Suspended slabs with Class L mesh

[sdz]

I just downloaded SRIA TN6 "Design to AS 3600:2001 of Suspended Concrete Floors Reinforced with Class L Mesh"

I found it interesting, useful, and free.

http://www.sria.com.au/

 

[rapt]

sdz,

What did you find interesting and useful about it?

 

[apsix]

rapt

That wasn't your reaction? (I haven't read it yet)

 

[OzEng80]

sdz

You might find the attached document interesting as well.

Cheers

 

[sdz]

Rapt,

I thought it provided a good overview. Personally, in the past I have designed many mesh reinforced suspended slabs. I haven't had to design any since those amendments came out.

 

[rapt]

Apsix, sdz,

Both articles are presenting one side of the argument. Both of those articles come from the same source. There are a lot of people both in Australia and internationally, who take the opposite position, and they do not have a financial interest in the material!

The biggest problem is with very lightly reinforced members with low ductility reinforcement. It is easy to show that the material performs reasonably if the steel ratios are significantly above minimum, say in the order of 3 or more times minimum. At those steel percentages, the strain in the steel is not as bad and there is an increased factor of safety to failure. The problem is when very small steel retios are used and the strain in the steel is much higher.

In my opinion, the 20% reduction in capacity is actually insufficient for low steel ratios, especially as low as minimum reinforcement. At these reinforcement levels, the rectangular stress block cannot be used for the calculations (irrespective of what these articles say). You have to do a strain compatibility analysis and reduce the compressive strain from the normal .003 to about .0011 to actually make it work, so the concrete is not going into the plastic range and the rectangular stress block is not applicable. At minimum reinforcement, I think the reduction from the capacity calculated using rectangular stress block should have been more like 40% than 20%.

And the real problem is that low ductility meshes are normally used in slabs with low reinforcement ratios.

There are gaps in the code that let you get away with the calculations as they are being done. The code puts no upper limit on strain in the reinforcement at ultimate when it should. We have been trying to get this addressed for years but without success.

If you are going to use the material, make sure you do the calcultions properly!!!

 

[hokie66]

I never have understood the motivation for going to higher strength reinforcing steel, higher than 400 MPa, that is. Many of our problems in buildings are serviceability cracks, and that is made worse by using less area of steel.

Why would the steel companies have wanted to sell less material? I doubt that the price went up 25%.

 

[rapt]

hokie66,

I think the initial motivation was that they were just about making 500MPa steel anyway and selling it as 400 so one company used it as a marketing tool to try to big name themselves, as is now happening with the Class L meshs and their "new" deflection calculation method which we have known about for years but they have now copyrighted a version of for themselves.

I think they were and are getting bad advice from certain expert sources.

500MPa steel is actually very good for heavily loaded columns and heavily reinfroced beams. Otherwise, I agree, we were better of with 400MPa steel. Better crack control, better deflections and less calculation problems for crack control. Unfortunately it is too late to go back know.

 

[asixth]

I'd definitely say that I am one of the junior engineers within our industry, however, I found the paper to provide quite a useful example.

Saying that, I did find it to be technically unconvincing in places. For example:

"10 Design Vertical Shear Strength

"The full cross-sectional area of Class L mesh may be used to compute ultimate shear strength, Vuc, in accordance with Clause 8.2.7.1"

One point that I did find interesting though is the recommended use of the uncracked section modulus for the structural analysis. Clause 7.6.6 states that the deflections must take into account the effects of cracking amongst other things. I would assume the best way to account for the additional deflections that result from cracking would be to reduce the I value of the cross-section, therefore, I would be using cracked transformed section, which would significantly reduce the I value for a lightly reinforced section, Icr < 0.5Ig I would imagine.

Within the general philosophy of AS3600 to keep the stiffness consistent throughout the design, this would mean using a lesser I value for both strength and serviceability checks.

 

[OzEng80]

asixth,

The recommendation is probably based on the requirement for 'stiffer' slabs with no moment redistribution. I too would take your approach and model with a reduced M of I (as long as all sections in the analysis are reduced accordingly it won't effect relative stiffness and the moment distribution).

