steel brace with intentional eccentricity (BIE) - verification problems

[dfd]

Dear Seismosoft team,

I've been trying for a long time to verify a research study on steel brace with intentional eccentricity and I really haven't been able to do it.

It's a simple model of a steel brace with a cross-section CHS114.3x3.5 subjected to cyclic loading.
Steel: fy=328 MPa & fu=490 Mpa

According to the paper the yield for the specimen 'G2-InP-60' happens at a lateral load of 300 kN and a drift of 3% (5 cm) according to figure 7-case (e), whereas in my case the brace is yielding at a lateral load of 195 kN of a drift > 4%.

The eccentricity of 6 cm was modeled through a element (steel plate 250x100), according to the paper. When i choose constraints instead the hysteretic curves (Node 4) don't seem reasonable.

I am trying quite a long time to find a solution using different steel models e.g. bl, Ramberg-Osgood or changing the parameters , but unfortunately i haven't found a solution.... but I don't know what else to do...

I would be really grateful if you could help me.

I attach the file below.

https://ln5.sync.com/dl/4e0cd31f0#r4kuc ... i-9wwybfp4

kind regards
Dimitris Douventzidis

[seismosoft]

Dimitris hi,

Unfortunately, we do not have the resources to work in detail on your model and duplicate the data from the paper.
Some comments and some tips you could work on:
1) The time step of 0.001 seems extremely small. Carry out a sensitivity analysis and see if it is really needed. You could employ more than one time history stages with gradually decreasing time step, if this is required.
2) You might want to increase the fracture strain of the steel material model
3) the strain hardening parameter also seems to change the results considerably.
4) Make sure that rotations are free in both ends, as in your model.

Seismosoft Support

[dfd]

Hi,

I have indeed increased and experimented with these parameters for more than two weeks and noticed that affect the results. I achieved a yield to 285 kN, but at a high drift 6%..

One last question. The Fy parameter refers to yield and not to ultimate strength? If i set this to fu=490 MPa i'm getting a gysterisi curve like this one from the Paper.

Anyway, thank you very much for your help!
Dimitris

[dfd]

If someone faced similar situation and problems, any help and advices would be appreciated.
Dimitris

[z.gronti]

Hi,

I have indeed increased and experimented with these parameters for more than two weeks and noticed that affect the results. I achieved a yield to 285 kN, but at a high drift 6%..

One last question. The Fy parameter refers to yield and not to ultimate strength? If i set this to fu=490 MPa i'm getting a gysterisi curve like this one from the Paper.

Anyway, thank you very much for your help!
Dimitris

Dear dfd,

The fy is the stress at yield. You may find the description of the required values depending on the material model you use in SeismoStruct's Help System:
https://help.seismosoft.com/seismostruc ... /index.htm

[dfd]

Dear Zoi Gronti,


By modifying the parameters of the material, setting the specific yield stress (fy=328,6 MPa) and choosing a strain hardening parameter high enough at 0.04 the curve matches that of the experiment, but continue to load over 300kN, where the fracture occurs according to the paper at a drift of about 3%.

Isn't there an automatic cross-section check according to Eurocode EN1993 in Seismostruct?

Should this be explicitly defined in the performance criteria field to determine when the fracture occurs?

And the specific stress-strain criterion is determined when the developing stress exceeds the ultimate fu=490 MPa ?

So it is defined based on the ratio between the ultimate stress fu and the modulus of elasticity of the material E=2.1 x10^5 MPa:
criterion stress-strain = fu/E=490 MPa / 2.1x10^5 MPa = 0.0023

Is this the right way of checking the cross-section or have I not understood correctly?

kind regards
Dimitris

[dfd]

I have to point out that the strain hardening parameter of 0.04 is out of the range from 0.005 to 0.15 which are recommended.

[dfd]

I have created some charts in excel if you'd like to take a look.

https://ln5.sync.com/dl/c125aea50/4tpc9 ... d-x748jkf4

Kind regards
Dimitris

[seismosoft]

Dimitris hi,

As we have already explained, given that we have thousands of academic users, we really do not have the resources to work in detail on the models of our users, especially with complicated issues as is the modelling of specimens tested on the lab.

You can find in the User’s Manual a description of all the performance criteria and the code base checks that are available in the program. Regarding all the other questions, regarding the duplication of the test results, these are questions that are probably better discussed with your supervisor.
Seismosoft Support