simple rectangle in 2D, to be fixed at 1 point/vertex ?, but rigid or to allow it to rotate ? If you look into the theory of beams, ofthen the beam physics is easier to handle than SOLID, as the beams solve for displacement and rotation angles, so you have all 6 DoF at hand. I solid you need to rethink correctly your physics. In SOLID if you fix a point, you ceate a singularity (except in 1D solid bt that is a beam without rotation approximation) any FEM programme will fail, as you cannot esitmate any derivative from a vertex point at the limit of two different materials. So always apply your BC to "boundaries" and use lower space elements only as exceptions, once you really know what you want to do. If you want to make a hinge, today the easiest is to use a rigid connector and enforce the point of rotation on the "point" / vertex of your beam. This will "make the small edge fully rigid" but it enforces that forces and moments are evenly distributed over a (often small) boundary. In similarity, you do it without doubts and without thinking over it, by sing Terminal or ports in ACDC, you say you give a voltage on a port, means you set the full boundary to an average value = desired port voltage (or current) without studying in thedetails the propagaion of the voltage/current inside this boundary One suggestion: use the latest version, it's far more stable than te early v4. -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 8, 2013, 9:42 p.m. EDT
Ivar thanks a lot for your answer and for your time. Since I am still a student there are things you said that I didn't understand, but if you forgive me for that I will go on :P

If I understand it correctly you advise me to use the Beam module for solving the beam. I know that it works fine but I would like to see how it works in Solid mechanics cause I think (and help me on that) that the results (eigenfrequencies let's say) will be more close to the experimental than the theoretical ones.

Later you talk about the Rigid Connector BC and that's something I'm very happy to learn about. How exactly do I enforce the point of rotation on the "point"? I find it too chaotic.
And something else, I read in another thread that you would not use the RBC for eigenfrequency or modal analysis, is that correct? Cause eigenstudies and frequency domains are the fields that I am currently interested in.

P.S. I use versions 4.3 and 4.3a.
Ivar thanks a lot for your answer and for your time. Since I am still a student there are things you said that I didn't understand, but if you forgive me for that I will go on :P If I understand it correctly you advise me to use the Beam module for solving the beam. I know that it works fine but I would like to see how it works in Solid mechanics cause I think (and help me on that) that the results (eigenfrequencies let's say) will be more close to the experimental than the theoretical ones. Later you talk about the Rigid Connector BC and that's something I'm very happy to learn about. How exactly do I enforce the point of rotation on the "point"? I find it too chaotic. And something else, I read in another thread that you would not use the RBC for eigenfrequency or modal analysis, is that correct? Cause eigenstudies and frequency domains are the fields that I am currently interested in. P.S. I use versions 4.3 and 4.3a.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 9, 2013, 2:48 p.m. EDT
Hi

you are right there were (not sure if still the case) some limtations with RC's and eigenfrequency analysis, you need to look carefully at the equations of the nodes for the different solver cases, and adapt the formulas. Example you adda body load on a domain in SOLID (graviy weight load) and you have an added mass in an RC node, then you need to add also a RC subnode force that is g_const"Mass (of added mass to RC) IN ALL DIRECTIONS, where it might apply.

The best is to try these BC on smple examples, one at the time, until you get used to them and master them well.

Concerning beams and SOLID, I would not say the eigenfrequencies are more precise for one rather than another. It all depends on the complexity of your model. But wat is sure, a beam moel will solve 10x faster than a solid, if not more, but might require a bit more time to set up, all depends ...

--
Good luck
Ivar
Hi you are right there were (not sure if still the case) some limtations with RC's and eigenfrequency analysis, you need to look carefully at the equations of the nodes for the different solver cases, and adapt the formulas. Example you adda body load on a domain in SOLID (graviy weight load) and you have an added mass in an RC node, then you need to add also a RC subnode force that is g_const"Mass (of added mass to RC) IN ALL DIRECTIONS, where it might apply. The best is to try these BC on smple examples, one at the time, until you get used to them and master them well. Concerning beams and SOLID, I would not say the eigenfrequencies are more precise for one rather than another. It all depends on the complexity of your model. But wat is sure, a beam moel will solve 10x faster than a solid, if not more, but might require a bit more time to set up, all depends ... -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 16, 2014, 5:15 p.m. EST
Hello,

I am back to this matter again. I put the mass and the forces to the Rigid connector node as Ivar suggested, but I am not sure how to define the restraints. With "prescribed displacement" node I usually check the displacement in the z direction to be zero.I tried the same here (the others unchecked) and the prescribed rotation left it free.

However, the 4th frexural mode I get is the one in the picture I attached (I also attached the model), and I think that's not a proper mode for a simple support boundary condition (is it?).

Please give it a try and check it out, I am really curious to compare the results of the rigid connector vs prescribed displacement.

Regards,
Nicolas T.
Hello, I am back to this matter again. I put the mass and the forces to the Rigid connector node as Ivar suggested, but I am not sure how to define the restraints. With "prescribed displacement" node I usually check the displacement in the z direction to be zero.I tried the same here (the others unchecked) and the prescribed rotation left it free. However, the 4th frexural mode I get is the one in the picture I attached (I also attached the model), and I think that's not a proper mode for a simple support boundary condition (is it?). Please give it a try and check it out, I am really curious to compare the results of the rigid connector vs prescribed displacement. Regards, Nicolas T.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 17, 2014, 7:40 a.m. EST
In addition to that, I just found out that the rigid connector is not available in the Plate element. I can assume that this is because of the thickness? (although Comsol says that can calculate Midlin Plates too)

So here we have only the options of Prescribed displacement/Pinned, however Ι think I'm not sure about the role of the prescribed rotation regarding the simple support boundary condition.
In addition to that, I just found out that the rigid connector is not available in the Plate element. I can assume that this is because of the thickness? (although Comsol says that can calculate Midlin Plates too) So here we have only the options of Prescribed displacement/Pinned, however Ι think I'm not sure about the role of the prescribed rotation regarding the simple support boundary condition.

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Nov 11, 2015, 9:51 a.m. EST
Very Nice and informative discussion guys,

My question what will be the boundary condition for simply supported beam in 2D??
Where, I use beam physics and stationary study. Also used Bézier Polygon to define beam.

Should prescribe displacement in y direction =0 ? Is there any other things I need to define ???
Very Nice and informative discussion guys, My question what will be the boundary condition for simply supported beam in 2D?? Where, I use beam physics and stationary study. Also used Bézier Polygon to define beam. Should prescribe displacement in y direction =0 ? Is there any other things I need to define ???

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.