Concrete modelling to predict crushing in compression

gemada

Hi Victor,
I've found a method to model reinforced concrete using a combination of Calculix and Mohr Coulomb stress.
According to the author, for concrete material it is necessary:
- Young’s modulus (used in the Calculix analysis to calculate elastic deformations and stresses)
- Poisson ratio (same)
- Uniaxial compressive strength (used during post-processing to calculate the Mohr Coulomb stress as an indicator for crushing or shear failure in concrete)
- Friction angle (same)

For steel: only yield strength

Three types of analysis are performed: 1) An elastic analysis using Calculix (only utilising the elastic parameters for concrete); 2) A post processing step to analyse the required reinforcement (only utilising the yield strength of steel) and 3) Calculation of the Mohr Coulomb stress (only using the strength parameters of concrete, i.e. uniaxial compressive strength and friction angle).

Is it possible to do this in Mecway?
Thanks

mmartin

I found that method implemented in FreeCad FEM with calculix inbuilt.
During the next week I will test the Mohr Coulomb formula for a reinforced concrete( compression only material)

gemada

Thanks for the reply!
Did you manage to test it in Mecway?

mmartin

I wrote MC formula to test a reinforced concrete beam failure. I modelled the concrete Using compression only material and the reinforce modeled with steel frames.
I Also reviewed the FC Fem source code. Very clarifing

Victor

Though I don't know much about Mohr-Coloumb, it looks like you could use the Solution -> New formula feature in the solution in Mecway to achieve this. If you find the formula for M-C equivalent stress in terms of principal stresses and those two parameters, you should be able to simply enter it and get the result evaluated for each node in the mesh.

The same goes for step 2, if you just need to calculate a function of stress and yield strength that's evaluated at each node, you can write a formula for it.

gemada

Thanks for your great tips Victor.
I'll try to do it but it may take a while because I'm no expert in FEM...yet...hehehe.

mmartin

I enclose you an excel file with the MOHR-COULOMB failure criteum for concrete.
I added a cell with the MECWAY formula you have to "copy and paste"(remember, the period is the decimal separator).
MANUEL

runinwind

Thanks mmartin for your concrete formula!

mmartin

RUNINWIND:
You welcome. I hope you find it useful.
I will test it in a clay/brick vault model in MW.

mmartin

Jalal

Hi Gemada, Mmartin and Victor,
Thanks for the method and formula.
I understand the first step, using only concrete properties and also understand calculating the M-C stress in each node using the formula. However, I am unclear on second step and inclusion of steel, do you actually model steel? and if so, how?
or you add a wide layer of concrete to model steel (like what we do in analytical methods in text books)?
I am a beginner in concrete modelling, but I have seen people talking about a separate layer of shell elements for steel (composite), or a layer of solid steel elements with contact properties, or steel as link elements node to node for steel.
I would really appreciate your help.
Jalal.

mmartin

Good afternoon JALAL:

I am also a beginner in concrete modelling.
In my case, the prototype I did was a 3D gmsh model with embedded steel reinforcement inside the concrete solid. I hand coded the input gmsh file with the whole geometry.
The output from gmsh was a 3d mesh with two groups of solids, one for the steel , and the other for the concrete. The rest of the preprocessing was done in MW.

mmartin

Jalal

Hi mmartin,

Thanks for your reply. Have you compared your results with any reference? just speculating, if the nodes are following each other, this means there is no friction (slip) between the materials and it could end up over predicting concrete stresses in vicinity of Rebars. I suppose other problems would happen if you define a contact between them.
I have seen people modelling the concrete only (in Stadpro) as shell elements, e.g. in a box bridge beam or concrete stairs. Maybe it is not necessary to model the reinforcements!
Maybe it is enough to make sure that concrete does not crush in compression (30Mpa) and does not crack in tension (3.5 Mpa). The Rebars are added for strength after concrete failure like cracking and to improve shear and torsion strength. I am still studying concrete... For example in concrete piles, the size of the pile is determined by Service Load Limit and then a known amount (percentage) of rebars are added.

With regards,
Jalal.

gemada

Hi. Thanks for all the replies. Thank you MMartin for your generosity. I've not tested the formula yet, but I intend to. Jalal, I'm also a begginer in concrete modelling...usually for steel I mesh an embedded layer inside concrete (with an equivalent thickness) and use bi-linear model; for concrete I've been using compression-only (with Calculix inside Mecway)...but I'm still learning in a slow process because of work. Best regards

Mishal

Hi Guys

Just my opinion on the aspect of designing concrete, being a structural engineer.

I previously did design work on coal pre-cast bunkers, in general we would model using beam and shell elements (without reinforcement), however design was never done though FEA. The FEA results were used in-conjunction with design codes as these documents were the design standards and ultimately stood up in court in the event of some sort of failure. Even when it comes to bridge design I've seen FEA companies show these fancy FEA models, however in industry I have seen them being analyzed using grillages and code calculations (this includes the design for pre tensioned concrete).

In addition I saw a comment about the only criteria being yielding for steel, which I think is untrue. When designing structural members (I-sections, angles, channels etc...) all the structural codes deal with a variety of failure aspects which cannot be assessed by means of a linear static analysis. With plate work I think the general criteria is not as well defined when compared to structural sections, however I do evaluate the need to perform non-linear analysis, as in cases there is buckling which could exist depending on the design at hand.

Mishal

A Bit expensive, but the below book is pretty good, works in-conjunction with design codes.

https://www.amazon.com/Finite-Element-Design-Concrete-Structures-2nd/dp/0727741896

gemada

Hi Mishal. Thanks a lot for your feedback. I understand your point. In my case I'm tryng to use FEM to check some special situations to help me complement experimental analisys, not specifically design...but I must say I'm still a begginer in FEM hehehe