[SUPER URGENT] Multiscale designer Discontinuous Fiber bond slip, pull out modeling issues

I really need some help knowing if the multiscale designer is capable of steel fibers in UHPC. Can bond-slip, pull out behavior be modeled in MDS??

I am attempting to model UHPC using Altair Multiscale Designer (MDS) with discontinuous short steel fibers. Since the tensile behavior of UHPC is strongly governed by fiber–matrix interfacial mechanisms such as bond–slip, debonding, and fiber pull-out, I would like to ask whether MDS provides any way to explicitly or implicitly represent these interface effects. In particular, can bond–slip or pull-out behavior be modeled through any built-in material laws, equivalent interface formulations, or recommended calibration strategies, or are discontinuous fibers in MDS always treated as perfectly bonded continuum inclusions that share load from the elastic stage? Clarification on whether fiber–matrix interfacial behavior can be physically represented within the current MDS framework, or whether this should be considered a fundamental modeling limitation, would be greatly appreciated. If it is impossible, then isn't creating uhpc in mds not appropriate?

Also I could use some help in selecting the nonlinear law for UHPC concrete matrix

Find more posts tagged with

Multiscale Designer

Accepted answers

All comments

altair_fukuoka

Hello,

MultiScale Designer only consider material non-linearity but does not consider interface/contact non-linearity.

What you can do with MultiScale Designer is to make an interface mesh and to apply an non-linear material on it to simulate the interface effect.

I myself do not have experience to validate those materials. Thus, I am unsure MultiScale Designer can satisfy your requirement. I hope you to try. If you want consulting/POC or some kind of special work by Altair, could you please ask your local Altair office or the account manager of your organization.

fig. example,

Regards,

Jay30

Thanks for your reply!

I found an article using MDS for UHPC, but they used the Iso damage and plasticity parameter for concrete matrix. I'm trying to use it too, but it seems I can not describe the compression strength and tensile strength differently. It is crucial for UHPC because compression strength is about 130~150MPa, and the tensile strength is about 8~15MPa.

the paper is

Computational analyses of flexural behavior for ultra-high performance fiber reinforced concrete bridge decks

January 2020

International Journal for Multiscale Computational Engineering18(4)

DOI: 10.1615/IntJMultCompEng.2020035500

Could I get some help on how the iso damage and plasticity non linear works? The manual too is quite hard to understand

altair_fukuoka

Hi,

To check ant test material models, single scale model with user-defined material is useful.

Iso damage bilinear and Iso damage 3-pieces have different tensile/compression strength.

On the other hand, Plasticity model (The stress-strain curve in yield region is a curve but not straight lines like damage models above) does not have different tensile/compression plasticity parameters.

Therefore, could you use Iso damage bilinear or Iso damage 3-pieces for your purpose?

Regards,

Jay30

Thank you for your reply!

Yeah I first use the iso damage 3piecewise for concrete matrix and rate independent plasticity for metal fibers. Under tensile loading, steel fibers in UHPC mainly act to control cracking rather than to strengthen the concrete matrix itself. Once a crack forms in the matrix, the fibers bridge the crack and continue to carry tensile load across it. As the crack opens, the fibers gradually debond and pull out from the concrete, which slows down crack opening and prevents sudden failure. Because of this mechanism, UHPC does not fail abruptly in tension but shows a more gradual and stable response compared to plain concrete. So for example, if the matrix tensile strength is 8MPa, 2% steel fibers can increase the tensile strength up to about 11MPa. It doesn't improve dramatically, but changes the concrete behavior to a more ductile type.

However, since the concrete matrix itself is very brittle, If I use iso damage 3piece wise, the homogenized material's tensile strength only increases about 0.5MPa even in 2% fiber volume.

So i'm having concerns that mds might not be the best way to homogenize UHPC for FEA.

altair_fukuoka

Maybe, just inputting the right material property to concrete and steel cannot represent well, because MultiScale Designer cannot consider geometrical effect like crack, bridge, debond, pull-out as I mentioned in my first reply.

If you knew the homogenized material behaviour, you can tune the material parameters in MultiScale Designer to fit the behavior. However, if you've already known the homogenized material behaviour, you don't need to predict it. Thus, this is a chicken-and-egg problem.

Sorry. I do not have direct anser for your problem.

Jay30

I always thank for your response!

I'm still trying a little bit more into creating UHPC material data. I am currently using the iso damage 3 piecewise law the concrete matrix. I have found that the fibers have almost no effect when the matrix is defined as too brittle.

I am trying to use Gf( fracture energy) Normally 0.1kJ/m^2 ~0.5kJ/m^2 for concrete. Gf represents the area under the softening curve after peak stress. Like the image below.

But since these Gf is based on crack width rather than strain, I am unsure how to correctly define the tensile strain to failure parameter.

I would like to confirm if Multiscale Designer automatically accounts for the characteristic length and mesh size when it generates the material cards for FEM, or if I need to manually calculate the strain by dividing a specific crack width by my intended mesh size. (crack width w = e (Strain) X L (characteristic Length)).

I would like to know if my manual conversion is a valid approach or if the software handles this regularization internally to ensure objective results regardless of the element size.

Always thanks for your response!

altair_fukuoka

Hi,

I am not sure about your question.

But anyway, the answer is no.

At least in MultiScale Designer, any material properties are independent to its unit cell size. Any material properties are for the relationship between strain and stress. Any relationship between length and stress are not exist.

For reference, you can check Radioss concrete materials. However, you cannot find any material which has parameters releted to length,

https://help.altair.com/hwsolvers/rad/topics/solvers/rad/concrete_materials_starter_r_2.htm

I don't think that is the material phenomina. That might be a phonomina on the test coupon which is made by that material.

Thanks,

Nobuyuki

Jay30

Thanks again. !

I looked through all the multiscale designer examples, and also requested any more examples through my altair local altair office but i'm currently not getting replys due to holiday seasons. It would be really appreciated to know if there are any concrete examples or demo files for MultiScale designer. It doesn't have to be UHPC, just concrete would be really helpful.

Always thank you for your response

altair_fukuoka

Sorry, I only know two examples in

<HyperWorks 2026 install dir>\hwsolvers\MultiscaleDesigner\win64\MDB\multiscale\