Caisson foundations and ties for Pre-Engineered metal building

[Jc67roch]

I am designing the foundations for a pre-engineered metal building. Due to fill and other soil conditions, we are founding it on 3 foot diameter caissions drilled 12 feet or so to bedrock. Grade beams span between the caissons. Piers for the building columns are to be constructed on top of the caissons and integral with the grade beams.

I have calculated that the caissons should have sufficient lateral load capacity to carry the loads from the building base plates, as well as the vertical loads. The column foundation piers should transfer these loads to the caisson tops. Do I still need to utilize hairpin bars in the floor slab, or cross ties between opposing column bases, to negate the lateral loads? Is it ok to rely on the lateral load capacity of the caissons as calculated? Additionally, it seems the grade beams extending 4 feet below grade also will offer a lot of excess lateral load capacity (safety factor) from passive resistance of the soil against them.

 

[SrVaro]

I will assume this is not "high" seismic where ties are required per code. If the caisson is able to resist, then I would think there is no need to add the ties. However, be sure to consider both load and what the deflection at the top of the caisson is. PEMB have relativelty high outkick forces and if the foundation has movement at the baseplate, it will add additional stresses within members of PEMB that have already been designed with little to no additional capacity.

 

[MiketheEngineer]

Hair pin bars are cheap insurance.

 

[msquared48]

I agree with Mike here...totally.

Creep could still be a factor over time with the caissons and hairpins will help mitigate that.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[a2mfk]

You have 4ft deep grade beams, or am I reading that incorrectly?

But no matter your foundation system, I agree with Mike.

 

[msquared48]

Oh, and there are two main reasons why you would ever need to use cross ties between the mainframes:

1. No slab situation, such as a riding arena, and

2. Lateral loads so high that hairpins cannot take the load.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[msquared48]

How big is this building anyway?

The "grade beams extending 4 feet below grade", do seem like an overkill here.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[Brad805]

We do these foundations without hair pins all the time. We have to deal with expansive clay soils and our geotechnical engineers typically object to hairpins or any connection between the slab elements and the perimeter foundation. I am curious if you are relying on the fill in your lateral load calcs. It always is a difficult to judge what to use for fill.

Brad

 

[Jc67roch]

Thanks All! I will use the hairpins - as you noted, cheap insurance.

The grade beams carry only very light exterior wall loads, and are extended to 4 feet below grade for frost protection. I am relying on the fill for the lateral load resistance, but using a relatively low allowable.

 

[hokie66]

I have never been involved with a "PEMB", but have done lots of big industrial steel framed buildings. I don't like the concept of tying the structural frames to non-structural slabs on ground.

 

[BAretired]

You could allow the slab to float and tie across with a tie beam under the floor slab. This is more expensive, but it satisfies hokie's concern.

BA

 

[Brad805]

Ties can be a problem in an industrial plant where they always seem to retorfit equipment every few years. We did a reno project one year that involved a lot of underground piping in a PEMB with ties like BA suggests. I have always sized piles to avoid ties since then.

Brad

 

[RHTPE]

To think that hairpins from the foundation into the floor slab satisfies the support condition assumed by the PEMB designer can be risky in MHO.

I am investigating a foundation failure under a PEMB frame. In the course of my investigation I modeled the PEMB frame, which is assumed to have pinned connections at the bottoms of the columns. I applied the full design snow load + the dead load and recorded the maximum moments in the frame. Then, I released one of the column's X-direction constraint (introduced a roller). Surprisingly, the column deflected horizontally only a little over 3/8" (50 ft span). But even more surprising was the increase in the positive moment in the roof beam - to 140% of the previous value.

The interior floor slab section connected to the foundation may offer enough drag resistance to the columns horizontal reaction. Or it may not. Then consider the amount of floor slab saw-cutting done to counteract slab shrinkage cracking - is there really continuity across the width of the building? In my situation, 3/8" horizontal movement of the column is within the realm of jobsite construction tolerances and lack of quality control. But ask yourself - can the roof frame safely sustain a 40% increase in positive moment?

If you're comfortable that the frame can sustain the moment increase caused by an outward movement at the bottom of the frame's column(s), then why worry about tying the foundation to the floor slab, or introducing a tie rod under the floor slab?

Guaranteed to elicit some interesting testimoney should a failure occur and the PEMB manufacturer wants to shed some liability.


Ralph
Structures Consulting
Northeast USA

 

[BAretired]

RHTPE,
I do not find it surprising that you would find a 40% increase in positive moment if you substitute a roller support for a hinge support. But the fact remains that a roller support is not realistic when the foundations are tied together by a tie beam or a tie slab.

BA

 

[hokie66]

I find it incredible, although I have done no numbers myself, that the rafter moment in a 50' span portal increases by 40% due to 3/8" movement at the base. However, I think a roller support is a better approximation than a pin if the only restraint is a non-structural floor slab. Even if the restraint is a tension tie, elongation of the tie makes the support less than a pin.

 

[msquared48]

Modelling one support as a roller, put in the tension tie in the frame model and see what happens.

If you get too much spread and moment, increase the section of the tension tie.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[BAretired]

Perhaps a movement of only 3/8" is incredible, but the 40% change in moment is not. With a pin-roller support, the positive moment is WL/8. Dividing that by 1.4, the positive moment with hinged supports is WL/11.2 which means the negative moment is WL/28 all of which seems plausible.

I cannot agree that hairpin bars tied to a slab is more like a roller than a pin, but there is some strain in the tie, so it is somewhere between hinge and roller. If the piles are designed to resist the horizontal force without ties, there will also be lateral strain at the PEMB baseplates.

If the PEMB designer expects the foundation to resist horizontal forces without strain, he is making an unrealistic assumption.

BA

 

[RHTPE]

BA - You are correct about the PEMB designer having perhaps unrealistic expectations of the foundation restraint condition for his columns. But then, they seem to consistently toss that issue to others - their responsibility ends at the bottom of the columns' base plates.

I really wish a PEMB designer would toss in his/her 2 cents. Both from their personal feelings and from the company's position. It would be beneficial to those who do design the foundations to have this insight.

Remember, the 3/8" movement was from a very quick approximation of the building's frame. I would really likt to have the members' fabrication drawings to do a better analysis, but the PEMB folks seem to be reluctant to share.


Ralph
Structures Consulting
Northeast USA

 

[BAretired]

RHTPE,

If the foundation designer uses a tension tie either below the slab or within the slab, the unit strain e in the tie is E/[σ] where [σ] is the stress at working load. For [σ] = 20,000 psi, e = 0.00069 and the total strain for a length of 50' is 0.41" which exceeds 3/8" by ten percent.

I cannot believe that substituting a roller for one hinge would result in a lateral deflection as low as 3/8". That would be one heck of a stiff frame.

BA

 

[hokie66]

Thus why I found it incredible.