Impact rolling clay fill as alternative to normal sheep-foot roller compaction

[Mad Mike]

Hi All,

I have a modest-sized earthworks site where the full depth of the proposed earthworks (cutting 10'; filling 10') is in a silty clay residual dolerite soil. The site is very gently sloping so the volumes are larger than the depths of the earthworks would suggest.

The clay soils are of low plasticity (PI < 15) but clay content relatively high (30 - 50%) and clay + silt content of 70%. So very little sand and gravel.

The soils primarily classify as ML in terms of Unified System and they are not highly expansive. Swell in the CBR mould is approximately 0.5%.

The proposed development on the platform is a large warehouse that the engineer intends to support on a mat foundation (effectively a beam-stiffened slab) and I gave a recommendation in my report that the clay should be compacted using a sheep's-foot roller, to a compaction range of 93 to 95% Mod. AASHTO. Construction has now started and the Contractor is pushing an alternative compaction method of "rolling dynamic compaction" or what we here call "impact rolling" (dragging the 3, 4 or 5-sided impact roller), compacting the clay in 500mm layers. He has produced density results showing 100% compaction (Mod. AASHTO) of the clay.

I am not happy with a clay fill being compacted in this method, but can find very few articles online to support my views. What are the experiences here with compacting low-expansive clay fill by impact? My main concern is that we could see re-bound of the clay after construction, or that even slight excess of moisture could jeopardise the compaction. Any thoughts or good references?

Best,
Mike

 

[zelgar]

I would have the following questions:
1. With the method of compaction, how are they incorporating the different lift together?
2. Are they achieving compaction within the entire lift?
3. Are they measuring the compaction of the entire lift or just the surface? About 25 years ago, I had an engineer who was meeting their density requirement using the nuclear densometer but couldn't verify the result using the sand cone method. It was found out that they were not using the nuclear test properly (they were using it in back-scatter mode) and were only measuring the effects of the upper surface of the lift. The sand cone method, while destructive, showed that the entire lift was not being compacted.

I would be concerned that the method they're using would not make a uniform support for the foundation. The sheepsfoot roller incorporates the new lift with the lower lift to make a uniform foundation. I would think their rolling dynamic compaction would result in multiple lifts that are not incorporated together.

Additionally, I would wonder if they're getting compaction throughout the lift. Are they testing the entire lift, or only the upper surface. Do the lifts need to be smaller using this type of equipment in order to achieve the required compaction?

 

[Mad Mike]

Thanks Zelgar,

1. lifts are 400 to 450mm. When compacted with the impact roller, this would produce a relatively smooth clay surface, much like you'd get from a smooth-drum roller.
2. Not yet confirmed. See point 3.
3. No - they've tested to a 300mm depth on their 400 to 450mm layers. You're probably aware that the impact roller has a substantial depth influence, at least 1m (3 to 4 feet) and I wouldn't expect any issues on the soil layer in the 300 to 450mm depth range which hasn't been tested.

Given that this is a relatively low fill embankment, I'm not too concerned about fill settlement. 450mm layers are very thin for an impact roller, with the effect that each layer ultimately gets compacted twice.

My biggest concern lies with over-compacting the clay soil and potential re-bound. Settlement would be a concern if it was a high fill, but not for just 10'.

As for moisture content, we seem to be at approximately optimum (20 to 22%) in the field which is still well below the plastic limit of the clay (> 30%). Pore pressures therefore unlikely to be an issue.

Best,
Mike

 

[EireChch]

I think you are over worrying Mike, I would not be concerned at all. Sheep foot rollers are seldom used in the UK as the soils are often too wet. Most clay is compacted with a smooth drum roller, no vibration.

I am surprised you are getting such good results with the impact roller as it is mainly used for granular soils. But the impact is not that extreme (compared to the three sided rollers that it is working similar to a smooth drum, its just self weight eh.

I have never hear of over compacting and rebound being something that needs to be considered!

Sounds like the contractor knows what they are doing tbh

 

[iandig]

I know there is great reliance on degree of compaction and air voids as a means of end product testing of fill, but ultimately the critical attributes for the fill would be by their end performance.
I have used both rolling dynamic compaction and high energy impact compaction (e.g. Bomag BW 226) to place and compact cohesive fills, but we would always include both end product (density, moisture, air voids) with end performance testing(subgrade stiffness) as part of the performance requirements to be met.
The equipment I would specify for the deep compaction would be expected to be fitted with continuous surface stiffness monitoring equipment. What we find is that as the compaction is carried out, whilst density increases, we start to get a reduction in stiffness as the fill becomes over-stressed. This is managed by using the inbuilt stiffness monitors to assess the ground stiffness before the fist phase of treatment, monitor the stiffness increase as the ground is treated, and suspend work if there is a reduction in the ground performance. You can then leave the ground to recover, and then re-assess once a suitable period of time has passed. This helps develop the best method of treatment of the fill. It also means you need to consider when testing is undertaken, as if you were to carry out say a plate load test immediately after a layer was compacted, the ground is likely to appear less stiff than if you tested it 24 hours later.
I have also found that the surface of the fill can appear to have a corrugated form, so would normally specify the final layer is re-compacted using a 'traditional' roller to provide a uniform surface.
I have also used the equipment for ground improvement of deeper cohesive fills, but this was coupled with the installation of pre-fabricated vertical drains (PVD, as if the site was to undergo preloading and surcharging) with the PVD used to accelerate the excess pore pressure, as part of a pashed ground improvement treatment. Depths up to 4m were treated in this way.

 

[iandig]

It should say 'accelerate the DISSIPATION' OF EXCESS PORE PRESSURE