Pile Support Condition

[cancmm]

I've created a structural model of a pile-supported bridge-type structure subjected to concentrated, cyclic lateral loads. The lateral load would be applied at a single pile bent and distributed to adjacent bents through the concrete deck. If I model the piles with lateral soil springs and an axially restrained toe, I find that the uplift on a pile local to the concentrated lateral load is quite high. If I use an axial linear spring allowing for ~2" of pile movement, the axial loads become negligible and the load is transferred to adjacent bents more effectively.

So my question... What is the proper way to model this pile? I know that any tension will require some level of displacement to mobilize the skin friction on the pile so the spring seems like a more realistic option. However, I don't want to count on the structure moving before I get capacity because of servicability concerns. There is a big cost difference between the two and I don't want to be overly conservative.

Any input is always appreciated.

 

[ishvaaag]

I recently read an article about the studies made for the foundation of the tallest skyscraper on Earth, and, essentially, they were even candid in acknowledging that they were tuning precisely the spring constant, i.e., quite likely an acknowledgement of expert designers on the current difficulties on getting all encompassing proposals for axial stiffness for piles.

I once studied a square foundation on (floating) piles, grill, mat and footings on elastic halfspace. The differences in settlements were between all cases maybe 10%, never 15%. This means that for this usual hypothesis, magnitude of load, point of application, and, what more meaningful for your case, soil stiffness, rule settlement. So in my view one is in sound practice thinking in a foundation following the expected settlement, and so a softer response that non-yielding support is more in accord to the true behaviour, and a more proper design.

 

[aeoliantexan]

If I understood correctly, you modeled the axial support of the pile with a spring that allowed two inches of axial movement.

For common soils and piles of ordinary size, skin friction is mobilized with small pile movements, perhaps 1/4 inch or less at allowable loads, surely less than 1/2 inch at ultimate. A pile that has moved two inches has failed.

 

[cancmm]

Thanks for the replies.

aeoliantexan, I agree that the 2" is excessive, but I wanted to get a clear understanding of how the structure would behave under a spring support so I exaggerated the deflection to see what would happen.

An elastic spring support (even if at 1/4" deflection maximum) would only be valid if the deflection was elastic - meaning that the pile would rebound back to its original position. Is this logical to assume?

 

[aeoliantexan]

I believe it would be logical if the safety factor is sufficiently high. There must be innumerable friction piles supporting railroad bridges that get loaded to nearly the full design load several times a day. If they were experiencing even small permanent settlement, the cumulative effect would be unacceptable, and the problem would be well known.

 

[DRC1]

Soil spring to model pile movement are a combination of elastic and inelastic springs. Pile diameter to depth and pile properties ar also very important. Water table and duration of and type of loading are significant. To get a good answer, have a geotech run the problem on L-pile. If ypou are in the preliminary stage he can do this quickly and get you an approximate answer. Ifr you have quality soil data and a final design, he can get you an accurate answer.