FEM Modelling 1

[InDepth]

The geotech on my project is curently FEM modelling a 35 foot tall 30" diameter secant pile wall system adjacent to an existing structure with pile foundations (within 2 to 3 feet). My design specified 5 levels of tiebacks on the secant pile (so its nice and stiff and minimizes soil dilation to maintain the existing pile soil friction). The geotech on the other hand took the EI stiffness per foot of the secant pile wall and modelled it without tiebacks. i.e. a 35 foot cantilever. The geotech also indicated that the top of this 35 foot cantilever only deflects a 1/4 of an inch. This is with W21x93 spaced at +/-54" OC. He also says he will provide me with the surcharge loads of the piles based on this cantilevered model.

So here are my questions:
1) Is it correct for the geotech to model the secant pile as a cantilever, without the tiebacks? In my mind I understand that he says that if the movement works as a cantilever than the tiebacks are insurance to minimize movement. In my mind, if a tieback is present it attracts load to the shoring and changes the secant pile movement behavior.

2) Is it correct for the geotech to provide a surcharge load based on the cantilever case without tiebacks?

 

[ishvaaag]

The position of the geotech may signify in this case (or other that admits such setup) a point of simplification. Maybe he has seen that it is more practical to design a pure (deeper embedded, or what he interprets needed) cantilever without tiebacks. Anchors are costly in means, manpower, money and time, much times reverted agains to money through cycle of investment or simply interest in borrowed capital. If he is proposing such thing must also be envisaging a scenario with small affection of neighbor properties, at 6 mm horizontal movement.

So except that you by whatever the reasons whish to keep your project for the construction, it might be reasonable to hear their arguments and if you want, further clarifications.

It must also be ascertained if through the change he wants or enters a process where undue differences in cost go -or risk to go- unwarrantedly to their pockets at no benefit at all to the works or the owner, taking amidst your professional stature.

 

[wedlmic]

The walls rigidity will influance the loads - a stiff wall may not defelect enough to mobilize the shear stress in the retained soils. The lateral earth pressure could approach at-rest conditions for a rigid wall. If the wall is reinforced as a cantilever and modeled as a cantilever the model must use the appropriate lateral earth pressure. If it is tied back it will use a different lateral earth pressure (LEP) and apparent pressure envelope. The drilled shaft size and reinforcement determine the EI and deflections. What ever is done, the only thing that matters is what is put in the ground and how it behaves during monitoring. I would not accept a modeling result with a high degree of confidence that does NOT depict what is actually in the ground. Did the model use reinforcement based on a tieback shaft or a cantilevered shaft and what LEP and AEP was assumed? It may "bracket" the expected results but why not model what is being built.

Wedlmic

 

[InDepth]

Wedlmic, that's exactly what I'm saying! Soil Structure Interaction (SSI) like it is built!

 

[ishvaaag]

Of course if you plan to build the wall with tiebacks their analysis is of no use to you nor the owner; only in the matter of a works' change makes sense.

 

[DRC1]

I am not sure why you need an FEM analysis for a cantilever wall. The effect of piles on the wall is highly debatable, but if the piles are close to the wall, it is possible some friction from the piles will transfer to the wall. I would think that 35 feet, even for a secant pile wall, is a bit high for a cantilever wall. On the other hand, 5 tiebacks seems like overkill.
Note that traditionally cantilever walls are designed based on developing full passive pressure. FEM generally uses soil springs. Depending on how good the model is depends on two things. First was the model appropriate? Did it capture the nonlinear nature of the soil or was it just a linear soil spring with subgrade reaction modulus or was it a spring with values that matchede passive pressures? Note that passive pressures develop on both sides of the wall. The other problem is data quality. FEM for wall analysis seems to be greatly by influenced by the quality of the soil data used in the model.
I would suggest being sure everyone has the same soil profile and soil values and then run some hand solutions for a few different support conditions.

 

[fattdad]

gigo. what is the basis for the model parameters? How were the value for modulus determined? Is the model using hyperbolic modulus values, if so, how were these values determined.

There's nothing wrong with this approach, but getting the input data "correct" is where the body is burried. My colleague used this exact method in 1980 or so, so it's proven technology, if properly applied. I think Prof. Duncan may have received his PhD on this at Berkley.

f-d

¡papá gordo ain’t no madre flaca!