AS3600 and ACI209 Creep/Shrinkage Models

[Trenno]

Hello everyone,

Has anyone previously had a crack at adjusting RAM Concept's creep and shrinkage inputs to more align it with values adopted/common in Australia, particularly creep coefficients and shrinkage strains according to AS3600?

I've been having a discussion with the Bentley support team about aligning the ACI209 (RAM default) and the methods of AS3600. Discussion found here.

After running numbers and looking at how each code accounts for factors such as initial time of loading, environment, size effects etc - I have come to conclusion that it is not a smart idea to try and combine the two different creep and shrinkage models, even though Bentley recommend it.

In regards to shrinkage strain, RAM adopts an average value of 400E-6. This value is below the design shrinkage strain recommended by AS3600 for favourable conditions (N40, Tropical, 200thk, basic 1000E-6). Not good.

The differences and attempts to merge the two different creep/shrinkage models raises alarms for me. I know concrete deflections are so highly variable but I want to assure myself I am estimating them as accurately as possible in the attempt to get close to realistic deflections. "Garbage in, garbage out."

Would be very interested in how Aussie designers are approaching this.

Thanks

 

[Trenno]

Now having a quick look at the shrinkage restraint %, which is explained here.

shrinkage restraint % = (f'ct.f x %Mcr) / (E x Sigma.cs). So you can simply calculate a shrinkage restraint % based upon a desired reduction in the cracking stress. But the question is, what is an appropriate desired deduction in the cracking stress? Setting a high percentage cracks the floor entirely and increases deflections. RAM does make some suggestions based on a shrinkage strain that I don't agree with.

Below are my estimates for a 200thk slab, N40, Interior enviroment (shrinkage strain = 610E-6):

%Mcr=0 SR%=0

%Mcr=25 SR%=4.75

%Mcr=50 SR%=9.5

%Mcr=75 SR%=14.25

%Mcr=100 SR%=19 (At this stage the restraint to shrinkage has caused the slab to crack fully)

 

[IDS]

I'm not a RAM Elements user, so I can't give any specific advice, but a general approach that should give reasonable results is:

- Find the total shrinkage and creep values from AS 3600, taking account of hypothetical thickness, humidity, etc.
- Replace the default strain with the AS 3600 value.
- Set the other values so that factored strain is equal to the input value

The shrinkage restraint tolerance is the only problematic input (I think). The suggestion in the linked document that the restraint is entirely due to external restraint is just wrong. The slab reinforcement restrains the shrinkages, and reduces the cracking moment. Also it's not a matter of a desired cracking moment reduction; you should be looking for the reduction in cracking moment that will actually occur.

A reasonably straightforward approach would be to calculate the cracking moment (Mcr) using the method in the notes under Formula 8.5.3.1(1) in AS 3600, then calculate the % reduction factor that gives the same result.

You should also be aware that deflections using Branson's Equation (as in AS 3600 and ACI documents) are likely to be unconservative, particularly for lightly reinforced sections, and ACI provisions also underestimate the effect of shrinkage after cracking, which is significant even for symmetrically reinforced sections.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[asixth]

I have always used default values and never really thought too much about it.

I guess that's not very helpful.

 

[asixth]

Someone tried to author an in-house manual once. I never really went thru this to see what they suggest and I would be careful with using what is suggested. Have a read and see if it helps.

 

[Trenno]

Thanks for the reply, Doug. I have been reading some of your papers on the topic. Very well written and just goes to show how variable the concept of long term concrete deflection can be (especially in that 28 storey case study).

A reasonably straightforward approach would be to calculate the cracking moment (Mcr) using the method in the notes under Formula 8.5.3.1(1) in AS 3600, then calculate the % reduction factor that gives the same result.

Sorry, but into the formula quoted above it should be "cracking stress" not "cracking moment." 0.6 root f'c. Could you please elaborate on the '% reduction factor that gives the same result' part of your suggestion?


Asixth, here I am thinking you would have been atleast one person to delve into how RAM does creep/shrinkage! I'll have a read of that document and see what I can find.

 

[Trenno]

After reviewing the Bird's document, I still don't think you can manipulate the inputs to truly align it with AS3600.

For one, ACI don't have a basic creep coefficient based on the concrete strength, rather they use a base coefficient (irrespective of strength). Derivation of additional factors is also different.

I'm also not sure whether the input creep and shrinkage should be final or 'with respects to time', as RAM iterates these values at each load history step.