The unloading stiffness of the test hysteretic curve of reinforced concrete column is always higher than that of the sim

[jiayanghnu]

Hello, the following is the hysteresis curve of RC column simulated by STKO and the hysteresis curve tested(The solid lines are tests and the dashed lines are simulations). The unloading stiffness of the simulated hysteresis curve is always lower than that of the test. Why is this? The concrete I used is an ASDConcrete3D constitutive , and the rebar is a hysteretic model. Can you give me some advice?

[kesavapraba]

Hi, the unloading stiffness is mainly related to the damage variables of the ASDConcrete3D material model. The level of plasticity and damage expected in the material can adjusted using the two parameters. I suggest you to use Concrete(6P) and play with the PScale Tension and PScale Compression parameters. You can also try DamageTC3D material model. You can share the model here if the issue still persists, we will take a look at it.

[STKO Team]

You can also try with the pinching parameters of your hysteretic material to make it match the curve of a steel02:

pinch.png (80.58 KiB) Viewed 3170 times

[jiayanghnu]

Thank you for your reply! I have tried the above methods but they still don't work. Could you please help me take a look at my model?

[jiayanghnu]

Hi, the unloading stiffness is mainly related to the damage variables of the ASDConcrete3D material model. The level of plasticity and damage expected in the material can adjusted using the two parameters. I suggest you to use Concrete(6P) and play with the PScale Tension and PScale Compression parameters. You can also try DamageTC3D material model. You can share the model here if the issue still persists, we will take a look at it.

Thank you for your reply! I have tried the above methods but they still don't work. Could you please help me take a look at my model?

[kesavapraba]

Ok, It seems that modeling of some physical behaviour is missing in your model. For example, it is not clear why varying vertical load is applied on the column's top - which can lead to strain penetration. If you don't model strain penetration, there can also be loss of accuracy in matching the curves. Similarly there can be other aspects. For me, your numerical curve seems to be fine. Getting perfect hysteresis curve is very difficult to achieve in cyclic response - a good match depends on many details.