Cracked/Uncracked stiffness in RSA

[seaedo]

Hi SeismoSoft team,

I'm doing a RSA in a model of a 12-storey RC building (Example from App.B FEMA P-2091). In the Project Settings, I have set a 50% of the cracked by the uncracked stiffness for columns and walls, and a 30% for beams.
After running the analysis, I want to check the total base shear, which I obtained from 'Global Response Parameters' and then 'Forces and Moments at Supports', calling this one BS1. Then, to double-check my values, I check the sum of shear in each vertical element at the base, which I obtained from 'Element Action Effects' and then 'Frame Forces', where I oriented each local axes of each element in the same direction in order to get a cumulative correspondingly, calling this one BS2.
I was expecting that both values should be the same or maybe some small variations but, comparing BS1 with BS2, I obtained BS1 = 2*BS2 exactly, in other words, 50%*BS1 = BS2. Then I tried with 70% instead of 50% for the vertical elements and I got 70%*BS1 = BS2. Also, I tried with 100% (uncracked section) instead of 50% for the vertical elements and I got BS1 = BS2.
Is it correct? should I get the same value for BS1 and BS2? How can/should I calibrate the model with cracked stiffness in order to get the same values in both cases?

Thanks in advance.

[seismosoft]

Hi saeaedo,
We will check things more thoroughly, but at first sight it seems that there is a bug in RSA in the calculation of the element forces, when the cracked stiffness is smaller than 100%. Hence, BS1 is the correct value, whereas BS2 is wrongly multiplied twice with the 50% value of the cracked stiffness.
We will investigate things further in the next days, and we will try to post a fix as soon as possible.
Apologies for troubling you and thanks for reporting this problem.
Seismosoft Support

[seaedo]

Thanks for your reply and I'll be paying attention to your conclusions

[fsapostolou]

If i want to run a pushover analysis , the default choice of Seismobuild is the uncracked stiffness.
The parametres E , I , change measure with each iteration of pushover or remains constant;

[seismosoft]

In pushover analysis there is not a distinction between the cracked and uncracked stiffness, since the stiffness is constantly calculated and updated, depending on the current level of deformation.
Having said that, in the necessary eigenvalue analysis (it is needed for the calculation of the target displacement) the stiffness in the uncracked one in ASCE-41, EC-8, NTC-18 and KANEPE, but the cracked one in TBDY (this is what is required by the different Standards).

Finally note that you should have posted the question on the SeismoBuild part of the forum, and not on the SeismoStruct one.

Seismosoft Support

[fsapostolou]

Thank you , for the very useful information , but i don't see on any KANEPE or EC8 paragraph that "the calculation of the target displacement is with uncracked stiffness

[seismosoft]

This is implied from the calculations to estimate the target displacement, whereby the period from the eigenvalue analysis is employed to calculate the effective period Teff.
Seismosoft Support