Let's discuss infinite slope boundary conditions 1

[fattdad]

Hi Eng-Tips:

A recent failure returned a consultant report. The failure on a typical bridge approach and the work was less than a year or two old. The consultant called the failure, "Shallow, sloughing failure" and told the contractor how to fix it. Contractor and consultant opined, "Well, of course it failed! We had four rain events in August!"

I'm the annoying fellow that then had to ask, "Just how many rain events did you consider in design?"

Well, I know the answer!

Returning to the original design, we captured the soil properties used in preliminary design. PHi=30, c=50. For the geometries, all was well in global stability. For the infinite slope equation (the form with cohesion), all was well, but. . .

When we considered seepage parallel to the slope, things went wonky. For top-down saturation to depth 18-inches, safety factor went to unity.

Bam! Just like the field observations.

So, to all - How many of you all set water filled tension cracks in infinite slope design? Cracks to what depth? We don't use cracks, we just make them consider a few feet of top-down saturation?

LA county requires 4 ft of top-down saturation!

Clearly, if it's free-draining embankment this is moot.

Seems to me, we are overlooking this area of slope stability and it's hitting our maintenance budgets!

f-d

ípapß gordo ainÆt no madre flaca!

 

[GeoEnvGuy]

I work in an area where we have blasted rock around all our abutments. The people here dont mind the look given the next kilometer or road is all blasted cuts and fills.

From the information provided a cu of 50 is being relied upon in a phi 30 material. In my practice I do not assume cohesive strength will develop in frictional materials. If the design was done with only frictional strength I doubt you would of had an issue, but it would have cost more to flatten the slope.

Working in dams I also do not consider the increased strength from lower effective stress as a benefit in the analysis, which leads to conservative slope angles compared to the infrastructure designs i see.

 

[fattdad]

Not quite sure how to respond. . .

Yes, if there is no cohesion the form of the equation is different. I'm not asking about that.

I also understand conservative design. It's competes; however. I'm sitting in the owner's seat. I can't just say, "Can't use cohesion!" Generally accepted practice allows it - refer to FHWA, etc. I can prescribe fully-softened strength. I can set boundary conditions - i.e., top-down to depth of 3 or 4 ft, etc.

Perhaps other's understand this topic?

f-d

ípapß gordo ainÆt no madre flaca!

 

[GeoPaveTraffic]

fattdad,

I have to agree with GeoEnvGuy, a strength of phi=30 and Cu=50 psf is high. As the owner, you can forbid the use of cohesion.

An easier answer may be to require the use of fully softened slopes, that would likly solve most of the problems.

Have you really studied the slope stability analysis that was performed? I've reviewed 100s of these analysis and most have a problem that the engineer just ingorned. This is a really big problem with GeoStudio's Slope/w because of the way it presents the results.

As for the owner's maintainence budgets; what slope do you use the abutment? I see lots of road abutments with 1.5:1 slopes, no soil is going to stand like that on it's own indefinately. Railroads consider it cheaper to just repair as needed instead of making all slopes flatter.

Mike Lambert

 

[GeoEnvGuy]

I believe that we should be using the

Can't use cohesion!

design line forced through zero on the "S" envelope.

I also believe in slope stability analysis today there is a common belief the triaxial compression test results are what should be used without the consideration for DSS or TE results.

 

[fattdad]

somebody will ultimately get it!

[ul]
[li]We use 2:1 slopes and require engineerings design.[/li]
[li]Until you both run a few numbers it'll be hard to see the problem. This is not a modeling exercise. That equation is older than my parchment![/li]
[li]You cannot get infinite slope stability over 1.3 with just 30 degrees on a 2H:1v slope - some cohesion would be required and then the problem goes away. . . unless there's top-down saturation.[/li]
[li]Cohesion in most txc doesn't reflect the low confinement of the infinite slope stresses. DDS at 200 and 400 would; however.[/li]
[li]Some seepage parallel the slope will develop - it's just how unsaturated permeability works![/li]
[/ul]

If these are topics you don't have direct experiences in managing, that's fine. Saying you can't use cohesion is not what's going to happen; however.

f-d

ípapß gordo ainÆt no madre flaca!

 

[GeoPaveTraffic]

I have to disagree with you fattdad. Your problem is you want your cake and you want to eat too. You are specifiying a 2:1 slope and a FOS of 1.3 and you are telling the designers that this should work, because we have been using it for years and it USUALLY works. I can set up a stability model and get any FOS you want and most engineers would not be able to tell there that the analysis has been faked/fudged without looking at the nitty gritty details; which is rarely submitted.

Mike Lambert

 

[EireChch]

Personally I have never used more than 5-10kPa in stability analysis. I am afraid! When I see c' of 50kPa I automatically think it must be a HOC clay, I would then think that Phi should be lower like 22-25 deg? Thats just based on my experience in soils in my areas of work. c' of 50kPa is a shit load of cohesion!!

