how to input column nonlinearity like P-M2-M3 interaction

[ankurjain1992]

Sir how can we input column nonlinearity like P-M2-M3 interaction for column for non-linear analysis. ??

For beam, we can assign ZeroLength Element with ModiMKPinching type Materials. But what about columns ? We need P-M interaction for that so how shall we input that data ?

[STKO Team]

Dear user,
In OpenSees the only way to obtain a P-M-M interaction is to use a fiber cross section. If you go with the phenomenological-hinge approach, the zero-length elements that you give for P, My and Mz are un-coupled.

We are working on implementing this kind of interaction for phenomenological-hinges as well in OpenSees. All users will receive a notification (and we will do a webinar on that) as soon as the implementation will be ready (in a month or so, I think)

[ankurjain1992]

Ok if I use fiber section then also there must be some field where I can define the column nonlinearity.... Where can I enter Moment capacity for columns and it's interaction with Axial force P.

[STKO Team]

if I use fiber section then also there must be some field where I can define the column nonlinearity

No, when you use a fiber cross section, you only define the nonlinearity of the uniaxial materials (concrete and steel) that will take part in your fiber section.

Where can I enter Moment capacity for columns and it's interaction with Axial force P

You can't. The fiber cross section is a discretized section that will compute automatically the tensile-resistence, compressive-resistance, and bending moment resistance, and the interaction among them.

Note: in the next STKO release we will provide a tool that the user can use to test a fiber cross section, for example:

• computation of moment curvature response at a gived P value
• computation of the P-My-Mz interaction domain

[ankurjain1992]

Then if i have to do Non-linear Time History Analysis for a RC structure, then i just need to define a fiber section for columns as well as beams.
Won't there be any requirements of ZeroLength Element or 'beamwithHinges' to be assigned at the start and end of beams and columns ?

I have seen one of your example in which you have solved a RC column with Distributed Plasticity approach and provided a zeroLength Element at the Start and end of the column. Was that type you referrring for Non-linear TimeHistory analysis ?

[ankurjain1992]

if I use fiber section then also there must be some field where I can define the column nonlinearity

No, when you use a fiber cross section, you only define the nonlinearity of the uniaxial materials (concrete and steel) that will take part in your fiber section.

Where can I enter Moment capacity for columns and it's interaction with Axial force P

You can't. The fiber cross section is a discretized section that will compute automatically the tensile-resistence, compressive-resistance, and bending moment resistance, and the interaction among them.

Note: in the next STKO release we will provide a tool that the user can use to test a fiber cross section, for example:

• computation of moment curvature response at a gived P value
• computation of the P-My-Mz interaction domain

Then if i have to do Non-linear Time History Analysis for a RC structure, then i just need to define a fiber section for columns as well as beams.
Won't there be any requirements of ZeroLength Element or 'beamwithHinges' to be assigned at the start and end of beams and columns ?

I have seen one of your example in which you have solved a RC column with Distributed Plasticity approach and provided a zeroLength Element at the Start and end of the column. Was that type you referrring for Non-linear TimeHistory analysis ?

[STKO Team]

First, do not confuse the following two aspects:

1. fiber-section versus a phenomenological moment-curvature or moment-rotation laws
2. distributed versus lumped plasticity

Regarding point 1:
You choose the fiber-section when you want P-M-M interaction, and when you have access to concrete and steel material properties. Then the fiber-section will automatically integrate the stress-strain response of every fiber to obtain the moment-curvature reponse.
On the other hand, if you don't care about P-M-M interaction, and you have experimental data about moment-curvature or moment-rotation, then you can use the phenomenological laws.

Regarding point 2:
You can choose the distributed approach, when you either use fiber-section or when you have phenomenological laws given as moment-curvature (not rotation). With the distributed approach, then the nonlinearity will be spread along all the beam.
On the other hand, if you only want to focus on nonlinearity at the beam/column's ends, or if you only have phenomenological laws given as moment-rotation(not curvature), you need the lumped approach.

We talked about both approaches in our 5th and 7th webinars:
#05 Learning Opensees: Nonlinear Modeling of RC Frames (Distributed Plasticity)
#07 Learning Opensees: Nonlinear Modeling of RC Frames (Lumped Plasticity)

[ankurjain1992]

ok thank you so much sir....I will go through those two webinars.