• Registered: 2010-11-21
  • Posts: 2

Topic: Concrete in tension and compression

Hello,
I am currently working on my thesis and I am using your programm to simulate shear walls. I am using the model SBETA to simulate concrete. In this model I need to define the σ-ε diagramm.While reading the ATENA theory I found out that the ultimate strain in both tension and compression is defined through a critical displacement  and the maximum strain of concrete changes in some load steps.
As far as compression is concerned i tried to use the values for ε provided in the Eurocode 2 for the diagramm in figure 3.2. In addition i made the assumption  of a uniaxial stress test so that the failure band has always at least theoretically a steady angle, which for compression is 0  degrees. Then I used the formula 2.17 from ATENA theory in order to find the length of the failure band. In my simulations I used  a FEM size 0.03m , only quadrilateral elements and  all the elements had the same rectangular shape and side dimensions. As a result I calculated L=0.03m. Then I tried through Eurocode values and L to define the w that is set to 0.5mm by default. However the results were much lower than the default value(actually 10 times less) and I can provide them to you if you need them. First of all were my assumptions and the whole calculation process correct?How could I calculate the proper w for my simulation?
In addition while trying to find the maximum strain in tension through formula 2.18 from ATENA theory I found out that it was about 4e-3, which I think is a large value.
Finally I would like to ask you why in this model the default strain values at maximum compressive strength are so low(less than 2e-3).Is it lower in order to provide better results or because of the diagramm used for concrete?
I would like to thank you in advance.
Dimitris

Quote Post 1

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Concrete in tension and compression

Hello Dimitris,
if you need to define your own working diagram for the material, you should probably look at the Nonlinear Cementitious 2 User material instead of SBETA. In the NLCem2User material, you define multilinear laws for tension, compression, etc. However, in most cases, this is not needed at all, and moreover, as long as you do not have made a lot of material tests to measure all the material data for the user functions, it does not make any sense anyway.

In general, it is recommended to use the values generated by ATENA or read from the material library unless you have measured them.
For nonlinear analysis, the most common approach is to use the mean values based on the concrete strength class.

Using critical displacement and not strain is a way to reduce mesh sensitivity.
I do not understand what you mean by "and the maximum strain of concrete changes in some load steps"?

Crack/crush band calculation, it seems you may mix the input values and the internally calculated "corrected" values used in each element?

I am not sure abouit SBETA, but in NLCem2, the default plastic strain at the compressive strength is 1/2 of the total strain at the point, i.e., plastic strain = elastic strain at max. compressive stress, which is based on some experimental data. On the other hand, in some codes, constant values for total strain at max. compressive stress are proposed. If you wish to use that, you have to calculate the corresponding plastic part and set the eps_cp value accordingly.

dpryl's Website

Quote Post 2

  • Registered: 2010-11-21
  • Posts: 2

Re: Concrete in tension and compression

Hello,
First of all thank you for your answer. Well i have to use the SBETA model and right now in order to finih my thesis I need the stress strain diagramms.
You told me you are using multilinear laws for tension and compression but which diagramms of the EC2 for compression are you trying to approximate (billinear,parabel-rectangle, the real stress-strain relation like Hognestad defined it) as there are different values provided for each diagramm.
In addition I just wanted to know the way you are defining each time the length of the failure crack/crash bands because in ATENA theory it is said that it depends on the angle between the element sides and the crack. As a result we are never sure about the εcu but only about the w. That is also the reason, why I tried to define them through the experiment of uniaxial compression/tension test. I took the values for strains and stresses for the real stress-strain relation according to EC2 and I calculated the w. As a result I asked you if it is possible to calculate the w through this experiment and then use it as a default value for the simulations. My question  is that having increased the values of stresses and  strains in some models in order to take into account the confinement phenomenon(according to Kappos model formulas)  do I need to increase w as well?
Thank you for your help

Quote Post 3

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: Concrete in tension and compression

Hello,

1. it is not clear if you wish to
a) define a material through prescribing a stress-strain law
or
b) create stress-strain plots for some points in the analysis
?

2. no, the material law are only defined by mulitilinear approximations in case of the User functions (NLCem2User material).
If the explanation in sections 2.1.2.4 Compression before Peak Stress and 2.1.2.5 Compression after Peak Stress of the Theory Manual is not enough, I have to ask my colleagues who have developed the SBETA material model - please be as specific as possible.

3. I still do not understand what you are doing/wishing to do.
If you have a direct tensile test, you are typically measuring the average strain over some relatively large distance (the shorter the distance, the larger strain you are going to get after a single crack localizes).
Do you wish to determine the critical displacement w_d from a uniaxial experiment? Why do you think the default value is not good enough for your case?
Or do you only wish to compare measured experimental results to your analysis? - then, displaying the plastic/fracturing/principal strains instead of crack openings may be a way.

4. confinement is already considered in the material model - please see section 2.1.5.1 Compressive Failure for SBETA material, or the corresponding sections for the Cementitious2 and Cementitious3 materials. The Cementitous3 material is based on the model developed by Papanikolaou and Kappos. If you need that confinement model, I suggest to use the NLCem3 material instead of trying to modify SBETA.

dpryl's Website

Quote Post 4