Topic: Cyclic Modeling of Low Rise Shear Walls in ATENA 2D

Hello,

I am currently matching the experimental results for four low rise walls (aspect ratio between 0.5 and 0.75) which were tested reverse cyclically using the ATENA 2D software.

Concrete: NonlinearCementious2
Steel: Discrete rebar, Multi-linear Cyclic Reinforcement

Currently, the models, when analyzed monotonically, provide accurate predictions for the backbone curves as compared to the cyclic experimental results and reasonable predictions of cracking damage and peak lateral strength. When trying to use the same material properties defined in the monotonic analyses in cyclic models, I am having issues with the models having reduced damaged stiffness and they are unable to run to the drifts reached during the experiments. From my analysis of the failure of the models, it seems that the first row of concrete elements has significantly localized damage from the cycling, resulting in lower minimum principle compression stresses as seen in the monotonic models, resulting in this reduced damaged stiffness and premature specimen failure.

Are there any concrete material parameters that would help with the cyclic behavior of this concrete material model?
Would using the 3D analyses in ATENA Science solve any of these issues, or would it be better to get improved behavior in the 2D analyses?
Would incorporating an interface help this behavior? In the tests, the interface was a cold joint, which had observed gap opening early in the test and low slip until peak load. Is there documentation on properly defining the cohesion, tensile strength, and coefficient of friction, for a model like this, where gap opening is important to the behavior and slip should not reduce the peak load behavior?

Thanks for any assistance, I can also provide more detail on my concrete and rebar models and variations I have tried to improve the cyclic behavior.

Thanks!

Re: Cyclic Modeling of Low Rise Shear Walls in ATENA 2D

Dear rdevine,
my first guess is that you probably have a problem representing the boundary conditions corresponding to the loading history, especially while switching load direction. It is a frequent error that tensile load is introduced at some part of the load cycle.
Please see ATENA Troubleshooting, 3.1.1 How to apply cyclic load? and 2.2.15 How to model a load combination, e.g., a constant vertical
force during horizontal displacement applied until failure?

Typically, to avoid these issues with tension at direction-switching, cyclic experiments are set in such a way that there is a frame, loaded at a single point, and bringing the load to both sides of the specimen. The best solution is typically to include this frame in your model (elastic, simplified geometry).

If you still can not identify the problem source, you can follow Troubleshooting, 2.1.1 to send us your model etc.

Regards.

Re: Cyclic Modeling of Low Rise Shear Walls in ATENA 2D

dpryl,

Thanks for these suggestions. For the cyclic application, I have followed the steps outlined in 3.1.1 and ensured that I take into account that the displacement increments are properly applying and removing the displacements.

As for a modeled frame, I currently have elastic plates on the specimen for displacement application, but the displacement increments are applied to each side of the model during the positive and negative cycles, so the model is only ever "pushed". I have tried moving both sides of the wall at once as well, as you mentioned to avoid inducing tension when switching loading directions, though this as little effect on the monotonic results and has not solved the cyclic issues.

From my analysis, it seems that the material model as I have it defined, especially for the specimens with higher concrete strength, is not developing the diagonal compression strut up from the leading toe of the wall.

I will look into sharing the models that demonstrate my point per Troubleshooting 2.1.1., unless you have other suggestions.