Hello, My name is Nathan Ruas Alves, I am a civil engineer graduated from the Federal University of Uberlândia and I am currently studying for a master's degree in Civil Engineering with an emphasis on Structures. I would like to ask for your kind collaboration in solving a problem I am facing in my research. I have tried several approaches, but so far, I have not been successful. I am studying smooth concrete slabs reinforced with non-metallic GFRP rebars. The slabs are being simulated by a quarter of the entire slab to facilitate processing. I am using a mesh size of 35 mm, viscosity of 0.0001 and dilatancy angle of 43º. The interaction of the concrete slab with the GFRP rebars is of Embedded region and the Actuator (element considered as rigid in numerical modeling) with the top of the slab where the load is applied was of Constraint type Tie type. For the constitutive model of the concrete I am using the fib Model Code 2010 model for compression and for tension I am using Hordijk (1991). It is worth mentioning that I am using symmetry boundary conditions on the faces of the slabs! The concrete compression and tension damage model that I am using is from the formulation proposed by Yu et al. (2010) which represents a simplified method, as it consists only of subtracting from the unit the ratio between the acting stress and the maximum resistant stress. What I don't understand is why in all the simulations I perform mixing dilatancy or viscosity angles or meshes the stress x displacement curve always looks like this before the peak. How can I reduce this peak to get closer to the experimental one?
Nathan, it sounds like you're experiencing some issues with your simulation results, specifically with the stress-displacement curve exhibiting an unrealistic peak before the actual peak. This could be due to several factors, including the mesh size, material properties, or the constitutive model used. Have you tried refining your mesh or adjusting the material parameters, such as the viscosity or dilatancy angle, to see if that improves the results? Additionally, you may want to consider using a more advanced material model or a different numerical approach to better capture the behavior of the concrete and GFRP rebars.
Manzar Masud, I hope you are well. Considering or not the GFRP reinforcement bars, this unexpected peak occurs. I simulated without any non-metallic or metallic rebar, and the answer was the same. Considering the concrete constitutive model of Mode Code 2010 or Carreira and Chu and considering the damage turned off in the CDP model, this is also happening. When I refine the mesh, the model becomes more rigid! The viscosity at its limit value of 0.00005 approaching zero only decreases the force values up to the maximum value of 100 kN as shown in the figure. Could you help me? I don't know how else to adjust this model, although it is simple, and the boundary conditions and mesh size values or plasticity information, provided by the author of the thesis I am studying, do not match or come close to the results of the processing of my numerical models. The numerical models were made following information from the thesis! Thanks for the feedback!
To address the issue of high stiffness in the elastic stage of finite element (FE) analysis in Abaqus, you might consider the following academic approaches:
1. Mesh Refinement
• Although you’re using a mesh size of 35 mm, the stiffness in the early stages can often be attributed to mesh sensitivity. Try refining the mesh near the regions of interest, such as around the GFRP rebars and loading areas, to improve local stress distributions.
2. Material Properties and Damage Models
• The use of the fib Model Code 2010 and Hordijk (1991) models for concrete is suitable, but ensure the parameters (e.g., tension softening, fracture energy) closely match your experimental data. A slight modification of tensile damage parameters might better replicate experimental stiffness.
• Revisit the damage evolution law in the Yu et al. (2010) model. If it is oversimplified, it may overestimate stiffness in the elastic stage. Incorporating a more gradual damage evolution law based on experimental calibration could help.
3. Viscosity Parameter
• Your viscosity value (0.0001) might be too high for capturing realistic early-stage behavior. Consider reducing it incrementally to avoid artificially stiffening the response.
4. Interaction and Boundary Conditions
• Verify that the Embedded Region constraint for the GFRP rebars accurately represents the bond-slip behavior. Overly rigid interaction may artificially increase stiffness.
• For the actuator-slab interaction (Tie constraint), evaluate whether this overly constrains degrees of freedom. Switching to a contact formulation with penalty methods or small separation might better reflect reality.
5. Dilatancy Angle
• The dilatancy angle (43°) may be contributing to the early stiffness. You can experiment with slightly lower angles while staying within realistic material behavior ranges to assess its impact.
6. Initial Imperfections
• Introduce small geometric imperfections or preloads in the model to account for experimental variances, which may reduce unrealistically high stiffness in simulations.
7. Calibration with Experimental Data
• Ensure all input parameters, particularly fracture energy, concrete tensile strength, and compression parameters, are calibrated precisely to match experimental results. The stress-displacement curve is highly sensitive to these parameters.
Additional Notes
• Early-stage stiffness discrepancies often arise due to idealizations in boundary conditions, interaction definitions, or material models. Conduct sensitivity analyses by varying one parameter at a time to identify the most influential factors.
• If the numerical results consistently deviate from experiments, consider alternative models, such as the Concrete Damaged Plasticity (CDP) model, which might offer more flexibility for calibration.
By systematically addressing these aspects, you can reduce the early-stage stiffness and achieve results closer to your experimental observations.
- Badges
- Science topic
- Similar topics
- Psychology
- Emotion