Zig-Zag Strain Pattern in NSM FRP Pull-Out Simulation (Cohesive Zone Model) – Need Help

[Roshaan]

I'm currently working on a numerical simulation of a Near-Surface-Mounted (NSM) FRP reinforcement in concrete using ABAQUS. I'm trying to simulate fiber pull-out behavior using the Cohesive Zone Model (CZM) approach to model the bond between the FRP and the epoxy adhesive.

Model Overview​

• Geometry: A concrete block with a groove made at the top mid-region (centered along the width).
• FRP rod/strip: Placed inside the groove.
• Bonding: The FRP is bonded to the surrounding concrete through an epoxy adhesive layer, where the cohesive elements are used to simulate the bond interface.
• Loading: A displacement-controlled load is applied to the free end of the FRP (pull-out direction).
• Boundary Conditions: The back face of the concrete block is fully fixed in the loading direction.

I’ve also attached a reference image of the model setup.

---

The Issue​

I'm observing a weird zig-zag pattern in the strain distribution along the bonded region of the FRP, especially near the epoxy interface. Instead of a smooth increase or gradient, the strain results are jagged and oscillatory.

What I’ve Tried So Far:​

• Applied Smooth Step Amplitude to the displacement to ensure gradual loading.
• Experimented with different boundary conditions and constraint combinations.
• Refined the mesh near the bonded zone.
• Double-checked material properties and cohesive behavior inputs.
• Tried increasing cohesive element resolution and checking the concrete-epoxy interface behavior.

Yet, the zig-zag strain profile still persists, and it’s affecting the interpretation of bond behavior and pull-out force.

---

What I Need​

I’m looking for serious suggestions or feedback from anyone who:

• Has dealt with strain noise or instability in cohesive simulations.
• Can help identify potential modelling errors or overlooked issues.
• Might know if this is a numerical artifact, and if so, how to regularize or eliminate it.

Is this a mesh sensitivity issue? Do I need to modify the cohesive law (like add viscosity or softening regularization)? Could the load application be too abrupt even with a smooth step?


Any help would be appreciated!

 

[FEA way]

Is that Abaqus/Standard or Explicit ? What if you plot stress vs strain or force vs displacement ? Are the other results correct ? Check the deformed shape and its animation while tracing the plot.

 

[Roshaan]

its abaqus explicit and yes rest of the results are correct the only problem is with strain and that for the frp in the bonded region only. Deformed shape seems reasonable, but i still can't find the reason for this unusual strain pattern.