Tension-only reinforced concrete members with lapped splices 1

[StrucCivEnv]

Has anybody ever designed a a tension only concrete member without any mechanical or welded rebar splices? I'm designing a building without an interior concrete slab, so I can't fan out rebar from the column bases to resist the spreading forces from the roof load. The idea is to pour a grade beam from column to column with adequate reinforcement to resist the spreading load. ACI covers lapped rebar design for tension reinforcement, but it seems to be implied that this is for tension in beams. Obviously I can weld or mechanically fasten the rebar together to make a continuous reinforcement, but is it necessary?

 

[KootK]

I've got a working theory on this that I'd like to throw out to the group for discussion. Consider:

1) Apparently you're off the hook if you've got a bunch of extra cover. Why? How much cover is required to be able to take advantage of this?

2) If you're not fully utilizing your tension rebar, you're also off the hook. Why? How low do your rebar stresses need to be?

In any tension tie member where the rebar is lap spliced, there will be some degree of eccentricity between the rebar on one side of the splice and the rebar on the other side. And that implies a moment in the tension tie member. If some of your tension tie rebar needs to be used to develop this moment capacity, then you would technically not have enough rebar left over to deal with your tension load. This assumes a very tight design for the tension tie reinforcement of course.

If you weld or couple your tie rebar, you accrue two advantages. Firstly, by virtue of the installation procedure, it's much less likely that there will be a significant eccentricity between the rebar on opposites sides of the splice. Secondly, if all else fails, the member can fail in bending and the tension tie can straighten out to safely satisfy equilibrium requirements. You'd still have all the required tension capacity. Presumably, the longitudinal axis of the straightened member would then be at a slightly different angle than the original tension tie member. No big deal.

Attempting to use my theory to explain the two points mentioned above:

1) If you've got enough extra cover, then you've probably got enough "outside the tie" flexural capacity to deal with any eccentricity induced moments. Consider the example of a major drag strut in a concrete floor slab. In the in-plane direction, you've obviously got plenty of moment capacity. In the out of plane direction, as long as the adjacent slab has a decent moment capacity, you should be okay there too. A way to apply the "extra cover" out might be to assume a conservative amount of eccentricity in the tie member and ensure that there's enough moment capacity in the adjacent concrete framing to deal with it.

2) If you supply enough rebar to resist both the direct tension and a conservative estimate of the eccentricity moment induced in the tie member, you should again be okay. Maybe one could assume an eccentricity of half the tie member depth etc. Again, this would provide a means of determining how much extra rebar you need before you can go back to using a lap splice.



The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Agent666]

There is no global moment in the member if the laps are in the same place on either face. There is no localised moment in the lap if you have detailed additional stirrups at the lap cranks in accordance with code requirements. In New Zealand we are required to supply additional reinforcement to deal with the change in direction equal to 1.5 times the nominal capacity of the bars (i.e. With no strength reduction factors. If you do this there is no moment as the forces are dealt with as part of a strut and tie arrangement across the section, provided that the laps are balanced on either side of the member which is usually the case.

Either way I am sure there are plenty of structures out there that are standing with laps in tension!

 

[KootK]

Agent,

Do you not agree that there will be an internal moment in the tie if the centroids of the bar groups on either side of the slice do not coincide?

Also, could you supply the NZ code reference that you mentioned above? You've piqued my interest. I'm going to see if I can find that.

KootK

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Agent666]

Will scan it for you. I agree with the moment is there. It's just taken out by the ties in a tension and compression force at the bend/crank in the reinforcement (which is resisted by a few dedicated stirrups).

 

[hokie66]

Methinks you are overthinking this. Tension tie members are not used frequently, and mechanical splices are readily available.

 

[KootK]

@ Agent: the attached sketch shows an exaggerated version of the moment that I see in a lap splice. It would not be resolved by ties in an STM fashion. That's basically the crux of the difference that I see between mechanical couplers and laps. With mechanical couplers, the tie can fail in bending and you can sort of get rid of the moment and still be left with the tie. With lap splices, you're stuck with the moment no matter what.

@ Hokie: I'm sure that I am one of the, if not the, principal "over thinkers" on this forum. Like Lady Gaga, I was just born that way. If I had my druthers, code commentaries would be an order of magnitude longer than they are now. I think that it's pitiful that I have to wonder about stuff like this. This code cause is particularly bad. They supply a bunch of situations where the clause applies without ever actually communicating the fundamental intent of the provision.



The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Agent666]

Will take a look tonight, here's the code reference I was referring too.

 

[Agent666]

If the laps are not symmetric then there would be some net moment. Reality is they are mostly placed symmetric. Without doing any numbers I would have thought the moment is fairly small as the eccentricity is one bar diameter which in the big scheme of the force required to yield the bars results in a fairly small moment in relation to overall flexural capacity of the member.

We never account for example in design that there is a reduced effective depth at the exact location of the crank in a beam or column bar. It's just one of those many secondary things we ignore in practical design.

 

[KootK]

Agreed. The moments would be relatively small and largely dependant on construction quality control. Thanks for taking the time to post the code clause for me Agent.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Agent666]

A similar clause exists in ACI 318 under 7.8.1.3. Slightly different wording but the same intent.

 

[sybie99]

In the British code there is no mention of "tension ties". There are however requirements for lap lengths for members in tension. There are also lap lengths requirements for compression members and members in flexure. So for beams for example, the lap length for tension bars would be taken from "members in flexure".

I have a need for a concrete hanger to pick up a slab edge. Cannot have column below. So I have designed the beam above to carry the force from this tension tie. The bars in this column then go down and bend into the bottom of the slab it is supporting. I have allowed double the amount of tension bars required in the column, and as the British Code does not require couplers to be used, I made use of splices, but increased splice length by 1.5 of that prescribed in the code (40 x 1.5 = 60 x bar diameter for 30 MPa concrete. Also I provide closed stirrups at the lap zone.

