Kent-Scott-Park (confined fibers response issue)

[felipe_carrasco]

Hi, regarding the Kent-Scott-Park concrete model (KSP), How the software take into account the confinement?. I have tried to use the KSP concrete model but I obtained the same local stress-strain response of a confined and unconfined fiber in a rectangular RC column. This occurs despite increasing the transverse reinforcement in the confined concrete.

According to the KSP concrete model, in the local stress-strain response of a confined concrete fiber, a larger maximum strain in the descending branch should be observed (less steep descending slope with respect to the unconfined concrete), i.e e_50c or e_20c.

Thank you,
Felipe

[seismosoft]

The con_ksp model considers explicitly the confinement factor in the max. compressive stress and in the strain at the max. compressive stress through the following expressions:
Strain_atMaxCompStress = Strain_atMaxCompStress_0*(1+5*(ConfinementFactor-1))
MaxCompStress = MaxCompStress_0*ConfinementFactor
The descending branch to the residual strength is not affected though.
Seismosoft Support

[felipe_carrasco]

Thank you for your answer.

I analyzed my model again, and I am still getting nearly the same stresses when comparing unconfined and confined fibers. I did the same analysis but using Mander concrete model, and in this case I am getting larger peak stress values associated with the confinement in the confined fibers, as expected.
What would be the problem with the Kent-Scott-Park material model?.

Thank you,
Felipe

[seismosoft]

Can it be that you are checking things in thr descending branch? If this is the case, then it is not unreasonable, because the strain at residual stress is not affected by the confinement factor (note that after your comment yesterday, we decided to change this in v2022).
In any case, note that, unless you have specific reasons to use the con_ksp model, you should better employ con_ma, which is far superior (more accurate and more stable)
Seismosoft Support

[felipe_carrasco]

No, in fact I checked the entire stress history response of an unconfined and confined concrete with con_ksp, and the stresses are nearly similar in the whole history. I wanted to use the con_ksp model because I was looking to regularize the concrete material to reduce localization issues (equal concrete crushing energy of elements with different lengths). With the Mander model is no that an easy task, because the input parameters are related to the first ascending branch of stress-strain response only, which are deemed to not affect the localized response.

[seismosoft]

Can it be that the confinement factor is 1.0 or close to 1.0? Can it be that you are checking the global response (whch could be dominated by the longitudinal steel rebars) and not the local parameters?
Check also in the same example the entire stress strain history of adjacent confined and unconfined monitoring points of the same section with the Stess Points (see Ooutput module of the pre-processor).
Seismosoft Support

[felipe_carrasco]

No, I even checked the very same problem with con_ma (Mander) and con_tl (Trilinear) and in those cases I am getting an increase in max compressive stress due to confinement, but when I checked the same concrete fibers modeled with con_ksp, there is no effect in the confinement. I have tried with very simple examples, and I obtained the same behavior. The confinement factors that I am using are > 1.4.
The parameters that I am checking are obtained in the stress and strain output tab, i.e. the local responses.
Try to model any simple example with con_ksp, and I think you will observe this.

[seismosoft]

Can you send one of your models to support@seismosoft.com, so that we check things?
Thanks.
Seismosoft Support

[felipe_carrasco]

Hi, I have just sent you one of my models with the KSP concrete issue.

Thank you,
Felipe

[seismosoft]

Hello Felippe and thanks for the model,
It seems that you are right, the confinement factors are not considered in the con_ksp model in v2021.
We were checking things with v2022, which is still under development and where the bug has been already fixed (the bug will appear fixed, when v2022 will be released in November). Thanks for reporting the bug and sincere apologies for not realising immediately what the problem was.
Seismosoft Support