 

[rapt]

Asixth,

Don't believe everything you read in technical papers, especially when they are produced by someone trying to sell something. There is too much logical arguement, from people who do not benefit either way from the arguement, against its use to risk using it. As I pointed out in an earlier posting my calculations show it to be inadequate and agree with test data, without having to go into

RE shear, interestingly some other codes have gone the other way, I think it is the Canadian code (from memory) that specifically precludes the use of low ductility steel for shear, while AS3600 still allows it.
I know they are cuurrently having the same arguements as us against Low Ductility reinforcement in Europe, and their minimum strain is 2.5% while ours is much worse at 1.5%.

RE the analysis,
That is just regurgitating the standard AS3600 rule that allows everything to be be based on an elastic analysis and as OzEng80 has suggested, it also fits in with the logic of no redistribution. Unfortunately redistribution occurs in all slabs as they crack during the loading cycle, even if you do not allow for it in design.

Design to analysis is different and cracking must be allowed for deflections. However, if you are using FEM analysis and start trying to reanalyse for cracked sections for deflections, what properties do you consider cracked values for? Obviously bending stiffness, but what about torsional stiffness and column stiffness. That is where current FEM design programs are still not telling all. What are they assuming to be cracked, only the slabs bending stiffness?? If you are going to start doing this then you also have to start allowing for creep and shrinkage in the cracked analysis and also construction seguence. All too difficult.

 

[sdz]

Rapt, going back to some of your comments above:
"The problem is when very small steel retios are used and the strain in the steel is much higher. ... At these reinforcement levels, the rectangular stress block cannot be used for the calculations (irrespective of what these articles say). You have to do a strain compatibility analysis."

I have just done a check using elastic theory at low reinforcement.
Elastic Mue=As fsy j d
Plastic Mup=Ast fsy d(1-Ast fsy/(1.7 b d f'c))

f'c=25 MPa; gamma=0.85; Ec=27500 MPa; n=Es/Ec=7.27
D=200 mm; d=160 mm; b=1000 mm; Ast=400 mm^2
p=0.0025; k=0.173; j=0.942

Mue=30.2 kNm; Mup=31.1 kNm; Mup/Mue=1.03
(capacity reduction factor not included)

So at low reinforcement ratio the lack of ductility makes very little difference in calculation of capacity. In fact since the calculation of Mue assumes only elastic strain any yielding will narrow the gap.

It might be that at intermediate reinforcement ratios there would be a bigger difference but I havn't had time to do further checks.

 

[rapt]

sdz

Your elastic calculation is still assuming you can achieve the full concrete compressive stress. To limit the strain in the steel, the compression in the concrete must be reduced. A lot of factors then start to come into it, including cover, concrete strength and percentage of reinforceemnt especially (you have used 400mm2 in your example where minimum is 290 based on the cracking moment rule!

With minimum steel, the strain in the steel is approximately 4 times the peak strain. This has to be limited to less than 1 times.

This results in a lower compression in the extreme concrete surface, deeper neutral axis depth and consequent reduction in capacity.

I will do some numbers and get back on this but last time I checked, the compression strain was .0011 and the reduction in capacity was 12%.

Add to this a capacity reduction factor for a non-ductile section of .6 compared to .8 for a ductile section and the reduced ultimate capacity is in the order of 40%.

 

[sdz]

rapt,
I think you are wrong but I will check over the weekend and post a results. The elastic method uses a transformed section and does not assume concrete will be at max compressive strain. I ignored capacity reduction factors because I wanted to show that capacity is not strongly dependent on yielding of steel at low reinforcment ratios.

 

[sdz]

OK I checked my figures. For the case above concrete stress=14.4MPa, assuming linear elastic behaviour.
I have also checked a range of reinforcement ratios and confirm that at low reinforcement ratios Moment Capacity does not depend much on yielding of the steel.

Design Capacity on the other hand depends on the capacity reduction factor used, and that can be somewhat arbitrary.

Perhaps this is why slabs reinforced with low ductility mesh seem to have exhibited satisfactory behaviour in practice. By the time yield strength is reached ~97% (in my example) of Moment Capacity is reached, without any yielding of the steel. Any yielding past this point provides some ductile capacity.