RAFT/MAT FOUNDATION - ONE WAY SHEAR 2

[rummaan17]

Dear All,
Do we design a Raft for One Way Shear, If Yes what is the criteria to check one way shear design of Raft/Mat foundation. ?

I am checking a Raft shear design with allowable shear stress in SAFE and doesn't seems to be reasonable. Most literature favors checking of shear forces rather than shear stress and to consider width that extend along the Raft.
ACI 318-08 SECTION 11.11 Provision for Slabs and Footing - 11.11.1.1 - "Beam action where each critical section to be investigated extends in a plane across the entire width"

Please comment on the recommended method of design if any. Thanks in Advance.

Find attached images of two method I adopted.

http://imgur.com/a/vEg1a

 

[KootK]

See page 15 of this document: Link.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[DaveAtkins]

I use RISA-3D, not SAFE, but I struggle with the same issue. The problem is, these programs cannot distinguish between two way shear and one way shear. On the output, all shear is shown the same way.

Here is what I do. Where I see large shear forces surrounding a concentrated load on the output, I treat that as two way shear. Where I see a large shear force not adjacent to a concentrated load, I treat that as one way shear.

Sometimes the one way shear is too high. Usually when this happens, the shear force tapers off quite quickly. If so, I look at the output to see if I can use an average one way shear over a reasonable width.

DaveAtkins

 

[bookowski]

I've posted a similar question Link here.

You can find some papers/opinions on this but nothing definitive. It doesn't make sense to use the entire width if that width extends well beyond the element(s) applying the load. I usually count 'd' on each side of the shear wall assemblage.

 

[geo321]

I follow the rule given by kootk link.
However if u have an irregular grid of columns should the shear plane be free of any vertical element ?

 

[JoshPlumSE]

I have personally used an shear capacity of 2*sqrt(f'c) in conjunction with an effective width of pedestal width + depth of slab on either side. Very close to the procedure described in that NEHRP document KootK linked to.

The use of such a low shear capacity (half that used in other places) described in that document gives me pause. My tendency is to think that value is just too conservative. But, my structures have all been ones where we knew 2-way shear was going to govern anyway. And, checking 1-way shear was really just a formality.

 

[KootK]

ACI 318-08 SECTION 11.11 Provision for Slabs and Footing - 11.11.1.1 - "Beam action where each critical section to be investigated extends in a plane across the entire width

I wouldn't actually support the use of this clause for a raft foundation. I believe that, in spirit, it was intended for single column footings and perhaps very simple combined footings. At a minimum, I would treat a raft foundation as we would an elevated slab, where one way shear is checked over design strips. And, like I said above, I generally follow the NEHRP guidelines for complex situations and where shafts and lateral loads are involved.

The problem is, these programs cannot distinguish between two way shear and one way shear. On the output, all shear is shown the same way.

Sometimes the one way shear is too high. Usually when this happens, the shear force tapers off quite quickly. If so, I look at the output to see if I can use an average one way shear over a reasonable width.

These things raise an interesting point in my estimation. Technically, I don't think that there's anything wrong with the computer output. From a plate mechanics perspective, there's no such thing as a distinction between one and two way shear. There's just shear. The one-way / two-way stuff just represents the particular "correlate to testing" bucket that we practitioners dump a specific problem into.

With FEM, I'd like to somehow get things down to a two step algorithm for shear.

1) Check two way at columns and such.

2) Check point stress one way at all locations and not bother with "planes of interest".

The trick, as DavidAtkins has alluded, would be finding a rational way to average out the peaks. I wonder if a rolling average might suffice. Say, the average value taken over three slab depths in any direction. That would be insane for a human to work out but a piece of cake for a computer.

Even in normal, two way, elevated slab design, the use of an average one way shear value taken over a design strip has always bothered me. When you look at it in FEM, the one way shear is all over the map as you approach your columns. It would seem to imply a great deal of capacity to redistribute shear laterally. And that seems at odds with the commonly held dogma that Vc shear failure is generally not so ductile.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[rummaan17]

Thanks KootK, DaveAtkins, bookowski, geo321 & Joshplum for your response.

I want through the document and your previous discusiion before posting my question. I am still unsure about the width to consider & If stress output from SAFE can be utilized to check shear.

I am also struggling to find the difference between shear output as if it oneway OR two way. Please see the attached Link for method, drawing & code reference I utilized.

http://imgur.com/a/8vIxm

Method 1: - I utilized shear force output based on ACI formula of Vc= Phi x 0.17 x SQRT (f'c) for one way beam shear and found my raft to be exceeding this value alot of places.

Method 2:- I put a design strip from Center to Center of span. Still the shear value I get in Span ( other than column drop Or thickened places ) is very high.

I am still unsure if increasing raft thickness for Oneway shear will be a viable solution.

 

[rapt]

Definitely need to check one way shear by the normal beam shear or flexure shear rules.

If the column pattern is reasonably regular then the normal rules would allow you to base it on shear over the full panel width (most Design codes require this) or in the worst case over the normal column strip width (approx. half of the slab width with approx. 75% of the shear) (what we do in RAPT software).

