Modeling of nonlinear pile foundation

[tsodk]

Hi,

I am trying to model a concrete foundation slab, supported by concrete piles. The sole purpose of the model is to find design bending moments in the slab when taking pile deformation under loading into account.

The approach I use has previously worked fine for smaller foundation slabs: The slab is modelled as a panel, which is then nodal supported with nonlinear, bilinear force-displacement supports, calibrated to the given deformation and load bearing capacity of the specific pile using some geotechnical considerations.

Untill now, my models have only considered compression of piles (no uplift from ground water), so i've used a symmetrical force-displacement relation for modeling of the pile support- Although this gave a model stiffness of the pile in tension that was much larger the reality, this method helped the calculation proces converge and all piles still ended up in compression, so no sweat there.

Now, I have a somewhat larger foundation slab with 400+ piles and upwards ground water pressure. The piles have different load bearing capacities in tension and compression - and there is 2 types of piles (one primarily for compression, one primarily for tension). Thus, I've created suitable nonlinear bilinear non-symmetrical force-displacement relations for the supports.

But my nonlinear calculations won't converge, not even for a simple test of uniform surface loading of the slab. Testing with pinned static supports rather than the nonlinear ones yields calculation results, so I guess the problem lies with the combination of geometry, nonlinear supports and the analysis parameters(?).

I've provided the mentioned test model (I'm using RSA Pro 2012). Can anyone point me in the right direction?

Regards,

Thomas

PS: For convenience is here listed the node lists:

"Compression" piles: 

1to183 185to191By2 193to215 225to247 259to274 276to317 319to323 328to340 342   343 347to415 425to436

"Tension" piles:

184to192By2 216to224 248to258 324to327 341 344to346 416to424

All piles: 

1to274 276to317 319to436

[Rafal.Gaweda]

Thomas

I've provided the mentioned test model (I'm using RSA Pro 2012). 

---

Have you attached correct file.

My results below:

[tsodk]

Very strange! Yes, it is the correct test model - I've even downloaded the uploaded version and ran calculations to check that the results still do not converge... which they don't. You haven't used difference analysis parameters?

I can see that i'm having version 25.0.0.3774(x64) installed (on windows 7), but could this really be the reason?

Regards,

Thomas

[Rafal.Gaweda · Accepted solution]

I can see that i'm having version 25.0.0.3774(x64) installed (on windows 7), but could this really be the reason?

Yes, please update this version with SP5.

http://usa.autodesk.com/adsk/servlet/ps/dl/item?siteID=123112&id=18534364&linkID=13140816

[tsodk]

Thanks for the quick assistance: updating RSA made the model results converge.

Thomas

[fenusn]

Hello Mr Rafal Gaweda,

Please, is there any materials, videos, autodesk forum links etc to understanding better on modeling piles as nodal supports (nonlinear, bilinear, force-displacement supports)?

I would also like to understand better, the terms... convergence of nodal reaction, calibration, modeling and identifying piles in tension and compression... Infact, I need to better understand what tsodk is saying.

Thanks.

[fenusn]

Hello Mr Artur Kosakowski,

Please, is there any materials, videos, autodesk forum links etc to understanding better on modeling piles as nodal supports (nonlinear, bilinear, force-displacement supports)?

I would also like to understand better, the terms... convergence of nodal reaction, calibration, modeling and identifying piles in tension and compression... Infact, I need to better understand what tsodk is saying.

Thanks.

[Artur.Kosakowski]

Hi @fenusn

I'd say the first thing t start with is to define how a pile is supposed to work and then convert that into the equivalent support type. So let me ask how is your pile intended to work? 

[fenusn]

Hello Artur,

Many thanks for your reply, and to answer your question, my piles are going to behave in tension and compression.  The major challenge  I'm having (converting these and all) is to come to terms with how the various pile behaviours could be captured in robot.

I actually would appreciate you assist me with resources (forum link, pdf, videos etc., like you have done in the past) to aid me in this regards, and also, if you could give the briefest of explanation to the terms I mentioned (convergence of nodal reaction, calibration, modeling and identifying piles in tension and compression) earlier.

You could be surprised how fast I learn... I ran a nonlinear analysis today... Read it up and did a thorough NL.

Thanks again.

[Artur.Kosakowski]

Hi @fenusn

If you define a (nonlinear) support with the function that describes the behavior of a pile under tension and compression and the nonlinear analysis converges then the value of the reaction in this supported node will give you the information if the pile is in tension or compression. 

If one or more of these posts answered your question, please click Accept as Solution on the posts that helped you so others in the community can find them easily.