RobPE

Geo-Strata has an article in the current issue discussing FEM-based stability modeling. I've never done it myself, but it sure looked interesting!

Jeff

RobPE,

Take a look at,

http://www.ggsd.com/

The webpage itself doesnt provide any real comparisons of software, but it is the most comprehensive listing of geotech software i have come across.  Also you might have some luck looking at the 'WWWlinks' page.

Farqy

RobPE

Virginia Tech has a report issued from their Center for Geotechnical Practice and Research titled "Comparison of Computer Programs for Analysis of Reinforced Slopes" that may be of interest.  Dated December 2000 and costs $50.00.

http://cgpr.cee.vt.edu/

goto research reports then bottom right of page for list of publications.

geotechPE

RobPE - are you talking about different softwares'
"accuracy" against other software for a specific slope using the same analysis method (e.g., Bishop vs Bishop but of G-Slope vs Slope-W)?  I've seen something and will look at it - but are you looking for 'glitches' that might be inherent?  Basically, in my view, if the method of analysis is the same (e.g., Bishop), the differenc from one program to another would be small - and much smaller than, say, how I would ever report the result (I would never report a result beyond 2 decimal places, for instance).  Still, I'll have a look and see if I have anything.

RobPE,

I agree with BigH that results found using different software packages, but using the same procedure, should be similar.  Recently, I was involved in a comparison using UTexas4-Spencer, SlopeW-Spencer, and FLACSlope-Strength Reduction Method.  On this project, all three returned very, very similar results for the same conditions and same failure surface.  What was different was the search mechanism to find the critical failure surface.  For this non-circular failure surface, the limit equilibrium methods required more user effort than expected to get to the critical slip surface.

Thanks for all the valuable references.  I was looking specifically for comparisons of the different method of slices software programs, and the references posted provide good info regarding the pros and cons of several programs.  I have a client that's looking to purchase one of these programs for several engineers and would like an unbiased opinion (other than my own, before making the big purchase.  BTW, I'm not interested in finite element modeling programs.  

Bigh and ntschwanz, I'd like to mention the "optimize critical failure surface" option provided in the latest version of Slope/w, in addition to your comments regarding the search routines for critical failure surfaces.  This option takes critical circles and moves points along the circle until a more critical Fs is calculated.  In doing so, I've found that critical circles can take on a more block shaped geometry with a flat basal plane shearing along a weak and flat lying clay bed.  These geometries can have significantly lower Fs values than the critical circular failure surface.  This is one of the biggest breakthroughs I've seen in method of slices program technology, and choose to use Slope/w for this reason in addition to a few others.

RobPE - GSlope permits you to have "truncated" slip circles - i.e.,, the failure surface is modeled in circular fashion until it hits a hard surface - then it slides along the hard surface until it 'daylights' on the other side (so to speak).  This has been in the DOS versions I have used for years.  Your block-shaped geometry is likely a varoatopm pm a Janbu-style non-circular surface .  I used to have to do these years ago by hand.  Remember that not all analyses are based on the simplified Bishops circular arc.  

Terzaghi and Peck had an intersting non-circular analysis in the 1967 book (Nonveliere - or something like that).  I tried it by hand ONCE and that was enough for me!  The Janbu worked better by hand.  Still, a lot of such programmes only give marginally different factors of safety and I don't look at anything other than 1 decimal point, maybe rounding to nearest 0.05.  I admit, though, the analysis critical failure surface may differ more.  

I am sure that you could "scope in" the critical failure surface almost before you start an analysis anyway based on your experience, the stratigraphy and geometry.  The programs just do a better job that trying to estimate the critical circle centre than does some of the older estimation methods.  On a 286, you could sit and watch the way the analysis looked at the various circles in a circle by circle view - and get a handle on which way things are going in the search - todays computers are just too fast for this.