Topic: Early age cracking in young concrete

Dear sir or madame

We are currently researching early age (0-28 days) cracking due to restraint for a thesis at KTH college (in collaboration with Ramböll) in Sweden. We are using Atena Science/GID to model this problem, using the transport data problem and importing our data to a creep problem. We have looked at the examples and tutorials for Atena Science/GID concerning this.
We have achieved a working transport model and imported that data. Our current problems are:

1. In the creep model, which Creep Concrete material would be appropriate for this? We have simulated a young concrete material by using the Cementitious2Variable material and applied gradually increasing E, FT and FC values according to Eurocode. Is it correct to then use this material as a base material in a concrete model?
2. The concrete heat expansion during hydration causes very high strains, which normally don't lead to cracking due to a very high creep factor and low E modulus in the young (<4 days) concrete. Is this possible to model, and if so how?

Thanks in advance
Martin Lagesson
Anton Hägerstrand

Re: Early age cracking in young concrete

Dear Martin and Anton,
for this kind of problem, I think the Creep module is needed, while the simplified representation of young concrete using Cem2Variable makes sense in different situations. Quickly thinking about it, I can not come up with a situation where choosing the Cem2Variable as base for the Creep model would make sense. The most typical approach is

I. Generate the "static" material properties corresponding to the age of 28 days using the EC2 or Cementitious2 template.

II. Create the Creep material using the above as Base.
If you have any measurements (losses, E, ...), enter them - when available, this can improve the results precision significantly.

III. Set the times at the Creep tab of the Problem Data dialog and in Interval Data accordingly.

Regards.

Re: Early age cracking in young concrete

Thank you for taking your time to reply.

We are currently experimenting on how atena handles temperature boundaries, because we observed some strange behaviours in our hydration model. We have created a simplified transport problem where two mature concrete objects with different initial temperatures are connected via fixed contact surface. One of the objects has an initial temperature of 30C and one 10C.

When we calculated this model the contact surface did not transfer any heat between the volumes. However, when we applied a Moisture temperature boundary (20C) to the surface of one of the objects there seemed to be some heat transfer, but the volume without a boundary surface got colder than it reasonably should.

Volume 1 has an initial temp of 30C
Boundary condition on that volume applies a temp of 20C
Volume 2 has an initial temp of 10C

Image album with results http://imgur.com/a/zQDHH

If you look at the images I'm linking you'll notice that Volume 2 gets much colder than it should, starting near the surface w. the boundary condition. Is this an unreasonable way to model our interaction? Should it be modelled in a different way?

I want to reiterate that this is not the actual experiment we are doing, but we modelled this because we observed strange behaviour in our curing model, where we have an early age concrete material connected to a mature concrete material with a master/slave surface connection.

Regards
Anton Hägerstrand
Martin Lagesson

Re: Early age cracking in young concrete

Please send us your model such that we can check it.

How large are the convergence errors?

Regards.

Re: Early age cracking in young concrete

We have a working model of a wall in the transport module (without a slab). Our problem now (in creep module) is how to model the walls restraint againts the slab in a lifelike way.  We would like to model a slip against the slab at the edges of the wall, to get a more lifelike cracking behavior. We now have just modelled whit a fix contact for surface against the slab. Our current cracking behavior on our model is shown in a attached picture. http://imgur.com/a/7UBcI

Re: Early age cracking in young concrete

Dear Laguess, please see ATENA Troubleshooting, 2.2.14 I have problems when using the Interface (GAP) material on contacts for the basic information related to using Interface elements (and connecting neighbouring volumes in general).
If you decide to use Interface elements, please follow ATENA-GiD User's, 5.3.6 Interface Material.
Regards.

Re: Early age cracking in young concrete

Thank you for your help.

When we run our creep with our transport results imported it appears the creep model does not take the temperature difference from the hydration into consideration. We want the shrinkage of our model to account for the temperature drop from the hydration difference. If we start our intervals from earlier on we get errors because we precede our curing time. If we set our curing time at an earlier point our material goes to failure because the concrete has very low tensile properties, and in reality it should act more like an elastic material at that point in time.

This is a link to our current model. https://drive.google.com/open?id=0B_7fw … FZleTdhekk

Thank you in advance.

Re: Early age cracking in young concrete

Dear Martin,
after downloading from the link, I see a GiD model named Transport but being of the Creep Problemtype + some mix of input and output files from Transport and Creep in the AtenaCalculation subdirectory. Please send us both models, without results but including the last .out+.msg+.err outputs, in 1 archive (or 2 archives if you prefer), following Troubleshooting, 2.1.1.

In the model I see, there is only 1 Interval with a single step, and importing the Transport results is NOT enabled in Problem Data (tab "Transport") nor in Interval Data ("Loading by Imported Temperature"). The 10x07x1 transport.inp (also of the Creep Problemtype) clearly does NOT correspond to the GiD model.

Regards.