The only time this needs to be reduced is for transfer slabs with very irregularly placed columns where you would need to break the slab down to more concentrated load areas as you should for flexure anyway.

To match design code shear calculation rules I assume you would have to convert it to shear forces rather than stresses!

 

[rummaan17]

Just need to inquire one thing, Please see the attached PDF

http://www.sefindia.org/forum/files/structural_design_of_raft_foundation_869.pdf

Page-15

Do we need to subtract soil pressure to get resultant shear value. Please comment. Much Appreciated.

 

[DaveAtkins]

If there is soil above the footing, then yes, it counteracts the uplift from soil pressure below the footing. But you should have included this as a dead load in your model.

DaveAtkins

 

[rummaan17]

Dear DaveAtkins,
I think a code clause clearly indicates when to check for one way shear . Please see the attached image. Your response is much appreciated.

http://imgur.com/PkzQudW

 

[KootK]

need to subtract soil pressure to get resultant shear value. Please comment. Much Appreciated.

I think that what they're doing there is subtracting the load/reaction that occurs within "d" of the support. I believe that's a legitimate and ubiquitous reduction to take.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[rummaan17]

yes Kootk, it took me a while to realize that he is just getting the shear at distance 'd' from support. What do you interpret from the Code clause I shared.

 

[KootK]

What do you interpret from the Code clause I shared.

Not much. I'd translate that to, essentially:

1) Always check for one way shear while recognizing that it's unlikely to govern in many two way situations.

2) Always check for two way shear near concentrated loads and supports unless proportions are such that you can exclude that failure mode by inspection.

I'd recommend thinking of things in this fashion:

1) Separate "analysis" and "design" as two separate activities in you mind. Analysis yields the demand side of things (applied shear force or shear stress). Design addresses the capacity side of things (allowable shear force or allowable shear stress).

2) Realize that, when it comes to analysis, there is no meaningful distinction between one and two way shear. There's simply... shear. This is why software doesn't make the distinction except for in post processing situations: there is no distinction to be made

3) When designing rafts for shear near areas of concentrated load or support, compare your shear demand (analysis) on a given plane to a capacity based on two way shear code design provisions. The allowable shear stresses are generally higher than those allowable for one-way shear. Why? Because testing says so.

4) Realize that one way shear capacity must exceed one way shear demand everywhere where two way shear capacity provisions to not apply. Everywhere.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[rummaan17]

Dear Kootk, Thanks for a detailed response.

May be I am trying to get away with one way shear check as it might result in increasing thickness or providing shear link. Secondly, I checked with alot of engineers who designed raft considering one way action and the structure is still sound. Though alot of checks are deemed necessary but an Engineer also has to focus of being economical with safety.

I can leave the ball in court for my seniors to decide now. Thanks alot for help

 

[rapt]

"Though a lot of checks are deemed necessary but an Engineer also has to focus of being economical with safety"

If a design code specifically requires a check or sets design limits, there is not much scope for an engineer to over ride that design requirement in the interests of economy! The design code has already been written to balance the needs for economy and safety.

Especially once there are problems and the legal system gets involved. In fact, even if the design code does not specify a limit but if it is generally accepted that it should or new research shows that it should, the engineer is responsible for keeping up to date and employ those extra checks!

 

[KootK]

or in the worst case over the normal column strip width (approx. half of the slab width with approx. 75% of the shear) (what we do in RAPT software).

I'd like to know more about this as it would seem to address the concern that I mentioned above regarding the lateral distribution of shear within a design strip. Is this something that was built into rapt in response to specific code provisions or is it just something that was thought to be prudent (I quite like it)?

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[rapt]

Kootk,

AS3600 has always been interpreted to suggest flexure shear in flat slabs be checked as the total panel shear on the total panel width, so averaging the shear over the total width. Though checking the wording now, it does not specifically say it, depends on interpretation. The wording is

" Where shear failure can occur across the width of the slab, the design shear strength of the slab shall be calculated in accordance with Clause 8.2."

The phrase "across the width of the slab" has always been interpreted to mean the use of the full panel width with the full load, even from memory within the Australian Concrete Standard Committee. I think the assumption is that for a collapse due to shear failure to occur it would have to be over the full width so redistribution of shear would occur to average it out.

I cannot remember the wording on this in other codes.

However, I was never willing to accept this so I took the conservative/consistent solution with RAPT and base it on shears at each cross-section consistent with the moments at that cross-section, so the column strip controls the shear design with 75% of the shear at the face of the support in the column strip and reducing to 60% in the column strip at the maximum +ve moment point (percentages by default from AS3600). This is logically required in PT design as there is an M/V term in the shear design in many codes and M and V must be consistent and I refused to use the total panel moment and total panel shear in this case!

This is more critical in a transfer slab of course where there are more design strips than column and middle required depending on the complexity of the column layouts above and below. Unfortunately, many newer designers seem to ignore the extra complexity in these cases because the FEM software does not do it automatically for them! In fact, many designing to ACI are simply using the banded/distributed average moment logic as that is the default that the programs provide.