Pattern loading for 2-way slab (ACI 318-8 method)

[Oversee]

ACI 318-8, clause 13.7.6 gives requirement for arrangement of live load on 2-way slab:

1) Pattern loading is not necessary if unfactored live load is less than ¾ of the unfactored dead load.
2) 75% of the full factored live load on appropriate panels is required if pattern loading has to be taken into consideration.

This clause is a part of 13.7, i.e. equivalent frame method. Is the above applicable only if the equivalent frame method is used? How about the more accurate method such as FEA?

Shall I use more onerous requirements based on clause 8.11 at this scenario.

I wish to thank you all in advance.

 

[rscassar]

What FEA program you are using for reinforced concrete design. It is my personal opinion that patterned loading should be considered for all designs, not just when the live load exceed 75% of the dead load.

Australia standards have similar requirements and the way I have approached it is to either do one of the following:

1. Consider 75% of the design load patterned and not use moment redistribution.
2. Consider 100% of the design loads patterned and utilize up to 30% moment re-distribution.

I have always found option 2 to be more economical.

http://www.concretecentre.com/online_services/publication_library.aspx

Go to this site, there is a paper available for download called 'How to design reinforced concrete flat slabs using FEA'

 

[Oversee]

Thanks Kikflip

I had read the paper done by Concrete Centre and used it as the guide through the FEA. General purpose FE software was used. Pattern loading was done manually, loads were applied on the appropriate bays to achieve the worst output. I did use Australian software RAPT to design one-way slab and pattern loading could be taken into consideration.
By the way, it may be a little bit brave to follow that Concrete Centre publish, e.g. use the average bending moment within 1 meter wide within column strips as the design value Mu. I instead followed ACI318 and use those obtained by column face.

In short, 100% live load will be used in the pattern loading. I at least know something is in my pocket and there is room for value engineering.

 

[rscassar]

Sounds like a safe bet. When patterning live loads you do generate design actions which can never co-exists (maximum positive moment can never co-exist with maximum negative moment). So it does allow that redundancy to redistribute from peak moment regions in the event that they occur.

You may be interested in reading this through this thread as well. It doesn't relate to patterned loading but has an attachment from the developer at RAPT software (who is an expert on reinforced concrete and will most likely be reading over this post). This technical note was also published in the latest Concrete in Australia magazine.

thread507-275796

thread727-109456

 

[rapt]

The latest Australian code covers this.

It requires live load be placed in a chequerboard pattern consistent with code requirements for alternate span and adjacent span loadings used for 2D frame analysis.

So then it gets to what your code requires. If you are designing to BS8110, it only requires alternate span loading. But it also requires a variable DL factor between loaded and unloaded spans. so you have 1.4DL + 1.6LL on loaded spans and 1.0DL on unloaded spans.

If you are designing to Eurocode, I think it requires both alternate and adjacent live load cases but not variable DL factors!

The alternate span cases are relatively easy to create on FEM. 2 chequerboard patterns should be able to cover this. It is the adjacent span cases that are questionable in how mthey can be created. I would think full bays would be loaded. So in one direction all of span 1 and 2 loaded. Then all of span 2 and span 3 loaded etc to cover all groups of 2 spans. Then the same in the other direction.

 

[Oversee]

Thanks Kikflip and Rapt

We are designing to ACI 318, rather than BS 8110. RAPT slab member (two way strips) gives identical numerical model compared to that required by clause 13.7 of ACI 318 for the equivalent frame method for two-way slabs (without beams). There is a reduction factor, 75%, for live load when pattern load is considered. This, from my point of view, is to consider the 2-way effect and not to give too much flexural reinforcement.

RAPT was however not used because column strips perpendicular to the equivalent frame span has to be designed as “torsional members”, see clause 13.7.2.3 of the same code. This job has to be done separately and we are about to hit the deadline…

 

[rapt]

Oversee,

13.7 defines the Equivalent Frame and Equivalent Column method of Anayisis. The objective of this method is to try to estimate real frame stiffnesses that would be achieved in practice but using a simplified 2D Sub-Frame analysis method.

The transverse torsional members discussed in 13.7.2.3 are an attempt to model the rotation of the slab away from the immediate column area compared to that at the immediate column area.

It is trying to model what actually happens but in a simplified analysis.

The method does not create torsions or torsional members that do not otherwise exist. It simply uses the logic to determine a realistic column stiffness so that realistic design moments can be determined.

No matter how the slab is analysed, torsions will exist in practice as the different relative rotations do occur. FEM will actually give you the values of these torsions if you want to delve into it a (and distinguish them from the Mxy momnets). But no-one designs for the torsional member transverse to the columns in a FEM slab analysis (hopefully they do include the Mxy moments in the design in combination with the Mx and My moments as has been discussed in other threads).

To say you cannot use and 2D frame analysis such as RAPT's for a flat slab design because you do not want to design for the transverse torsional members simply shows a complete and worrying lack of understanding of what you are doing, in analysis and design!

 

[Oversee]

Thanks RAPT

Yes. You are right. Another reason scared me off using equivalent frame method is a lack understanding of these torsional members. Do you mean “torsional member” does not physically exist and ain’t need shear links to withstand the twisting arising from a) unequal frame span on both sides; 2) pattern loading? Clause 13.7.5 is only provided to ensure that the torsional stiffness of these "torsional member" is properly calculated.

The analysis software I am using allows for Mxy in the design. My major concern with FEA is about the F.O.S. of dead and live loads used for pattern loading. They have not been clearly given in ACI 318 (unlike BS8110 and EC2). I am not sure if I can borrow the F.O.S. of loads from the equivalent frame method section and therefore posted this thread.

Conservatively, I decide to use fully factored live load when pattern loading is considered at ULS state.