Internal Settlement of Rock-Fill Embankments

[Mad Mike]

I'm currently running the geotechnical control on a large-scale industrial development on South Africa's east coast.

The larger development platforms are in the 10 - 20Ha range, developed over a site underlain by hard rock sandstone mantled by sandy soils. Due to the enormous volumes of hard rock, the previous geotechnical consultant incorporated a few rock-fill embankments into the development, which are now being carried over to my involvement.

The rock fill is being constructed by blending sound, blast rock sandstone of acceptable size with the sandy soils and compacting in layers of 750mm using heavy impact plant.

I have drawn the engineer's attention to the fact that rock-fill limits the founding options for the top-structures (which in this case are generally warehouse structures but often with fairly heavy gantries) and he is unfazed, being of the impression that all elements can be founded shallowly on pad footings.

Under normal circumstances I wouldn't be concerned- the fill layers are being engineered to a very high standard. However, the fill thickness in several areas approaches 30m (100ft).

I am now being asked by structural engineers to estimate the settlements beneath shallow foundations, specifically for a platform that has been standing for 1 year, with several good rains having occurred since construction. Note that there are no compressible soils beneath these enormous rock-fill embankments and my gut feel is that internal fill settlement will be negligible.

I would greatly appreciate any advice on fill shrinkage in well engineered rock-fills...I'm aware of all the rules of thumb for settlement estimation but this is not a conventional fill and I'm wary. My recommendation to the engineer / contractor was to monitor the embankment settlement at platform level, which they did, but the readings consistently came back as around 0mm and I have grave doubts as to the accuracy of the survey. I am now in a position where time is of the essence so will have only my own judgement to go by.

Sincerely,
Mike

 

[Mad Mike]

Okiryu, the settlement monitoring was being carried out by survey beacons at various points on the platform margins. I'm not familiar with this work so can't give you specifics- however it was not a geotechnical specialist doing the survey, just some run-of-the-mill surveyor. It was supposed to be accurate to 1mm.

My reason for doubting it is just my gut feel- monitoring at monthly periods showed that an 80ft high earth embankment fill settled less than 1mm. This was within the first couple of months following construction of the embankment and goes against every rule of thumb and geotechnical advice I've ever been given. Had I seen 10 to 20mm of settlement I would have had faith in the result.

Since hearing all the input into this thread, I've advised the civil engineer and developers to avoid using rock-fill beneath any structures, completely. In embankment shoulders, excellent for stability, but beneath structures- unpredictable settlements and cannot be piled through. I'm okay with it beneath conventional warehouses frames and jointed floors, but when it comes to super-flat floors and heavy gantries spanning the cut bedrock / rock-fill, I will not stake my reputation on predicting differential settlements. Finally, I would not be nervous if these embankments were less than about 30ft...100ft is my main reason for concern!

To date, all of the existing structures are built on conventional earth fill. But within the next few months, they will be looking to develop the rock-fill platforms so I'm strategizing in advance!

Sincerely,
Mike

 

[Mad Mike]

Sorry Okiryu,

I meant to add- I've advised the engineers that where the existing rock-fill platforms are to be developed, a dynamic compaction testing programme should be carried out on the finished platform surface. Doing a few heavy compaction tests on a grid pattern should allow us to gauge whether we have a real problem or not. Whether the developer agrees to pay for the testing is another issue entirely!

Mike

 

[Okiryu]

Mike, I think that is important to follow your gut feeling. If you feel comfortable with your recommendations that is what really matters. Let us know how it goes... Good luck!

 

[Okiryu]

Hey Mike, how did the fill go?

 

[Mad Mike]

Hey Okiryu- thanks for checking in- still haven't reached the development stage for any of the high rock-fill platforms. I have taken the conservative approach and informed the developer and engineers that some settlement problems can be expected, not with conventional portal-framed structures, but with the internal gantries and tolerable floor grades i.e. super-flat floors will not be feasible. A new fill-monitoring programme is being initiated on the high fills and will be checked independently for accuracy. No results as yet.

We've come to a fair internal resolution- the engineers will attempt to partition materials on future platforms such that granular earth-fill is favoured within the structural footprint (thus allowing sensitive components to be piled through to bedrock). This will be difficult since the site has very little soil and an excess of rock. Where insufficient earth-fill is available for this, we will use rock-fill up to a predetermined level in the structural footprint, say 5m below final grade, then follow it with a granular fill of sufficient thickness to hopefully "cushion" the structure against any patchy settlement relating to the rock-fill at depth.

Even with this construction approach I will be assuming that some residual settlement will occur (say 0,25% of the fill thickness) and any ultra-sensitive equipment (machinery or laser-racking) would need to either be moved onto the "cut" portion of the platform and founded onto bedrock, or alternatively some very expensive form of deep founding would be required in the rock-fill areas. A dynamic compaction programme and raft/matt foundations are still being considered a possibility for such sensitive structures.

So in conclusion- I still cannot predict with any accuracy the long-term fill shrinkage of a granular rock-fill embankment in which no control testing has been carried out to validate the compaction methodology, however I'm confident that the recommendations given above will assist in minimizing the settlements and from my side I will be sure to disclaim any settlement predictions I am forced to put into writing!

All the best,
Mike