Strut & Tie Models - Superposition of Column & Reinforcement Forces 1

[Drapes]

In developing a strut and tie model, is it necessary to first split the load b/w the concrete and reinforcement and merge 2x models into 1x model?

I came across a worked example of a strut and tie model b/w a column and pile cap (refer attached), where 2x models (representing the load carried by the concrete and load carried by the reinforcement) were superimposed to give a final model and strut angle. The centreline of the nodes for the 'reinforcement' model was based on half the compressive development length of the bars, leading to a sharper strut angle that the 'concrete' model.

I suspect this superposition would only apply if the reinforcement is required to satisfy bearing. But if you have a column with say 2 to 2.5% reinforcement, and that reinforcement was only required to give the column adequate capacity under combined axial and moment over the height of the column, but under pure bearing the concrete alone was deemed sufficient at the column/pile cap interface, is there any need to split the model into two? In this case, would it be acceptable to simply create 1x model at the outset, apply the full load and go from there?

For example, simplistically if we assume a column with an axial capacity of 10000kN over the height of the column (with 8000kN contributed by the concrete and 2000kN contributed by the reinforcement), and the design load is 7000kN so there are no issues with bearing using the concrete capacity alone, can the load taken by the reo be ignored as far creating a strut and tie model is concerned?

 

[Drapes]

In case the link doesn't work.

 

[canwesteng]

I can't say I've ever done it, although I've never needed reinforcement for bearing at the column pile cap interface. I'd say the cost of slightly larger piers so bearing works without rebar is less than the cost of you doing this analysis.

 

[KootK]

In developing a strut and tie model, is it necessary to first split the load b/w the concrete and reinforcement and merge 2x models into 1x model?

No, for most models it is unnecessary. The only practical situation in which I can think of it being required is the vert one that you mentioned: where compression needs to be transferred, in part, by the rebar. In that situation, the two model approach is a useful method of getting to a reasonable, final model that accurately (hopefully) reflects the reality of how bearing stresses will actually be delivered to the cap.

Can you point us to the document that contains the attachment that you provided? This is the first time that I've seen something like this in print and I'd very much like to check it out in greater detail.

...can the load taken by the reo be ignored as far creating a strut and tie model is concerned?

Good question.

1) I suspect that most engineers would make that assumption. I certainly have in the past.

2) In a heavily loaded, multi story column where much of the compression is designed to be taken by the rebar:

a) Creep effects will result in even more of the compression being taken by the rebar. So a lot of the compression may well arrive at the joint through the rebar.

b) Before the concrete only bearing mechanism can kick in, some combination the following may need to occur:

i) Bar forces may quickly need to transfer over to the adjacent concrete.

ii) The compression bars may need to initiate a bond failure if designed for less than the full compression development length.

iii) The compression bars may need to initiate a punching shear failure beneath the diagonal compression struts.

c) After all of the bad stuff listed in happens, you do still have a valid, competent load mechanism for that particular load action.

 

[Drapes]

Thanks KootK, that's a helpful assessment.

After all of the bad stuff listed in happens, you do still have a valid, competent load mechanism for that particular load action.

Do you think a redistribution of load of this nature from the reinforcement back into pure conc bearing at the column/pile cap interface (due to say bond failure in the reo) can be justified?

The worked example can be found from 'fib Bulletin 61: Design examples for strut-and-tie models: Example 1-3 Pile cap for precast concrete piles'

 

[KootK]

Do you think a redistribution of load of this nature from the reinforcement back into pure conc bearing at the column/pile cap interface (due to say bond failure in the reo) can be justified?

That's up to the designer given the particulars of the situation. If there's to be tension in the column bars at some point, it won't due to have failed them in bond prior to their seeing that tension. I also don't love the idea of a mini-punching shear breakout at the bottom creating a vector for moisture to get in from the bottom. I usually ensure that, at a minimum, the column bars are compression developed into the foundation.