Infills modelling

[mirko]

Hi,
I'm implementing 2D analysis of an infilled dual wall-frame building.
Infills has been implemented using Crisafulli hysteresis rule available in Seismostruct.

Before undertaking nonlinear analysis I have done a simple Eigenvalue analysis, in order to compare my results with other ones previously gained with a simplified modelling. The simplified model is an elastic model featuring the infills simply as an inclined elastic truss. Obviously mechanical parameters of infills have been assumed coherently for the two different models.

The input parameter, in Seismostruct, characterizing the area of the strut has been chosen equal to the overall area of the equivalent infill strut, as described in the Manual.

After this comparison I have realized that analogous results, regarding eigenvalue analysis, yield only if the area of the strut chosen for Seismostruct input represents the half of the overall area.

After this comparison I have run a Pushover analysis, Response control, and I gained some contradictory outcomes. Through output deformations and output forces, associated to the first step of the PO, I have computed the initial elastic modulus of each strut. What I get is a value close to the half of the one I have assumed in my input.
I've also tried to increase the step of the PO, from default value of 50 to 100, but still getting same outcome.
Analogous comparison has been performed also ffor the pushover analysis.

Resuming, according to eigenvalue results, everything seems fine using half of the equivalent area (parameter "strut Area 1"). According to the pushover analysis only the chosing the overall equivalent strut area as input I get reasonable results.
I'm afraid that I've missed something in my calculations but I've checked several times what I did and everything seems ok.
Could you please help me to solve this problem?

[mirko]

I'm sorry,
one more useful information is missing.
I haven't explained in detail my model since analogous comparison has been undertaken without infills and everythins was ok.

[mirko]

Hi,
I've made a step ahead trying to address eigenvalue issue.
Following analysis have been performed both for a simplified 1bay2storey frame building and with my "complete" model.
Results yielded from Eigenvalue analysis and Pushover analysis would be both consistent with the result gained from calculation result gained from other software only if EigenValue analysis is performed taking into account, just for the initial elastic state of the analysis, both struts in compression and in tension.

I've realized this simply modifying tensile strenght of infills strut as input Crisafulli parameters. First I had assumed zero tensile strenght, and after that I've repeated the analysis with tensile strength equal to the compressive strength.

The outcome, in term of eigenvalue results, is exactly the same, apparently meaning that even strut in tension are contributing for defining the initial stiffness.

The confirmation of this has been given performing pushover analysis.
What I get is that only if the input area of the strut is the total one the elastic modulus of each strut derived from the given first step strut outcome yield the same of the input.

Resuming, my conclusion could be that "factor 2 issue" regarding the input area of Crisafulli parameter is solved treating this input area as the overall one: in this way Pushover analysis are consistent with the result gained from other calculation.

Still Eigenvalue analysis results seem not correct, and turn out to be true only inputing a "fake" area equal to the half of the total one.

Could you please check if my gained assumptions are correct?

Thanks

[Stelios_Antoniou]

Hi Mirko,

I really doubt if your assumption is valid, since the same code is used to calculate the stiffness both in pushover and eigenvalue analysis. Can it be the case that you are somehow misled by the fact that Crsafulli's model employs 4 instead of two struts (refer to the help for more details)

Seismosoft Developer