My model is 3D reinforced concrete beam in ABAQUS software. I knew one reason which may be the development length of longitudinal steel. but i want to know if there are any other reasons because the behavior in nonlinear zone, there is a difference between the experimental test and FE
I think the one more reason you getting the variation may be when the concrete is poured in beam with min 0.5 to 1 m height, the heavy aggregate will settle first at the bottom. So, there would be variation in density assignment to concrete values. You better try with adopting those variations taken into considerations.
Your question is so general that it can take hundreds of (right) answers! If I understand right you have differences between numerical and experimental results. The model is a RC beam under bending or something like that. First let's assume that the experimental results are "right". This means that the experiment was successful and it was recorded successfully so that there is no doubt about the experimental results that you have.
Now let's move to the numerical model. I assume that you are performing non-linear analysis and you are using 3d elements for concrete and linear elements for steel? is that right or you have a simpler model? You need to check the following:
1. Strength of concrete and non-linear constitutive law for concrete. Are you sure that the strength you have set for your material in the FE model matches the strength of the real specimen? Note that nominal values for strength are not accurate. A C20/25 concrete can have real strength values of 35 MPa easily so nominal values are not to trust when you want to validate an experiment. You should do compressive tests to the specimen and find the real strength value (again with variations).
2. strength of steel and non-linear constitutive law. With steel things are better than with concrete, but again you have to make sure that the parameters you set for strength and ductility are the right ones.
3. Bond-slip between steel and concrete. If you push the material to its limits, bond slip will occur. You need a special FE model with contact elements to achieve that otherwise your numerical model cannot follow the experiment after some point.
4. Crack model for concrete. Real concrete will crack when big tensile stresses are applied. If you want to follow the behavior of concrete after the first cracks appear, you will need a numerical model for cracking. Smeared crack models can help in that but they are rather complicated compared to simpler concrete models that do not include cracking.
Point 1 above can be a major parameter. Make sure that you are using the right values for concrete strength, the values that match the real concrete strength of your experiment. Other than the above, there are many other parameters that can affect your results. In order to give you a better answer, you need to describe in detail the experiment and the numerical model, otherwise we need to guess...
Thank you very much , Vagelis Plevris and Hojjat Badnava and Radwa Defalla
- Badges
- Science topic