Menu

Discussion Forum

New Discussion
Filtra

Modeling Issues in Post-Tensioned Prestressed Concrete Beam with Bond-Slip Interactions

Posted 25 feb 2025, 10:12 GMT-5 Materials, Studies & Solvers, Structural Mechanics Version 6.3 2 Replies

Zhou Siiii
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Model Setup

I built a 1/4 geometric model of a post-tensioned prestressed concrete beam for a four-point bending test.

Key details

Prestressing tendons differ from ordinary rebars: Their ends are anchored to the beam’s end faces via anchorage devices. To simulate this, I applied a "Pinned" boundary condition in the Truss interface.

Concrete models

Study 1: Ottosen failure criterion

Study 2: Damage-plasticity coupled model

Bond-slip interactions

(defined inEmbedded Reinforcement under Multiphysics Coupling)

Spring connection between ordinary rebars and concrete (setting the bond slip model);

Spring connection between prestressing tendons and concrete (setting the bond slip model);

Rigid connection between prestressing tendons and concrete.

Results & Issues

Study 1 (Ottosen model)

Bond-slip for rebars + rigid connection for tendon and concrete (ignoring "Pinned" condition)

Result: Load-deflection curve (yellow) aligns with test data (blue) in overall trend and ultimate deflection but overestimates ultimate load.

However, it's better to include both bond-slip model and "Pinned" ends for tendon, but this causes computational nonconvergence.

Figure: http://easylink.cc/qcg2c5

Study 2 (Coupled damage-plasticity model)

Same bond conditions as Study 1

Result: Premature non-convergence; green load-deflection curve shows unrealistically small deflections. Theoretically, this should better match tests than Study 1.

Questions

(1) Constraint Conflict: Is it impossible to simultaneously apply bond-slip interaction and "Pinned" boundary condition to prestressing tendons? Do these settings conflict fundamentally?

(2) Coupled damage-plasticity Model Instability: Why does the damage-plasticity model converge poorly and produce unrealistic results compared to the Ottosen model?

(3) Segmented Tendon Modeling: Goal: Divide the prestressing tendon into three segments:

Segment A: Rigid connection to concrete

Segment B: No surrounding concrete (free/unbonded)

Segment C: Bond-slip interaction

Implementation Options:

-Split tendon geometry into segments and assign different Embedded Reinforcement settings (rigid/spring connection)

-Use parameter modulation (e.g., set parameter c0 and c_h for bond-slip in Emmbedded Reinforcement to 0 in Segment B via position-dependent variables) Which method is more stable and physically consistent?

Reply
2 Replies Last Post 1 mar 2025, 15:39 GMT-5
Tobias Gasch COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 4 days ago 27 feb 2025, 10:38 GMT-5

Question 1 There are no such restrictions. However, note that the Pinned condition in the truss is not consistent with the Rigid Connector in the solid to which the truss is connected. You could try to connect the end point of the tendon to the rigid boundary of the solid using a General Extrusion coupling instead.

Question 2 The Coupled damage-plasticity model includes more aspects of the mechanical properties of concrete compared to the Ottosen model, for example softening in tension. Inevitably this means that the model introduce more nonlinearities and is thus more sensitive to an under resolved mesh and mesh elements with poor quality. When a model includes damage it is also recommended to use a linear discretization for the displacement field.

My suggestion is to :

  • Improve the mesh, see attached image for the mesh I used.
  • Use a linear displacement field
  • Specify the tension softening part of the model using fracture energy or a displacement limit to reduce mesh bias.
  • Allow a smaller minimum step size by modifying the continuation settings of the Parametric node in the solver sequence.

With these changes I can solve study 2 and get a ultimate load of approximately 225 kN compared to 260 kN from study 1.

Question 3 I would add one Embedded Reinforcement coupling for each segment with the following settings:

  • Segment A, use a rigid connection
  • Segment B, use a spring connection with zero axial stiffness
  • Segment C, use a spring connection with the bond-slip model

If you have more questions about the model, I recommend that you contact support.

**Question 1** There are no such restrictions. However, note that the Pinned condition in the truss is not consistent with the Rigid Connector in the solid to which the truss is connected. You could try to connect the end point of the tendon to the rigid boundary of the solid using a General Extrusion coupling instead. **Question 2** The Coupled damage-plasticity model includes more aspects of the mechanical properties of concrete compared to the Ottosen model, for example softening in tension. Inevitably this means that the model introduce more nonlinearities and is thus more sensitive to an under resolved mesh and mesh elements with poor quality. When a model includes damage it is also recommended to use a linear discretization for the displacement field. My suggestion is to : * Improve the mesh, see attached image for the mesh I used. * Use a linear displacement field * Specify the tension softening part of the model using fracture energy or a displacement limit to reduce mesh bias. * Allow a smaller minimum step size by modifying the continuation settings of the *Parametric* node in the solver sequence. With these changes I can solve study 2 and get a ultimate load of approximately 225 kN compared to 260 kN from study 1. **Question 3** I would add one Embedded Reinforcement coupling for each segment with the following settings: * Segment A, use a rigid connection * Segment B, use a spring connection with zero axial stiffness * Segment C, use a spring connection with the bond-slip model If you have more questions about the model, I recommend that you contact support.

Attachments:

Said Bouta
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 days ago 1 mar 2025, 15:39 GMT-5

How did you mesh this mesh with the tendons? Also there is no symmetry with these tendons!.

How did you mesh this mesh with the tendons? Also there is no symmetry with these tendons!.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content