Secondary settlement / heave

[jpoha]

I have a question concerning secondary compression and secondary heave. I have a deep structure (approximately 35 ft. deep) to be built on a site with 20 ft. of sand and gravel over up to 80 ft. of compressible soils. This will result in a net unloading of the soil.

Based on the consolidation test results, I have calculated that during excavation, the soil will heave past the point where it will end up, once the structure is built and backfilled. I call this the "final resting place". At some point during construction, the soil will stop heaving and start settling back the "final resting place".

One fellow engineer argues that since the construction will result in a net unloading, once primary settlement is finished, the secondary movement will be a heave.

I argue that since the last primary movement is a settlement, the secondary movement will be a settlement.

Any thoughts?

 

[GeoPaveTraffic]

In theory, I agree with you jpoha; that since the final movement is settlement that any secondary movement will also be settlement. However, it is very likely that any movement once the structure is started will be very small and hard, if not impossible, to detect. Also, remember that one dimension consolidation test results only approximate what is going to occur in the field.

 

[dgillette]

Interesting question! I don't recall seeing anything like this (secondary de-consolidation in unloading) in the literature. Have you done oedometer tests with unload-reload, and a long periods of holding for the secondary to show up?

It may be that at the field scale, the primary consolidation is slow enough that the secondary gets lost in the decimal dust. That is, the dissipation of pore pressure is slow enough that the effective stress is applied to the soil very slowly (over months, rather than minutes or hours as in the oedometer), so the secondary can, in a sense, keep up with the increase in effective stress. Is this a real concern for the final resting place with an overconsolidated clay, or is it intellectual curiosity? With "forgiving" structures, like highway embankments, secondary is generally neglected (quite reasonably), except sometimes with organics.

You got me curious, too.

DRG

 

[oldestguy]

Has any one ever experienced such a thing? No me and I have been involved with many such sites.

It would be my guess any "heave" occurs before the structure is built

Then too, once backfill is in place the structure is held down and, in this case, that backfill may cause settlement due to loosening of subsoils due to dewatering.

The closest I have seen with heave is unloading a site that has been surcharged and some "rebound" has been measured.

 

[DRC1]

Heave in a deep excavation is actually a bearing failure. Elastic rebound does occur and can be theoretically calculated as an elastic response using primary conslodition characterisics. Once it has unloaded, it will begin reconsloidating. However, many factors will effect it including the depth of the sheets, disruption and grading during construction and changes in drianage paths. Note that in clays, the bottoms are generally over excvated (wether called for or not) and backfilled with about a foot of stone. I have never seen where the rebound of the soil has had any measurable impact on the grade of the structure.

 

[fattdad]

If the soil is pre-consolidated and acting on the reconsolidation curve, you may be just o.k. looking at this from the soil modulus perspective (i.e., elastic theory). I would expect the extent to which there will be vertical upward movement during excavation will occur with the offloading (i.e., negligable time lag). With the application of new loads, I would then use bousinesq stress distribution divide the subgrade into layers with depth and then calculate the elastic compression that would likely be realized. Bear in mind that the modulus will increase with depth and the stresses will dissipate with depth, so the seat of settlement will be limited by the size of the structure.

Now that I've said all of this, I'd also look at the 1-d reconsolidation index and calculate settlement that way also. You may want to look at strain distribution; however if you do this depending on how limited the size of the structure. Mike Duncan's settlement manual (Virginia Tech) has information on this.