I am confident that the spice connection is fine as it is overdesigned by effectively 100% to British Standard requirements ( I am aware this is not what ACI prescribes)

My only concern is at the connection where the vertical tension steel goes down into the flat slab that is hanging off the column. I have detailed the column bars to go down into slab with hooks.

If I calculate the bond stress from the straight section of bar anchoring as well as that generated by the hook, I have sufficient capacity. I have added top steel in the slab at this point (hogging moment reinf.) as if there is a column below the slab.

I have checked punching as if a column is below the slab. The difference being that the force is not transmitted from bearing of a column below pushing up, but by the stresses that come out of the column bars anchored to the slab concrete.

Has any one actually designed a concrete hanger column that supports a flat slab and how do you deal with the connection of slab to column. It is similar to a point load on a flat slab I would think.

 

[KootK]

I can't locate the thread but this exact issue has come up here at least once before. You're on exactly the right track in terms of the failure modes that you're considering. Punching shear is the really tricky bit.

How thick is your slab? What are the dimensions of your columns?

I would favour a solution that gets as close as possible to a bearing style punching shear resistance as you mentioned above. Some options, listed in order of preference (for me):

1) Get a beam or some nib of concrete cast under the column deep enough that you can develop the tie bars in it. Your architect and contractor may not thank you for this.

2) Terminate your tension tie reinforcing at a weld plate cast into the underside of the slab and sized for bearing.

3) Use ACI Appendix D provisions for post installed rebar. It's not as inspiring as bearing for punching shear but at least you'll have addressed concrete breakout in some fashion.

As you've already intimated, this is a good connection to design conservatively.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[CELinOttawa]

I think Sybie99 is touching on the truth here, and all this talk of moments strikes me as a little silly, and also of the human tendency to "run home to Momma". All of us try first to make a new problem fit the mould of what we expect, prior to trying to think of a new reason for what we see. I'll give it a go as well...

I'm a fan of reading the ACI reports (and other code committee background research) that goes into making the codes. The ACI people have got to be the most transparent, and actually it is an ACI report on laps in bars that I want to talk about.

ACI 408.3R-03 (still need to purchase and read 09, *blush*) specifically addresses in detail tension splices... More specifically it talks in detail about how splices are actually believed to work, and suggests what approximations may be appropriate to go into the codes. In terms of the transfer of a tension load between two bars, the load acts as a stress on the concrete through the mechanical deformations on the bars. That concrete then transfers the load to the adjacent bar through both the tension capacity of the concrete and a keying effect. The next bar over then adopts this load in reverse to the first (load originating) bar. There is no discussion of any internal moment, and I believe the reason for that is simple...

The eccentric effect of the bars being next to each other but not colinear is minor. For a normal bar (say 15M) for a standard lap distance, the confining force of the concrete is applied over the full length of the lap, and the eccentricity over approximately 16mm. As the gap increases, so does the length of the required splice. There is no comparison, and I think this is the major reason why there is nearly never an issue, except for where there is insufficient cover. The minor applied moment is overwhelmed by the tension capacity of the confining concrete, and so long as this very rare failure mode does not occur, the loads will transfer as through a standard stress lap effect in the PCC.

 

[sybie99]

KootK,

The scenario is as follows, also see attachment: 300mm thick slab, 35MPa concrete, hangs off a 450 x 230mm column (edge column).

There are eight 16mm bars in column. Yield stress of bars is 500MPa. The bars go straight down into slab for 150mm before bending with a 120mm radius bend to join into bottom mat reinforcement in slab. The horizontal part of the bar is 1000mm.

The tension force is 190kN (ultimate limit state force). The top steel in the slab at the column location is 565mm2/m perpendicular to slab edge and 1000mm2/m parallel to slab edge. I have checked punching using the following input: 300 x 200 column area, 35MPa concrete in slab, 250mm thick slab (actually 300mm thick but force is not applied at slab soffit. No punching rebar needed.

 

[KootK]

You might want to move this question into its own separate thread Sybie99. We're drifting a bit from the OP's original question.

What you've detailed is spot on for dealing with punching shear. The question here, however, is whether or not punching shear is applicable to your situation. Were you able to employ methods #1 & #2 that I mentioned previously, I think that you could argue that punching shear applies.

As you've detailed your situation, unfortunately, I don't think that punching shear is the appropriate way to look at things. It leads to a check based on the wrong shear crack in my mind. Check out the attached sketch for my thoughts on the matter and a proposed solution. Do the British codes have hanger steel requirements for beams supported by other beams? Canada's code does and I see that as being the applicable design check here.

Another thing to consider is integrity steel design. In both Canada and the US, you have to have integrity steel installed that would provide an alternate load path if punching shear doesn't get the job done. With your detailing scheme, I don't think that you'd have anything to satisfy that requirement. Since you're using BS that may not be an issue for you.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[hokie66]

KootK is correct about the critical crack plane.

 

[TLHS]

KootK,

A23.3 has the same clause regarding tensions ties and mechanical splices. In the 2004 version of the spec it's 12.15.5.

 

[KootK]

Thanks TLHS. For exposed tension ties, I've also applied clause 8.12 of the Canadian Highway Bridge Code (crack width).

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[sybie99]

KootK

I have decided to also have a steel plate with 4 No 20mm bars butt welded to the plate, to sit below the slab. These bars are then lapped into column. The plate pushes up against slab soffit and then gives a scenario similar to a column that sits below slab.

 

[KootK]

I'm glad to hear that sybie. I think that the plate is a good solution. Bonus tip: in my neck of the woods, that plate would have to be fire proofed.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.