Seismosoft | Earthquake Engineering Solutions

Thank you very much Huffle for your reply.
yes, I have verified all the materials properties with experiment. I am providing same material properties for experimental and numerical analysis.
I have verified that the boundary condition also same as the experiment. However, the experimental results shows the strength is deterioration after 3% drift, whereas for as built column numerical results shows that there is no degradation on strength, and it is becoming uniform after 4% drift. In addition, numerical results shows more load carrying capacity of as built columns compared to experimental results for as built column. As I understand, the experimental and numerical results of as built column should be more or less same. Anybody, could please help.
Thanks

Regarding to the Confinement Factor option for applying FRP jacket, it seems that if the Jacket thickness is greater than 1.2 mm the Confinement Factor will be increased, whereas it can be found in the practical articles and retrofitting methods, thinner sheets have been used as well (e.g. 0.117mm) and their structures have shown appropriate seismic performance after retrofitting by this thickness. Is it correct?

You are correct. Indeed the thickness you are referring to is the Nominal thickness of fibers, for instance like BASF Co. CFRP sheets nominal thickness is 0.0065 in/ply = 0.165 mm/ply. However, the actual cured thickness of a single ply laminate (fibers plus saturating resins) is 0.02 to 0.04 in. (0.6 to 1 mm) for the aforesaid CFRP system.

The use of either nominal fiber thickness or cured thickness has already been discussed in the forum in the link below and the answer that I concluded is to use the cured thickness and got really good results, one may disagree though.

http://www.seismosoft.com/forum/topic.a ... fakharifar

"Being concerned about the thickness of the FRP material in the con_frp model, this is what I found in the literature and assumed it might be helpful for seismostruct users:

"Chen et al. [a] concluded that the utilization of the actual (cured) thickness of the Fiber Reinforced Polymer (FRP) (instead of the nominal thickness provided by the manufacturer) is more appropriate in the Finite Element (FE) analysis."

[a] CHEN, J. F., LI, S. Q., BISBY, L. A., AI, J. “FRP rupture strains in the split-disk test.” Composites Part-B: Engineering, 42(4), 2011, pp. 962-972.

Thanks SeismoStruct for your already right answer to my question, believing the use of the Actual cured thickness of the fibers, impregnated into epoxy resin."

quote:Originally posted by anantparghi

Thank you very much Huffle for your reply.
yes, I have verified all the materials properties with experiment. I am providing same material properties for experimental and numerical analysis.
I have verified that the boundary condition also same as the experiment. However, the experimental results shows the strength is deterioration after 3% drift, whereas for as built column numerical results shows that there is no degradation on strength, and it is becoming uniform after 4% drift. In addition, numerical results shows more load carrying capacity of as built columns compared to experimental results for as built column. As I understand, the experimental and numerical results of as built column should be more or less same. Anybody, could please help.
Thanks


Dear anantparghi,

Regarding your observation, if trying to capture the degrading hysteresis behavior/pinching, bar pullout/slippage etc ... one may not only model a column cross section, fixed at based and expect to accurately simulate the hysteresis behavior matching the experimental models. SStruct has a versatile library of link elements that upon appropriate calibration can be accurately utilized to simulate the bridge pier (similarly any other element) hysteresis behavior, irrespective of the confining material (steel, FRP, SMA). The below discussion on using modified Takeda or Sivaselvan and Reinhorn link elements (multilin or smooth link elements) should help. Besides, even one may use the simple reinforcement buckling/fracture strain to limit the post capping rotational capacity of an element, as this is an important factor to capture deteriorating behavior. Moreover, the performance criteria is really helpful as if you try to model a lap-spliced column and obtain reinforcement strain of 0.06, that is unlikely an unattainable level of strain (unless lap-length is beyond 36~40 time bar diameter), as spliced column has failed due to bar slippage already. Good luck on modelling!

http://www.seismosoft.com/forum/topic.asp?TOPIC_ID=1821

Hi there,
Thank you very much for your help all the way to understand the topic. Still, I have few questions.
In new version of SismoStruct-7, for FRP confined concrete we need to provide the FRP jacket properties, such as 1)Jacket thickness (m):0.000111, b) FRP Elastic modules (GPa): 230, and c) jacket strain: 0.0163.
Once, I click the Help button, there is not help.
Another question is that: How do I know, which FRP materials model I am using. I mean to say Kawashima et al. (2001, Ferracuti and Savoia model etc..
Another question,: can I apply my own FRP confined model in SeismoStruc-7 ?
Please advise me, thank you very much for your help.
Regards,
Anant

Hi Antant,
The FRP properties are used for the calculation of the confinement factor values only. The program calculates the confinement factors based on the properties specified in that window, which are somehow independent from the rest of the project input.
The material used is the one specified for concrete in the Materials and Dimensions page of the Sections Module. Finally, if you want to introduce your own FRP confined model, refer to the following topic:
http://www.seismosoft.com/forum/topic.asp?TOPIC_ID=32

Last but not least, thanks for spotting the bug with the Help button.

Seismosoft Support

Dear SeismoSoft Support Team,
Thank you very much for your help.

Hi,
I am modelling a CFRP retrofitted cantilever circular RC columns with displacement control quasi-static time history load with constant lumped mass at the top of column. The as-built column was model without the CFRP jacket while the other specimen is modeled using one, two, three and four layer of CFPR wrap with the thickness, modulus of elasticity and ultimate strain of 0.11mm, 230 GPa and 1.63%, respectively. With the different thickness of CFRP, the confinement factor was also increased.
I used the infrmDB beam column element, and Link element with Takeda link element for stiffness degradation; and for bond slip model of rebar and concrete I reduced the young modulus of elasticity of longitudinal rebar.
If the confinement factors are different with different thickness of CFRP, then why the load carrying capacity of column is not increased compared to as built column? However, stiffness degradation is significant after 4% drift of CFRP wrapped column. But, no stiffness degradation and load carrying capacity is changed with the increase of CFRP layer. Even, for all the columns the load carrying capacity of columns are more or less same for all CFRP wrapped column.
I would request you could you please help me what is wrong with my column modelling precess ?
Thanks
Anant

Can it be that the response is controlled by the Takeda curve, i.e the maximum load is not reached in the column, but rather in the link element below? What happens if you replace the Takeda urve with a linear elastic one?

Seismosoft Support

Hi there,
Thank you very much for your kind response to my query. As you recommended, I replaced the Takeda curve and used the lin_sym, but there is no any change on load carrying capacity or stiffness, only the load increased 0.58 kN.
I again modeled, without the link element, even there is no any change in load carrying capacity and stiffness, only 1 kN,load is increased.
While experimental results shows the increased in load carrying capacity about 12 kN to 15 kN.
Thanks
Anant