Eigen Value Analysis

[ankurjain]

In a 3D model, I have done a Eigen Value analysis without applying any load over it. The results came out approximately matches with the SAP2000 model.
But when I apply loads on the members and then I do the Eigen Analysis, I found that OpenSees is ignoring the Loads that I applied and giving the same Eigen value results as previous. While SAP2000 is considering the Loads that I applied and giving different Eigen Values.

SAP2000 add mass automatically....Just we have to define the mass source....In opensees we have to provide the mass density...
After I add load, the periods gets longer as soon as I add loads on the elements in SAP2000 but the same is not happening in Opensees....
In opensees, the periods remains unchanged.....So It seems to me that adding load on the elements in Opensees won't affect the Periods....
Is there any way that I can consider the loads also in opensees for Eigen analysis.?


Can anyone help why OpenSees is ignoring the loads that I am applying for Eigen Analysis.?

Thanks in advance

[mhscott]

OpenSees treats load and mass separately. So, you have to add mass when you add load if you want OpenSees to do what you're describing in SAP.

[ankurjain]

OpenSees treats load and mass separately. So, you have to add mass when you add load if you want OpenSees to do what you're describing in SAP.

are you suggesting that I have to convert load in to mass first and then apply it ?

Because I have already assigned self mass of the elements and the results are coming same as that of same. The problem is only when I add some additional loads on the elements in opensees.

[mhscott]

Yes. Add additional nodal mass at the element ends (based on the load you add to the elements) and see what happens.

[anuragiitg]

OpenSees will not convert load to mass (or the other way) for you. You could assign load as a function of the mass so you don't have to worry about it.
For example,

load 42 0 0 [expr -$massCol*$g] 0 0 0

[ankurjain]

OpenSees will not convert load to mass (or the other way) for you. You could assign load as a function of the mass so you don't have to worry about it.
For example,

load 42 0 0 [expr -$massCol*$g] 0 0 0

I have tried what you suggested but still the answer is the same as previous. This time also Opensees has not considered the load I applied after converting it to mass.

[ankurjain]

Yes. Add additional nodal mass at the element ends (based on the load you add to the elements) and see what happens.

After adding the equivalent mass at the nodes, I am getting the same results as that in SAP2000. Thank you so much for the help.

But one thing.......in case of a very big structure with so many loadings, how is it possible to convert all the loads in to mass and then apply at respective nodes. Won't there be differences in the reaction forces in the elements.

[anuragiitg]

[/quote]

I have tried what you suggested but still the answer is the same as previous. This time also Opensees has not considered the load I applied after converting it to mass.
[/quote]

You have to have a conversion both ways. In this case, you want the load to be converted in to a mass. For example,

set load 10; # assuming the load you are applying is 10 kips
set mass [expr $load/$g]

mass 42 $mass $mass $mass 0. 0. 0.

load 42 0 0 -$load 0 0 0

[geoStru]

OpenSees treats load and mass separately. So, you have to add mass when you add load if you want OpenSees to do what you're describing in SAP.

Can there be any differences of periods by SAP2000 and OpenSees (3/4%)?
I am attaching the results:
SAP2000: 0.453 0.414 0.380 0.127 0.120
OPENSEES: 0.435 0.399 0.370 0.121 0.115
I checked the total mass, the value of E and the constraints of the 2 models and they are the same.

[mhscott]

Are the models completely elastic? Are you using hinge models in SAP?

[geoStru]

In OpenSees I used nonlinearBeamColumn. Should I use the elasticBeamColumn to compare the results of the 2 models?
I didn't use hinge models in Sap

[mhscott]

Yeah, use all elastic elements and see if you get the same eigenvalues.

[selimgunay]

These periods look already close and with the elastic elements, it should even be closer. One other thing is that SAP2000 adds distributed masses due to the masses of the elements when you have nonzero densities for the materials. You can make the density zero in SAP2000 if these masses are negligible compared to the story masses due to slab, etc.