Multiple Support Excitation

[fmk]

you have to remove the homogenous single point constraints from the nodes in the dof direction at which you are applying the ground motion .. the vel and accel at these nodes will always be 0 (unless you apply vel and accel response as well) as these dof do not appear in the equation of motion (unless using penalty method).

[mnazari]

Hi,
For a simple cantilever column, if I excite the base with SIN excitation (displacement time history), I have moments at free node, which is not correct, What's the problem here, since model works well with free vibration analysis.
Thanks

[fmk]

your model is wrong or you convergence test is not tight enough .. try using NormUnbalance

[mnazari]

Thanks Frank, But it also does not work.
My complete model consists of different parts but I am going to excite just one support with a SIN Excitation so I used :

pattern MultipleSupport 2 {
set DispSeries "Sine 0. 5 0.5 -factor 0.01"
groundMotion 1 Plain -disp $DispSeries
imposedMotion 1 1 1
}

Point 1 is that support which I want to apply excitation. but it also jointed to node 2 with a zero length element link modeling a 2D spring, like as:
element zeroLength 1 1 2 -mat 1 -dir 6 -doRayleigh 1
fix 1 1 1 1 ; # node DX DY RZ
equalDOF 1 2 1 2 ;

Is it something make problem in case I have this type of excitation? since same model did not have any problem with free vibration analysis.
Thanks a lot

[fmk]

post the entire model.

[mnazari]

Thanks Frank, I have looked at Multiple support Excitation command in wiki and found out type of constraint handler is so important, so changed the constraint handler and now it's working!
But I have another question. When we apply excitation to one of three supports of a structure, all nodes' responses are absolute values? and for relative responses they should be reduced from support excitation?

[vesna]

Yes, you are right.

[mnazari]

Thanks Dear Vesna !

[andreascj]

Hi fmk,

That means, by applying velocity, we can get displacement. Hence the response after applying only velocity or only displacement is quite different. For example when we use this for SDF with spring and dashpot only, applying velocity will see damping as well as restoring forces but applying displacement will only see restoring forces. Strange!

Thanks.

fmk wrote:
> the vel and accel at these nodes will always be 0 (unless you apply vel and accel
> response as well) as these dof do not appear in the equation of motion
> (unless using penalty method).

[fmk]

if you only supply velocity the default is to use a simple trapezoidal integration to come up with what the displacements are (if only accel, trap to get vel and then trap to get disp). these will not be the same as the actual displacements. if you provide the dip the disp is set at the node, if vel or both vel and displacement are provided, vel and disp is set at the node. if accel ......... . all the constraints are imposed as non-homogeneous single point constraints at the dof at the node.

if you have the displacement record it is highly recommended that you use it. you only need to provide the vel and accel if the model is such that the vel and acccel impact the elements (stiff, resisting force) attached to whatever node you are constraining.

[andreascj]

Makes sense. Thanks fmk!