How to model a tieback wall in a FEA program?

[pootypeters]

I am a new engineer at a structural engineering firm. My first 3 projects have all been walls. First a cantilever, Second a Tieback, Third a soil nail. The 60% design I submitted for the tieback needs to be updated for our next submittal. I did some rudimentary calcs for the tieback wall by hand but am wanting to run a more in depth analysis. The problem is that I'm not quite sure how to model it. At first, it seemed simple. Create a mesh of kirchoff plate elements, apply the loads on the backside (Earth, Water, LLS, Seismic, etc.) and then pin the structure where the anchor heads are located. But as I got to thinking about this, this is not correct. The anchors are post-tensioned to 60kips. I would somehow need to apply a 60 kip force at each of the anchor head locations. The question then is how do I support the structure? Where do I apply the restraints and how? Just to be clear, I am only interested in modeling the wall itself and have no interest in the soil interaction with the tiebacks. That configuration has already been chosen by the geotechnical engineer. I need to get my internal bending moments so that I can design the reinforcing in the wall. Any advice would be greatly appreciated.

 

[Splitrings]

Are these anchors post tensioned after the wall is placed? Tensioned against the face of the wall? If so, I would model my backfill with soil springs and apply a 60K load at your anchor locations. Has your geotech given you the wall displacement required to develop passive soil resistance to resist these post tensioning loads?

I suppose you could model your tiebacks as pins and make sure the reactions at these pins don't exceed 60k. The drawback to this is your model won't correctly indicate the stresses in the front face when tensioning is applied.

 

[TXStructural]

Don't complicate a simple problem. Design the wall as a uniformly-loaded, continuous, one-way beam/slab in the direction of the more closely-spaced anchors. You could consider 2-way effects in the design, but this is probably not needed. You will be applying temperature and shrinkage reinforcement in the perpendicular direction. Use the higher ACI 350 minimums if the wall will be visible or cracking needs to be reduced for corrosion resistance (as would be the case near salted roadways or marine exposure. Reinforcement should be uniformly-spaced bars of one size in each direction to retain economy of construction. Minimizing reinforcement seldom achieves economy if the layout is complicated. Uniformity saves money and time in construction. Where the wall is thick, consider placing part of the T&S in each face (2/3 on the exposed side, 1/3 on the soil side.) You may also end up with two layers of flexural reinforcement, but I would start off expecting one layer, mid-depth, unless the wall is thick.

The anchors should be examined for punching shear, as for columns, and this may determine the wall thickness. Again, uniform thickness will usually keep costs lowest. Formwork and labor is most of the expense of concrete constructon, not the amount of concrete or reinforcement.

The weight of the wall should be placed on a strip footing (or other foundation) to avoid laterally loading the tiebacks. There will be a friction component between wall and soil, but this could vary significantly with changes in soil moisture and with creep of the anchors.

 

[pootypeters]

TXStructural,

Thank you for the detailed information. I did already model it as a two-way slab with column-strips and everything. I was wanting to compare these results with a FEA model to see how accurate or overly-conservative my calculations were. In my opinion, modeling it as a flat plate with the anchor heads treated as columns is not correct since the tiebacks are post-tensioned. I feel as if the stress concentrations around the lock-off zone will be much greater. The geotech gave me loading configurations for the soil, water, seismic, etc. but these were decided independently of the stress in the tie back bar. You are probably correct, I am complicating this problem but I want to feel confident that my design is accurate.

Splitrings,

Yes, post tensioned tieback bar tensioned against some initially poured stressing blocks with steel anchor plates. No, geotech has not given me wall displacements, just loads. I would like to model it with springs but I'm not sure about the stiffness to apply to the springs. I definitely agree with your last comment.

 

[Splitrings]

Your geotech must be able to give you a reasonable estimate for the initial creep of the stressing block into the fill? I would take 60k divided by the estimated displacement of the stressing block for a starting spring value. When you run your model you'll want to make sure your wall displacements at these points don't exceed the initial creep distance you used. You might also want to look a more long term creep behind the stressing block. In the end compare these wall stresses to the plate model stresses and see how they compare. Then you have to use your engineering judgement and comfort level to pick the appropriate design.

 

[IDS]

I am only interested in modeling the wall itself and have no interest in the soil interaction with the tiebacks.

But if you want to know the load distribution on the wall you need to have an interest in the soil interaction with the wall, even if you just model the tiebacks as point loads. You could adequately model this with some "soil springs" if you knew how stiff to make them, but since you don't (and published typical figures can be highly misleading) why not model the soil with plate elements in a 2D plane strain FE analysis?

Or get the geotech engineer to do it and provide you with a load distribution on the wall.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/