concrete-reinforcement bond

[roshen.rj]

In seismostruct fiber-based modeling, I understand that inelastic DB and FB elements consist of a number of fibers (meshes). However, I wanted to ask how is the concrete-reinforcement bond modeled in RC section? Moreover, in the meshing approach by the software, the mesh is not regular and the finite elements are not coordinated (have no common nodes). Could you clarify these points?

[z.gronti]

Dear roshen.rj,

As i replied to your email:

The model assumes full bond between the concrete and the rebars in the analyses. However, the chord rotation checks are carried out employing the advanced parameters for lap splices (code-based checks module).

I am not sure what you mean with `the finite elements are not coordinated (have no common nodes)`. The fact that the monitoring points are triangular and not rectangular is a modelling choice that (i) does not play a significant role in the obtained results - for a mesh of 100-120 m.p.s or more the differences are negligible (ii) for certain numerical reasons (that cannot be explained in a short message) the meshing with triangulation provides much stables solutions.

[roshen.rj]

Dear Zoi

Thank you for your response. Could you please share the references that could give some more discussion to :

(i) concrete-rebar bond in seismostruct analysis
(ii) meshing with triangulation monitoring points provides much stable solutions in seismostruct analysis

Also, could you clarify what you mean by 100-120 m.p.s?

[seismosoft]

Hi roshen.rj,

There is no reference on the bond between concrete and steel (only the the Bernoulli-Euler principle that plane sections remain plane, I would guess).
Regarding the meshing, this is not published, we have observed it and it is attributed to the fact that in 2D analysis (or loading in one direction) there are not lines of monitoring points that yield/lose strength simultaneously. Finally, m.p. is a monitoring point
Seismosoft Support

[roshen.rj]

Dear roshen.rj,

As i replied to your email:

The model assumes full bond between the concrete and the rebars in the analyses. However, the chord rotation checks are carried out employing the advanced parameters for lap splices (code-based checks module).

I am not sure what you mean with `the finite elements are not coordinated (have no common nodes)`. The fact that the monitoring points are triangular and not rectangular is a modelling choice that (i) does not play a significant role in the obtained results - for a mesh of 100-120 m.p.s or more the differences are negligible (ii) for certain numerical reasons (that cannot be explained in a short message) the meshing with triangulation provides much stables solutions.

Dear Zoe

Thank you for your response. For the second point, I meant the following:

ii) For example at the edge between the 'unconfined' and 'confined' concrete fiber sections there are some triangular fibers that do not share common nodes like the ones represented in the attached figure. So does Seismostruct consider these edges to be constrained? How does the section act as a whole if at certain nodes such as these the nodes are not common for the 'confined' and 'unconfined' concrete sections?

link to the figure: https://i.imgur.com/Wl75NeS.jpg

[seismosoft]

The meshing of the section to monitoring points is a completely different procedure from the discretization of a structure to create the FE mesh. There are no nodes. Each triangle in the section is represented by the mechanical quantities (strain and stress) at its centre of gravity. Strains are calculated from the axial strain and curvature of the section, and stresses from the hysteretic laws of the material that corresponds to each m.p. Please refer to the fundamentals of fiber modelling.
Seismosoft Support