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.

Help setting boundary conditions for solid mech structure

Please login with a confirmed email address before reporting spam

Hi, kinda new to comsol so please be gentle.

I have a structure consisting of a plate with 3 holes that is to be mounted on a solid concrete wall. two beams are welded to this plate and load is applied to these beams.

The plate and the beams have been modeled as one solid for now, i'm not interested in the stresses on the bolts or the wall. I have tried setting different conditions like fixed constraint and prescribed displacement but I cant quite figure out how to model the plate in a realistic way so that it's allowed to deform but now go "into" the wall that doesnt exist.



3 Replies Last Post 19.03.2011, 18:36 GMT-4
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 19.03.2011, 09:04 GMT-4
Hi

for the plate face (the back one against the wall) use "roller" constraints, this will block it from moving in the direction perpendicular to the plane X (but all over its a best case stiffening). Equivalent to a hard wall and stiff plate

To be more realistic, you could cut your plate by two cylinders about 3-4 times the diameter of the bolts and block only these in roller mode, then you have a soft wall and the stiffness is only from the plate. other extreme

You could also add some blocked displacement in the direction perpendicular of the plate face for the lower edges

Then you need to block the Y-Z motion (if I assume the plate normal along X) for this you could block the interiour cylinder of the bolt holes in Y and Z direction. Or even try a RBE (rigid body constraint on the bolt cylinder area) and constraint the centre (and perhaps the torsion)

It is worth to see the deformations and stress (von mises) cases for all these fixing conditions, one by one, then by combining them, to understand how they apply, sometimes one overconstrain, other times one forget some degrees of freedom.

Fixing rigid is not a trivial case, and its a good exercice to learn how to handle blocking of complex shapes

Another interesting thing: with a full roller constraint (as first mentionned) add the surface arrow of the reaction forces along X (solid.RFx) it will give you some usefull information how the wall load (and in which direction) is applied. Then you can scan through all 3 reaction forces AND reaction moments solid.RMx,y,z
--
Good luck
Ivar
Hi for the plate face (the back one against the wall) use "roller" constraints, this will block it from moving in the direction perpendicular to the plane X (but all over its a best case stiffening). Equivalent to a hard wall and stiff plate To be more realistic, you could cut your plate by two cylinders about 3-4 times the diameter of the bolts and block only these in roller mode, then you have a soft wall and the stiffness is only from the plate. other extreme You could also add some blocked displacement in the direction perpendicular of the plate face for the lower edges Then you need to block the Y-Z motion (if I assume the plate normal along X) for this you could block the interiour cylinder of the bolt holes in Y and Z direction. Or even try a RBE (rigid body constraint on the bolt cylinder area) and constraint the centre (and perhaps the torsion) It is worth to see the deformations and stress (von mises) cases for all these fixing conditions, one by one, then by combining them, to understand how they apply, sometimes one overconstrain, other times one forget some degrees of freedom. Fixing rigid is not a trivial case, and its a good exercice to learn how to handle blocking of complex shapes Another interesting thing: with a full roller constraint (as first mentionned) add the surface arrow of the reaction forces along X (solid.RFx) it will give you some usefull information how the wall load (and in which direction) is applied. Then you can scan through all 3 reaction forces AND reaction moments solid.RMx,y,z -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19.03.2011, 15:20 GMT-4
Thank you, been trying out different boundaries now, i dont quite understand what you mean by the cylinder-approach. Should i cut out 2 holes in the "middle" of the plate and apply roller boundary on the inside of the holes allowing the plate so deform around the holes while restricting deplacement along the x-axis?
Thank you, been trying out different boundaries now, i dont quite understand what you mean by the cylinder-approach. Should i cut out 2 holes in the "middle" of the plate and apply roller boundary on the inside of the holes allowing the plate so deform around the holes while restricting deplacement along the x-axis?

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 19.03.2011, 18:36 GMT-4
Hi

for the cylinders, you have now 2 holes I see, you make another cylinder about 3 times the bolt cylinder diameter, but you cut only the volume, or even just the back surface. This is just to define a smaller area to which to apply a roller condition (or a restricted X displacement).

I often cut up my volumes and leave some internal boundaries, this helps to define control volumes/surfaces, for metrology and for more realistic BC definitions. Three is not much extra numerical load by adding a few of these boundaries. Meshing becomes often even easier, as one can select the domains independently

--
Good luck
Ivar
Hi for the cylinders, you have now 2 holes I see, you make another cylinder about 3 times the bolt cylinder diameter, but you cut only the volume, or even just the back surface. This is just to define a smaller area to which to apply a roller condition (or a restricted X displacement). I often cut up my volumes and leave some internal boundaries, this helps to define control volumes/surfaces, for metrology and for more realistic BC definitions. Three is not much extra numerical load by adding a few of these boundaries. Meshing becomes often even easier, as one can select the domains independently -- Good luck Ivar

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.