Drilled Pier Foundation - TIA & ACI

[azcats]

I'm working a lot with the analysis of existing and design of new drilled pier foundations for telecom tower footings. Very light axial loads and relatively high lateral loads. The foundations are typically modeled in LPile.

The foundation section of the tower code, TIA-222 (Section 9.4 for Rev G & 9.7 for Rev H) has a Φs factor for design. For piers this factor is 0.75. The corresponding note says, "For foundation analyses which model the lateral stiffness of the soil, factored reactions for the analysis shall be divided by Φs."

ACI 336.3R-14 in section 4.1 states, "Normally, service loads are used to calculate the resulting moments, shears and axial forces, which are then multiplied by the appropriate load factors for the various cases of loading for structural design of the pier."

The two approaches seem to contradict each other. I'm not sure I agree with modeling soil structure interaction using factored loads that area again factored up. What useful information does this model provide? Deflections are meaningless. Forces in the member may not be accurately represented as the soils will behave differently under different magnitudes of loading.

If I were not working within this TIA code, I would model as the ACI document indicates and factor resulting forces for RC pier design.

For now, my approach is to model with (3) load cases: factored-factored loads, unfactored loads, and TIA service loads (lower wind speed). I take the member forces from the unfactored loads, factor them up (including the Φs), then compare to the output from the factored-factored load case and use the worst for design.

As for deflections, Rev. G indicates that deflections can be ignored if the TIA service load case at grade deflections are less than 0.75" which is typically the case. The factored load deflections are unreasonable and often multiple inches.

Does my approach seem reasonable? Anything else any of you would consider?

 

[BridgeSmith]

If ACI's load factors are similar to those in AASHTO (1.4 for wind), the result would be similar whether you increase the loads (ACI) or reduce the capacity (TIA) to get the factor of safety. It's still a FOS of approximately 1.4 (1.33 for the TIA approach).

 

[Settingsun]

Forces in the member may not be accurately represented as the soils will behave differently under different magnitudes of loading.

If I were not working within this TIA code, I would model as the ACI document indicates and factor resulting forces for RC pier design.

Soils behaving differently at different load magnitudes is why neither service/ultimate load analysis is necessarily more correct than the other.

I believe that the 'reactions' referred to in TIA are actually the applied load in the Lpile analysis. Do you agree? Ie they assume you've done an analysis of the above-ground structure with a spring (or something) representing the drilled pier, and then are transferring the reaction from that analysis to the Lpile model.

In that case, what I think they're getting at is that Lpile is a soil-structure interaction analysis and not a limit-equilibrium analysis. Lpile doesn't tell you the ultimate geotechnical capacity of the foundation unless you increase the load until it fails to converge. So, to demonstrate that you have a geotechnical safety margin, you have to apply a larger load than the usual factored load. Essentially moving the phi,s factor from the capacity side to the load side as 1/phi,s. On the other hand, if you use Broms/Brinch-Hansen/other, you get an ultimate geotechnical capacity (independent of the applied load) which you can multiply by phi,s.

The TIA method of then multiplying the bending moment from Lpile by phi,s is how they've chosen to arrive at a single number for design. I personally prefer to do some sensitivity checks and make a judgment - I'm not a fan of 'hard-coded' capacity factors for geotechnical work. It's too variable compared with manufactured materials like steel and concrete. If you've done a gold-plated site investigation and lab testing program and your structure presents low risk, you should be able to use a higher capacity factor than a small-scale investigation with no lab analysis and a more-critical structure. Australian Standard 'AS 2159' is on the right track in this regard IMO.

For now, my approach is to model with (3) load cases: factored-factored loads, unfactored loads, and TIA service loads (lower wind speed). I take the member forces from the unfactored loads, factor them up (including the Φs), then compare to the output from the factored-factored load case and use the worst for design.

In the bit I've underlined, what factor would you use? I'd use ~1.5. I don't think you need to consider the phi,s in this because the TIA code tries to cancel it out: you increase the applied load by 1/phi,s then reduce the resulting bending moment by multiplying by phi,s.

And you don't need to check ultimate/factored deflections unless there's a serious problem that would result from the excessive deflection like a supported structural element would fall off (ie an ultimate limit state stability failure).

Overall, sounds reasonable. You've ticked the code compliance box and also done a sensitivity check of sorts.

 

[BridgeSmith]

I agree with steveh49. If you're using a P-y analysis program, such as Lpile, factoring the loads is the better approach, which is what both methods are doing as you've presented them. They just have different load factors. Some methods put an increase on the design wind speed and leave the other factors as 1.0 (ASD method from AASHTO Standard spec for Signs, Luminaires and Traffic Signals).

 

[azcats]

Please excuse my delay in following back up here. I appreciate the input from both of you.

In my hastiness, I didn't fully realize that the TIA phi factor was backed out when determining pier capacities.

I was also somewhat frustrated at analyses that showed relatively huge deflections and called them "OK". As it's not particularly OK, but not applicable. The deflections really multiply from the TIA service load to the factored loads. In the one on my screen now, the service load deflection is .011" and the ultimate is .57" where the lateral load is about 5x larger.

In the couple of piers I've modeled since my original post, I've found that my two approaches yield almost identical moments in the pier and it would probably take some odd soil condition where the factored load didn't converge for this method to really do anything.

I believe that the 'reactions' referred to in TIA are actually the applied load in the Lpile analysis. Do you agree?

Yes.

In the bit I've underlined, what factor would you use?

I would factor them up to ultimate load factors for concrete pier design. So 1.0 or 1.6 depending on the code and wind load development. Would you put an additional 1.5 on top?

 

[Settingsun]

Since soils are almost all strain-softening, the factor-up/factor-down analysis should give a larger bending moment.

1.5 factor was my estimate at the overall load factor, not an extra factor. The old factor of safety eg allowable stress in steel design. I'm not old enough to know how allowable concrete design worked.

Why is your ultimate load 5x higher than service load? Incidentally, non-linear soil response is why I think limit state design is worthwhile for geotechnical work. But it does need to be thought through better than a lot of codes do. And gets complex when the ground is both load and support but that's not an issue here.