piled bases - pile loads + forces (springs-pinned)

[oneintheeye]

as has been discussed many times a lot of people model piles as springs in order to spread load as would truly happen. But what restraints do you use in your model to model moments etc? The spring model or one based on rigid (pinned) supports?

 

[kslee1000]

We usually assume the interface of pile cap and piles is pinned. By either pin or spring, you are going to collect a set of vertical and horizontal loads for use in the design of piles. The reason to use spring model is to study the settlement and lateral displacement to adjust the number of piles and size of pile caps.
Once the pile configuration is set, you can design the cap use the method for spread footings.

 

[oneintheeye]

dont think that is what I asked with repect. I am aware of load evaluation for pile design. I am asking, (just to clear up) in a FEA model do you continue to use your spring analysis in designing the slab or use a pinned condition. A pinned condidition for pile loads will result in inaccurate vastly increased pile loads particarly if concentrated loads are applied, the effect is less pronounced in UDL's.

Hope this clears up any confusion on my query.

 

[kslee1000]

If the slab is cast integrally with the pile and pile cap, then you have to be consistent in using one type of support - pin or spring.

 

[asixth]

I would be interested to see how others respond here. My initial thought is to try and keep your model as consistant all the way through the analysis, so if springs are modelled to analyse the piles, springs should also be modelled when determining slab moments over.

An aside note, my pile design approach has been to model the foundations as pinned. The reactions that are returned by my analysis are then linked into a seperate pile design program that works on the simple broms method for pile analysis.

 

[kslee1000]

Due to the amazing speed and capability of modern day PC & software, I personally prefer to spend some time to put my foundation on a 3D model with spring supports for a test. It, the results, provides better feel of the entire sub system, and understanding of load path and distribution. The most beautiful thing is we gain better control in horizontal movements, if it is of great concerns.

For smaller project with less concern of foundation movement, I too, use pin supports to simplify the design. Actually, I don't think there exists THE CORRECT/INCORRECT way to model/design the foundation system, as long as you have set the loads properly, and taking into account of the subgrade conditions, the system tends to balance itsself, and reach equirbrium ultimately.

Finally, the choice always depends....

 

[oneintheeye]

but what if the two models have differing slab moments that will effect the design?

 

[kslee1000]

If I have time to fool around computer model, I tend to trust the spring model better, because it provides more information, such as foundation settlement and deformation of structures, from which, I can exert certain control by detailing techniques.

The pin model, by neglecting the relative movement, in my opinion, is less accurate. However, it has been done over the years by manys, including me. But in order to ensure integrity/longitivity of the foundation system, understanding on behavior of the soil-structural interface, and sound engineering judgement are indispensable.

If you have time to play both, be conservative (but not overly), envelop the maximum if possible. Otherwise, select one method based on the nature of the project, budget and any other constraints, then be consistant throughtout. Your slab will survive if all load case are considered and applied properly. Also important is specify the subgrade preparation properly.

 

[csd72]

If the pile configuration is simple such as 2 lines of piles and a moment exerting forces on these, then I would expect traditional rigid hand calculations to give accurate results.

If it is more complicated, e.g. odd layout, multiple spans e.t.c. then I would expect a full computer analysis would be more accurate.

I would only bother with springs in the latter case.

 

[kslee1000]

csd is right on the point.

 

[oneintheeye]

yes this is a more complicated base than a simple 2 line case which agreed I would use hand calcs for. I am actually trying to prove an existing base for additional load. I do not have the original calcs. There are over 150 piles in this structure so you understand the computer model. I have proved the pile capacities using spring application. If i model as pinned however and compare the moments in the spans are a lot greater. I suppose as long as my spring coefficent is valid then the forces will distribute as given in spring. Just a bit concerned on the reasons the pinned are much greater in some locations.

 

[csd72]

I would expect lower loads in most cases with the spring supports.

Take for example a double span beam on three piles. On fixed supports the middle one takes about 62% of the load. Now if these are on springs then the middle support would give more and thus more load would transfer to the outer supports.

Infinitely soft springs would transfer the load equally between the three supports.

 

[oneintheeye]

yes but im not trying to determine the pile loads, im lookng at slab moments, my spring stiffness is quite high so im guesing i am conservative anyway but if the piles act as rigid supports im back to my pin model.