Long Term Stability - Drained vs Undrained

It makes a difference if you have normally or lightly overconsolidated soils (where pore pressures are induced and then with drainage, p' will increase (i.e., you gain strength) or if you have heavily overconsolidated soils where upon "drainage" which is really a suction of water into the soil due to the negative pore pressures induced, the soil will lose strength with time. Normally, in my experience, with the former, the short term or undrained strength is the most critical and that is why most times, one doesn't carry out effective stress analyses (for in the short term it is quite difficult to estimate the induced porewater pressures). For the latter case, the effective stress analysis is crucial.

In addition to what BigH has stated, I would ask the following:

1. What are the details of the analysis?
2. Are you adding material to an existing slope?
3. Are you cutting material from a slope?
4. Are you constructing a water retaining structure?

These questions to the point of how the load/stress on/in the slope are changing. If there is no change in stress and has been no change in stress for a long time, then drained strengths are going to control becuase any drainage has already occured. However, if you are doing something to increase the stress in the soil, then either drained or undrained strength may control depending on the specifics of the problem.

Mike Lambert

To: GeoPaveTraffic
Thanks. Your questions relate to short term (changes) conditions, but I specifically asked about the long term stability assessment where pore water pressure corresponds to steady seepage conditions or in your words “there has been no change in stress for a long time”.

To: BigH
Thanks for your response. I still don’t get it both formally and mentally . The US Society on Dams “Strength of Materials for Embankment Dams” ( USACE Publications such as “Stability of Earth and Rock-Fill Dams”, Duncan, J., & Wright, S. (2005) “Soil Strength and Slope Stability” all direct to use drained shear strength for the long term stability assessment. Are they all incorrect or not conservative? Moreover, something tells me that once drainage occurred and there is no excess pore water pressure, the initiation of the slope failure must be in drained manner. Can this be the real reason behind the drained strength approach?

Your responses indicate the need to understand where loading of the slope comes from. You need to identify what is causing the change in stress. If there is no change in stress there is no failure.

I think you need to post information on what kind of structure you are looking at and the history of the structure. Your questions are general enough that there can be many different answers depending on the assumptions someone makes.

Mike Lambert

CEMAB - you should take a look at stress paths (see Lambe and Whitman, for instance) . . . you have the "drained" line which all loading will eventually head towards - the normally consolidated and lightly overconsolidated soils will "drain" towards the right - or towards the drained line . . . hence, the short term stability is more critical than the long term stability. If the short term stability is safe, the long term will also be safe. For Heavily overconsolidated soils, the "drainage" - i.e., the sucking in of water - will move the left towards the drained line - hence, the short term strength is less than the long term strength. Once porewater pressures have equalized - no more water expulsion or water absorption - the drained condition governs. But until that time is reached, you will have water movement (out or in) and you need to consider that as well. And, what does "long term" mean? In what you are describing, it means the drained (no water movement) condition.

To: BigH

I appreciate your answer and thanks for the reference. Just clarify a few points. Here is the definition given by Duncan and Wright “Soil Strength and Slope Stability” (2005) on short term conditions: “Short term refers to conditions during or following construction—the time immediately following the change in load.” The long term conditions described as: “After a period of time, the clay foundation would reach a drained condition, and the drained analysis would be performed because long term and drained conditions carry exactly the same meaning. Both of these terms refer to the condition where drainage equilibrium has been reached and there are no excess pore pressures due to external loads.” This is how I see the definition of the long term conditions as well. In other words: If the slope didn’t fail in a short term following an introduction of any change in load/stress and drainage equilibrium is reached after a while, the slope is under the "long-term" conditions and should be checked for failure in a drained manner.

CEMAB - the quotation is almost exactly what I had intimated . . . It seems to me that Duncan and Wright's comments are for normally and lightly overconsolidated foundation soil. It would not necessarily apply to a heavily overconsolidated soil where the shear stresses might be higher than where the drained stress path establishes the failure criterion shear strength - and hence, with suction of water into the soil, it will move towards the drained stress path - but hit the failure criterion before it can be fully equalized.

Another point to consider is whether you are discussing first time "failure" or subsequent where the residual strength may be at play.

Thanks BigH. The thing that is still confusing me is that none of the sources I’ve read including Duncan and Wright attempts to differentiate between OC, NC or Slightly OC materials for long term stability purposes. The guidelines are clear to use drained strength for all materials. In any case, I’ll try to do more research on the topic in a few next weeks. Thanks for the interesting discussion.

What BigH is describing is the basis of Critical State Soil Mechanics. For simplicity separate pore-water pressure generation into 2 parts, the first being related to compression/rebound and second related to shear. If you apply an all around pressure to a fine grained soil sample, like in an isotropically consolidated triaxial test, it will change volume but no shear is induced (some will argue differently on this) and you can picture that pore-water pressure may take time to dissipate. Now if that specimen is normally to lightly overconsolidated and you load it rapidly in pure shear, with drainage closed, it will likely generate positive pore-water pressure. If the specimen is highly overconsolidated, then negative pore-water pressures will be generated as BigH describes.

Now think of an embankment, say a levee, that has been in place for many years so that all pore-water pressures from consolidation and shear have dissipated under existing conditions (long-term condition). A flood comes along and rapidly loads the levee and foundation. If the levee and foundation are normally to lightly overconsolidated they can still develop pore-water pressure due to shear and so the shear strength would be based on undrained response related to pre-flood effective stresses. This example is a long term condition that is suddenly loaded, but what if you don't have a change in load, can you have an undrained soil response? Probably not, and you would design using drained strengths. Steady state seepage conditions (no excess pore-water pressures associated with compression or shear) is synonimous with drained shear strengths. Now, you may want to look at the assumption of no loading closely to be sure that assumption is correct as slope instability can happen fast so that shear induced pore-water pressures may not dissipate (think of static liquifaction of mine tailings that have a collapsing structure).