Dear Dpryl:
        Can you give me you E-mail address? Then it will be convenient to send the graph of my test result and material properties to you. Thank you very much.

Dear yuanfang, can you send us a graph comparing the Load-Displacement graph calculated with NLCem2 w/increased Gf to the measured data?
And could you also include a sketch of the setup, emphasizing why you think 2D modelling does not make sense?

I think the next step is to add model the fibres as smeared reinforcement (in 3D, you should define it in 7 directions - the 3 axes + 4 pointing in the middle of the octants; in 2D, 3 should be enough - one axial direction + in a 60 and 120 degrees angle from it). This modelling is of course not good in capturing the fibre bond influence, but dedfinitely should give you the general response including the strain hardening.

Dear Dpryl:

I have tried to use the normal NLCem2 material with Hordijk/exponential softening with increased fracture energy (100x compared to normal concrete) to model ductile fiber reinforced cementitious composites, there is still substantial different between the two models. I think NLCem2User or NLCem2SHCC is more suitable for defining FRC, even the result is good enough with a selected fracture energy, because the normal NLCem2 is to define concrete especially.

So i have to define all the functions for the NLCem2User material. I need your help  with the functions and parameters to define. I will send you the stress-strain relationship of the material if nessessary, or you can tell me the way to determine the parameters of εf, and characteristic size lch. Thank you very much!

Dear yuanfang,

1. Even if you think the plain NLCem2 with increased Gf can not be enough, I recommend to give it a try to see how big is the diiference.

2. The choice of tension softening laws is quite wide in the SBETA material, and "SFRC" and "SFRC local strain" have been developed specifically for FRC. If you can work in 2D, SBETA offers the most straight forward way.

3. If you have tried 1. or 2. and the results are not good enough (after having checked the boundary conditions and material modelling by a colleague or by us), you have to define all the functions for the NLCem2User material. We are ready to help you with the functions and parameters if you come to that point.

Dear Dpryl:
Thank you for your reply, I want to say that the fiber reinforced cementitious material i model has the properties of strain hardening and multiple cracking, so i think it may be not suitable to use NLCENT2 or SEBATE to model it, if i use NLCENT2User, in the tension and compression function, εf and εpl must be input, are they refer to fracturing strain and plastic strain? Can i input the stress-stain curve directely which i obtain from the test? The characteristic size lch is also hard to determine, can you give me some suggestions? Thank you very much.

The NLCem2SHCC material model is only available through the ATENA-GiD interface (i.e., ATENA Science, not ATENA Engineering) or by editing the .inp file manually. However, this material is only a small modification of NLCem2User for some special FRC types. The only difference compared to NLCem2User is generating the shear response functions for some special kind of fibres.
For "normal" FRC, the NLCem2User material is recommended in ATENA 3D, and NLCem2User or SBETA (where you can select the corresponding softening law).

In many cases, it is enough to use the normal NLCem2 or SBETA material with Hordijk/exponential softening with increased fracture energy (e.g., 100x or 1000x compared to normal concrete).