Can you explain how you consider top-down saturation in the stability analysis? Are you assuming the upper 18 inches is saturated so failure is in an undrained manner? This layer considers undrained parameters and the rest drained?

Or is groundwater at 18 inches depth?

 

[geotechguy1]

It's an unsaturated soil mechanics problem, and in practice everyone uses saturated soil mechanics to solve slope stability problems but we neglect the word saturated and forget the limitations. The slope is stable in the infinite slope case likely because of suction / negative porewater pressure effects increasing the strength, but this effect goes away under heavy rainfall.

Curious to know what the recommended fix was.

 

[fattdad]

c=50 psf. Sorry about the labels.

I do not want my cake and eat it too. Unsure how that thought even popped up! Clearly, if the correct boundary conditions are in use, some secondary geogrid could address the failure mode.

One inch of infiltration can soak 4 to 8 inches of unsaturated soil. Back-to-back rains can provide further flux. Unsatured permeability is lower than saturated permeability. Unsaturated strength is higher than saturated strength. I get these things.

Shrinkage cracks form. Water enters, stresses change and we have failures.

Should the owner define depth to shrinkage cracks? Should the owner define saturated depth?

How deep can the infinite slope actually go - my benchmark is L/20, where L's the hypotenuse.

We can model this in Slide by setting Ru to appropriate value. We can replicate those results in model via shrinkage cracks.

This is not a glib conversation. Unsure of some folks reactions, that is unless you are not familiar with the equation.

f-d

ípapß gordo ainÆt no madre flaca!

 

[MTNClimber]

F-d - are these failures common in your area? If so, the standard of practice needs to be updated to include consideration of shrinkage cracks or saturation of upper layers. Most of the time these types of updates in practice are spearheaded by the local DOT or another local government entity. I definitely think requiring these considerations in slope stability analyses are justifiable by an owner if failures have become too common.

Just be careful of what you require and be specific. Maybe do similar back calculations of other failures you guys have seen in order to come up with some reasonable requirements. I can tell you from experience that some local codes are crazy. The Denver local code requires a fire truck loading for SOE designs. The loading is absurd (18.5 ksf!) without making some interpretations of the code.

 

[fattdad]

Thanks MTNClimber.

We have been back-calculating on a recent failure and begin to see the issue. There was/is actually no considering in geotechnical design (that I see) for any measure of surface saturation in infinite slope design. It's almost a design oversight.

Our policy says design for all modes of failure, so there is a contract obligation to run some numbers. If those numbers do not include prescribed saturation, then we get no assessment.

Clearly, this is related to the free-draining nature, or lack thereof, of the embankment fill.

What we see in parametric study is safety factors plummet when saturation depth goes from zero to 4 ft. We did increments of 0.5 ft and after 4 ft it's asymptotic.

Then too, what defines an infinite slope? Prof. Duncan told me that it's L/20. So, if you have a 60-ft long hypotenuse length, the infinite slope would mobilize the top 3 ft. After all at some depth, the toe buttress comes into play and it's not free to be infinite.

In Virginia, we have slope lengths of hundreds of feet! No joke, what's an infinite slope when the length is 300 ft! That'd be the upper 15 ft!

I'm assembling specifications for a d-b project right now and imposing 3 ft of top-down saturation. This is a project in limestone and the residuum is all clay with limited sand. Borderline CL/CH and ML/MH. They want to build embankments with this stuff and I get it. Don't want to see it fail, even if I'm retiring!

f-d

ípapß gordo ainÆt no madre flaca!

 

[GeoEnvGuy]

I believe if you specify to design 2H:1V slopes and accommodate 3 feet of top down saturation you will get the designer to show you reinforcement of 3 feet parallel to the slope. If you were to specify 4 ft the designer will reinforce 4 ft and so on and so on.

If you have full saturation with parallel flow lines to the ground you will require 60 to 65% of Height reinforced parallel to the slope or until you get to the 0.5 tan(phi)/tan(slope) angle.

 

[fattdad]

Hi GeoEnvGuy:

It'd be more than those lengths of reinforcement. We are in the middle of a parametric study. We are also submitting a research topic to the TRB.

I'm actually looking for comments related to folks that are in this area of engineering.

I'm thankful you are interested; however.

f-d

ípapß gordo ainÆt no madre flaca!

 

[geotechguy1]

As an alternative to a geogrid reinforcement impermeable surface coverings can be used to prevent the slope from becoming saturated in which case the strength due to soil suction effects can be retained. A shotcrete or soil-lime mix can be used or some other form of impermeable layer.

 

[fattdad]

yes, and impermeable surface, that doesn't desiccate, would be one approach.

Thanks!

f-d

ípapß gordo ainÆt no madre flaca!