RigidLink beam & RigidDiaphragm issues

[giovannimilan]

Dear fmk,

I saw you are the developer of rigidLink and diaphragm. ive got a lumped plasticity models with rotational springs at the 2 ends of columns.

I am getting a very weird interaction between the EqualDOF necessary for the rotational spring to beam, and the the rigid link or diaphragm that i am using for modelling the spandrel/floor.

I am wondering if there is an interaction between the 2 somehow. It seems like the last one is overwriting the equal dof and the deformed shape is really messy.

Could you think of a reason for it?

[fmk]

transformation handler cnnot handle one node constrained to another constrained to another. you should have been seeing warning messages. try Penalty method, but you may have numerical problems due to conditioning as some of the penalty factors are going to add up if your values are poorly chosen.

[giovannimilan]

thanks a lot fmk

[waskita]

I see.. thanks a lot fmk

[giovannimilan]

fmk wrote:
> transformation handler cnnot handle one node constrained to another
> constrained to another. you should have been seeing warning messages. try
> Penalty method, but you may have numerical problems due to conditioning as
> some of the penalty factors are going to add up if your values are poorly
> chosen.

fmk,

Thanks for the prompt reply. I observe the same issue when using "sp" for the simulation of the loading protocol from the lab.

In this case i dont even get any warning, but with the post processing tool I see that the equalDOFs are not respected at all.

I ve tried both with penality and transforamtion.

This means that in OpenSees is not possible perform a loading protocol with displacement control for a lumped plasticity model (such as "Dynamic Analysis of 2-Story Moment Frame" in which at floor level the nodes are constrained to the bottom and top nodes via equalDOF) ?

Regards,

[giovannimilan]

fmk wrote:
> transformation handler cnnot handle one node constrained to another
> constrained to another. you should have been seeing warning messages. try
> Penalty method, but you may have numerical problems due to conditioning as
> some of the penalty factors are going to add up if your values are poorly
> chosen.

Solved it with very high values of penalty coeff (e.g. 1.E18), deformed shape and F-disp response seems clean)

Cheers,

[giovannimilan]

fmk wrote:
> transformation handler cnnot handle one node constrained to another
> constrained to another. you should have been seeing warning messages. try
> Penalty method, but you may have numerical problems due to conditioning as
> some of the penalty factors are going to add up if your values are poorly
> chosen.

fmk, do you know if such a configuration ("one node constrained to another constrained to another"), could cause problems also with the Eigen command? I am plotting the modal shape with the window command and seems that the EquaDOF's are not fulfilled. I wondering if it is the same issues u described above (constrains handling) or I am messing things up with the window command (I am using an existing script, tho).

Regards,

[wuhaoshrek]

fmk wrote:
> transformation handler cnnot handle one node constrained to another
> constrained to another. you should have been seeing warning messages. try
> Penalty method, but you may have numerical problems due to conditioning as
> some of the penalty factors are going to add up if your values are poorly
> chosen.

I see rigidDiaphragm command can only be used for 3D problem. What if I want to model the rigidDiaphragm effect in a 2D problem? Other than using elastic element with large stiffness, if there any options? thanks.

[giovannimilan]

wuhaoshrek wrote:
> fmk wrote:
> > transformation handler cnnot handle one node constrained to another
> > constrained to another. you should have been seeing warning messages. try
> > Penalty method, but you may have numerical problems due to conditioning as
> > some of the penalty factors are going to add up if your values are poorly
> > chosen.
>
> I see rigidDiaphragm command can only be used for 3D problem. What if I want to model
> the rigidDiaphragm effect in a 2D problem? Other than using elastic element with
> large stiffness, if there any options? thanks.

Maybe with the rigidLink beam.

[wuhaoshrek]

giovannimilan wrote:
> wuhaoshrek wrote:
> > fmk wrote:
> > > transformation handler cnnot handle one node constrained to another
> > > constrained to another. you should have been seeing warning messages. try
> > > Penalty method, but you may have numerical problems due to conditioning as
> > > some of the penalty factors are going to add up if your values are poorly
> > > chosen.
> >
> > I see rigidDiaphragm command can only be used for 3D problem. What if I want to
> model
> > the rigidDiaphragm effect in a 2D problem? Other than using elastic element with
> > large stiffness, if there any options? thanks.
>
> Maybe with the rigidLink beam.
that's incorrect cause rigid link will constraint translationnal DOFs.

[giovannimilan]

wuhaoshrek wrote:
> giovannimilan wrote:
> > wuhaoshrek wrote:
> > > fmk wrote:
> > > > transformation handler cnnot handle one node constrained to another
> > > > constrained to another. you should have been seeing warning messages.
> try
> > > > Penalty method, but you may have numerical problems due to
> conditioning as
> > > > some of the penalty factors are going to add up if your values are
> poorly
> > > > chosen.
> > >
> > > I see rigidDiaphragm command can only be used for 3D problem. What if I
> want to
> > model
> > > the rigidDiaphragm effect in a 2D problem? Other than using elastic element
> with
> > > large stiffness, if there any options? thanks.
> >
> > Maybe with the rigidLink beam.
> that's incorrect cause rigid link will constraint translationnal DOFs.

I am quoting the wiki page

$type beam both the translational and rotational degrees of freedom are constrained.

source: http://opensees.berkeley.edu/wiki/index ... nk_command

[wuhaoshrek]

giovannimilan wrote:
> wuhaoshrek wrote:
> > giovannimilan wrote:
> > > wuhaoshrek wrote:
> > > > fmk wrote:
> > > > > transformation handler cnnot handle one node constrained to
> another
> > > > > constrained to another. you should have been seeing warning
> messages.
> > try
> > > > > Penalty method, but you may have numerical problems due to
> > conditioning as
> > > > > some of the penalty factors are going to add up if your values
> are
> > poorly
> > > > > chosen.
> > > >
> > > > I see rigidDiaphragm command can only be used for 3D problem. What if
> I
> > want to
> > > model
> > > > the rigidDiaphragm effect in a 2D problem? Other than using elastic
> element
> > with
> > > > large stiffness, if there any options? thanks.
> > >
> > > Maybe with the rigidLink beam.
> > that's incorrect cause rigid link will constraint translationnal DOFs.
>
> I am quoting the wiki page
>
> $type beam both the translational and rotational degrees of freedom are constrained.
>
> source: http://opensees.berkeley.edu/wiki/index ... nk_command

YES, I know. But RigidLink beam will constrain all the DOF while RigidLink bar will constraain all the translational DOFs (vertical and horizontal) in a 2D problem, which is not right for modeling a rigid diagphram. Rightnow I believe in a 2D problem, if you wanna model rigid diagphram, just use EqualDOF command and make all the rotational DOFs of the nodes in that plane to be the same, that's it. The horizontal DOFs should not be constraint, rather depending on the axial stiffness of beams. That's what I thought.