Finite ELement or Classical Rankine Analysis 4

[cap4000]

I have the latest RAM Advanse 3-D F.E.A. for a Retaining Wall Design. When I compare the results with the typical way of designing a retaining wall with my F.E.A. the results are very different. I am wondering if anyone has found this to be a very conservative approach or is it more realistic. An article by T.C. Goh in 1993 shown in the Braja Das book confirms a much higher lateral pressure value than the classical Rankine Ka active pressure behind a wall fixed at the base to a large footing.

 

[MSEMan]

I think the FE analysis may be telling you that the retaining wall (because of its stiffness) can't yield enough to generate active pressure. You should be able to look at the deflections of the wall stem to check if this may be the case. How do your FE results compare with Ko? Another issue may arise if you are analysing the wall in construction stages - the compaction stresses get "locked in" because the wall doesn't yield enough to relieve them. Classical analysis usually ignores this feature.

Good luck.

 

[cap4000]

MSEMan

Thanks for the response. I have the Goh analysis research paper done with finite element analysis in 1993. He basically says from the top 1/2 or 1/3rd down the wall you should use 1/2 the way between Ka and Ko say 0.4. He does not heavily compact the bacfill for his 12 foot high wall. His results look pretty accurrate as does mine with RAM 7.0. The real trick is which method do you use. Either Rankine, Coulomb, EFP or At-Rest and how the failure wedges develop directly behind the wall. I like for years the IES Quick R Wall software program that uses simple statics like AASHTO and CRSI and does a real professional looking job.

 

[fndn]

Cap4000- In my experience, Walls of height less than 18 feet tall designed with the Equivalent Fluid Pressure method performs well if the backfill is granular and adequate drainage is provided. Rankine method is very conservative since it ignores wall-soil friction. Since your wall is fixed at the bottom and not restrained at the top, I would lean on Active pressure numbers.

If you want to use the failure wedge theory and your wall is long, you'll have less steel and be able to produce economical design. As for compacting the bakfill, I would not be concerned with a 4 foot roller that has the vibratory turned off and 9" lifts of granular backfill. To learn more click on

http://www.usace.army.mil/inet/usace-docs/armytm/tm5-818-1/ch-14.pdf

 

[cap4000]

fndn

Appreciate the nice link. Personally I think the EFP method is the "quick & dirty" too easily understood way to analyse a retaining wall. I lean towards the Coulombs method since most of my footing heels are not that long and the wall friction clearly helps the stem design to be economical as you indicate. All that said, I do check all the different methods anyway and make a choice comparing all the results to the CRSI Tables. The F.E.A. method highlights the complex stiffness, rotation, deflection and translations a retaining wall actually goes through.

 

[DRC1]

FE is based on soil modulus, where as classical analysis is based on lateral pressure coefficients. Although not apple and oranges, it is somewhat of tangerine oranges approach. The problem is hat for FE to work well, quality data is needed. Lateral pressure coefficents have been around so long that given some familiarity of the soil, reasonable values can be assumed even with marginal information. Actual pressures require sufficent deflection and rotation of the wall to move the soil wedge behind the wall. This is generally permissable for flexible walls such as sheet piling, but not for rigid walls such as permenant concrete walls. These are better designed with at rest to prevent excessive wall deflection.

 

[cap4000]

DRC1

Your absolutely right about quality soil data. Every retaining wall job I have ever done working for an Architect or an owner,I get no soils info/report at all much less a soil modulus. I just indicate the required soil/rock conditions I used for the design. I think that Goh and Casagrande had it right when they say, the typical "safety factors" used for overturning, sliding, shear and stem moments are enough to keep the wall from actually failing as the pressures developed behind the wall clearly exceed the classical Ka values.