Different result of eigen value between ETABS and STKO

[BuuNguyen]

Hi STKO

I try to compare the natural period of building between ETABS and STKO they are different. can you suggest a reason for it?

I also attach etabs file and STKO file.

Thank for your support.

[kesavapraba]

Hi, A cursory glance shows that there could be some errors in assignment of mass in your models. Initial three modes of the ETABS model show large time period - A period of about 20 sec seems to be too large! Can you please share the total mass of the structure from ETABS file? Please also reshare the STKO model after cleaning the unused properties. One more point: a gravity analysis should fall in Static Analysis category, so you can avoid using Transient in your analysis settings.

[BuuNguyen]

Hi Prabakaran Kesavan,

1. The period of the first mode in ETABS is in the X-direction. The first and second periods in Y-direction (4th and 9th in modes in ETABS) should be compared with those of Opensees model.

2. Transient analysis I used in the model with ramp-load (quasi-static) also works as static analysis

Please refer to attachment for new STKO file without unused properties.

Thank you.

[kesavapraba]

Hi, Thank you. Please find my suggestions below.

1. The period of the first mode in ETABS is in the X-direction. The first and second periods in Y-direction (4th and 9th in modes in ETABS) should be compared with those of Opensees model.

By default, OpenSees uses lumped mass matrix calculated from the uniform mass assigned to the line elements. Alternatively one could use consistent mass matrix in OpenSees. You need to check in the -cMass option to use that option in STKO. I did use that option for your model, now the modal frequencies are matched pretty well. Please see the attached figures. You might also check ETABS manual to understand how the mass matrix is calculated.

2. Transient analysis I used in the model with ramp-load (quasi-static) also works as static analysis

Fine, But Static and Transient analyses are fundamentally different. I suggested to use Static analysis to avoid any inertial and damping effects produced by transient analysis.

[BuuNguyen]

Thank you for your kind support,
I have 1 more question:
I want to get the stiffness matrix of the building frame. Is it possible in STKO?

For example in ETABS, I applied unit force for each floor --> get displacement --> get deformation matrix
--> inverse deformation matrix to get stiffness matrix.
But it takes much time to do it manually in STKO.

Thank you,

[kesavapraba]

Hi, Yes, it is possible to obtain tangent stiffness matrix in STKO. You will have to use TCL custom command (printA) for that purpose in the analysis section. Please see your model attached below - I have included it. Please note that the stiffness matrix obtained so is formed from current tangent from the previously converged step in the analysis. The matrix also depends on the type of constraint handler. For more details, you can go through the link below.

https://opensees.berkeley.edu/wiki/inde ... tA_Command

https://portwooddigital.com/2020/05/17/ ... fness-too/

[BuuNguyen]

Hi,
Regarding the natural period using STKO,
1/ I checked ETABS and it also computes mass a lumped mass.
2/ I also compute the deformation matrix (C) and inverse C to get stiffness matrix K as you can see the the attached photo. Then, I compute manually the mass of each floor, using eigen command (eig(K.M)) in Matlab to get eigenvalues, and also get natural periods, which are similar to ETABS.

So, I think with the chosen cmass in STKO, it means lumped mass. Can you check the source code?

Thank you

[kesavapraba]

Hi, I'm not entirely certain about how ETABS computes the mass matrix. However, in the case of OpenSees, the default procedure involves the formation of a lumped mass matrix for elasticBeamColumn. If a consistent mass matrix is required, one can activate the -cMass command in OpenSees to generate that matrix. Additional details can be found in the provided links for your reference.

https://opensees.berkeley.edu/wiki/inde ... mn_Element

https://portwooddigital.com/2023/11/14/ ... vibration/