Wall friction angle - Seismic case

[calky117]

I'm currently doing a design for an anchored sheetpile (combi) wall. Retained height approx 15m, marine environment so mostly submerged soil conditions, cohesionless granular material.

I've setup spreadsheets using pseudostaic loadings and limit equilibrium analysis. As well I'm performing a WALLAP analysis of the wall to confirm results. When running WALLAP it requires the wall friction angle to be zero during seismic conditions.

When thinking about it, it makes some sense that due to the shaking you can no longer rely on the friction, possibly. Ignoring wall friction makes a large difference to the passive pressures in particular, basically reduces it to rankine. Thus significantly increasing wall embedment.

Any thoughts/advice appreciated.

 

[calky117]

Sorry just to verify, the question is, should wall friction angle be ignored for a seismic load case?

 

[dgillette]

Interesting question. I suppose at one extreme you would have Mononobe-Okabe (basically Coulomb active pressure with horizontal acc. in addition to gravity). Is that what WALLAP uses? I have never seen MO done with water pressures, and I believe the original derivation was for a nonsaturated backfill. I don't recall what MO does with wall friction, and it might have been set to zero in the original for simplicity back in the sliderule era. (I'm home, not close to my files.) A very hasty Google search turned up this paper extending MO to be more general:

http://www.ejge.com/2012/Ppr12.045alr.pdf

It suggests that the conservatism in neglecting wall friction on the active side isn't too awfully large. If a thin water film developed at the contact, the friction would be lost, so maybe it isn't conservative at all to neglect it, with saturated backfill.

At the opposite extreme would be liquefied sand with phi=0 and c={something small}, or acting like a fluid that weighs 120 lb/ft3.

Haven't got a good answer for you - just compounded the question.

 

[calky117]

Thank you for the response, I'm still having trouble with the issue (I'm a graduate marine structures engineer so my geotech knowledge is limited, and our senior geotech guys can't give me an answer on this).

WALLAP lets you input both a horizontal and vertical acceleration, the seismic load case uses 'wedge stability' to compute active and passive pressures, it doesn't use M-O, you are required to manually input Ka and Kp parameters as well. After looking further into WALLAP it appears that you can run the analysis with wall friction if you set the vertical acceleration to zero (this is a limitation of the software rather then a strict requirement). I'm happy to Kv to zero, though it still gives a stern warning when analysing that the wall friction may not be able to be relied upon.

Regarding M-O, yes I do believe the original was for a non-saturated case, as you say there are in fact two issues one being the submerged condition and the other being the wall friction during shaking (dry or saturated, I can't find a good answer). Originally I accepted that the wall friction should be neglected, but it seems all the examples I've seen are allowing its use, either through the M-O equation or modifications of it, or the log-spiral method (I believe is a modified M-O equation as well), the log-spiral charts available in AASHTO, FHWA etc all include a wall friction angle.

For my hand calc/spreadsheet I'm allowing for submerged by following guidelines from USACE - Seismic Design of Waterfront Structures, which modifies the M-O equation and effectively increases Kh using: Khe=(Gamma.dry/Gamma.buoyant)x Kh. Basically doubling the Kh. (this is for 'free water' case, slightly different for a 'restrained water' case).