ROBOT Modal analysis results are different from other programs

[erginc]

Hi all,
We are currently working on a project requires Russion SNIP code. In order to speed up things we decided to try Robot Structural analysis which is compatible with SNIP codes. But we are very new to Robot Structural, so we don't know the inner workings and assumptions that is made during the analysis.
So we decided to do a simple study to learn it. However, we have encountered a problem in ROBOT with modal analysis results (different periods with SAP2000). Which we cannot figure out why it happens.
We have modeled the same simple 3D structure in SAP2000 with all material, section properties and loading being identical. So we hoped to see very close results in terms of support reactions, M diagrams, and modal periods.
Support reactions are identical for both ROBOT and SAP2000, M diagrams (for BEAMONLY model) are quite close, still I would expect closer results. But modal analysis results are different.

Building Models:

1. BEAMONLY: This is a model with concrete beam-column frames. Loading is a little different, simply all beams have DL2 loading with value 0.5t/m, and LL loading with value 0.75t/m. No floor diaphragm is defined.

Since this structure is very simple, one can expect closer results from both programs.

1. BASICFLOOR: This model is same with other model but beams does not have loading instead defined floor members are loaded with DL2=0.2t/m², and LL=0.5t/m². However, in ROBOT we have chosen rigid-floor option so no finite element and no contribution to stiffness matrix. Only simple load distribution algorithm is used to calculate beam loads transferred from floors. On the other hand, SAP2000 does not have such option (at least i don't know how to) therefore we have chosen membrane type element in order not to change lateral stiffness, added diaphragm for rigid floor. And loads on floor is transferred to column tops directly (no beam loading in SAP). Therefore, M diagrams will not agree while total reactions are. But still mass matrix should be similar as a result periods too for both program.

Model	T1	T2	T3
ROBOT_BEAMONLY	0.71	0.51	0.48
SAP_BEAMONLY	0.59	0.45	0.40
ROBOT_BASICFLOOR	0.48	0.23	0.23
SAP_BASICFLOOR	0.62	0.45	0.43

Some thoughts:

1. Since M diagrams are quite close between Robot and SAP models, I think analytical models are similar. No auto rigid end zone for ROBOT or any other automated model modification affecting the results. Materials definitions should be correct too.
2. May be mass definition is somehow different. I actually defined in both program to convert loads to mass in the same manner. Or I think I did it correct.
3. Also results of two ROBOT models, BEAMONLY and BASICFLOOR, are unexpected. For model BEAMONLY, total load is less than BASICFLOOR model. SAP results show an increase in period for basic floor model. However, ROBOT results goes other way. It gives higher period values for BEAMONLY model. I am not sure but when I created BEAMONLY model from BASICFLOOR model, I deleted floors and also loads defined for floors. Somehow mass conversion table was deleted. So I redefined them. May be mass is doubled even though I cannot see it on the table. Is there a way to output total MASS in ROBOT?

Any help would be much appreciated, since ROBOT is quite foreign to me at the moment. Thanks.
Best Regards,
Ergin Ceyhan
P.S.: I have provided SAP2000 models in s2k files, and ROBOT 2013 files as attached. M diagrams of BEAMONLY model is also attached.

[kmarsh]

Hi Ergin,

First of all Thanks for this question. I'm still learning a lot about the software myself and I want to give a big thanks to our indefatigable Robot support wizard Pawel P. over in Poland for helping me out with this one.

The first situation with beams only: You are correct, there is an effective doubling of the mass of the structure. Robot has an option in Modal Analysis settings called "Disregard Density". This little parameter, unless checked, asks Robot to calculate the mass of the story based on actual density specified in the material definitions for the volume of material at each floor. If you check this parameter then only the dead load which is already considered as mass in the "Load to Mass Conversion" tab of Analysis Types and the natural frequencies will be much closer to what you are getting with SAP2000. (revised results below)
Here is where this setting lives:
Density.png

In the second situation with the added floors: The settings in the Panel Calculation Method wind up having a very large effect on the structure. Using full rigid stiffening for the panel (“Slab-Rigid diaphragm” model corresponds to full rigid link in all degrees of freedom (3 translations and 3 rotations)) results in increasing the overall stiffness of the structure to the point where the period is far shorter than you would expect even with all the additional mass of the slabs. I've attached an image with settings that will result in much closer periods as calculated by SAP2000 for this situation. As you might expect, understanding these settings is very important and I'm learning a lot too so thanks for asking these great questions.
PanelCalculationMethod.png

After above changes the results are as follows:
Model T1 T2 T3
ROBOT_BEAMONLY 0.57 0.41 0.39
SAP_BEAMONLY 0.59 0.45 0.40

ROBOT_BASICFLOOR 0.61 0.41 0.39
SAP_BASICFLOOR 0.62 0.45 0.43

revised models (courtesy Pawel P) are attached for your review.

I hope that helps you get a better feel for the different settings in Robot that affect these calculations. Please don't hesitate to ask more questions! We're happy to help.

Thanks,
-Ken

(and a big thanks to Pawel P. too!)

[erginc]

Hi Ken,

I knew there is this little parameter somewhere, but don't know where. It is good to understand more about inner workings of the program. Thanks for your time and detailed explanation. And thank to Pawel P. too...

Regards,
Ergin