Research by Pavement Technology is looking at sub-base and capping improvement using geocells. Jack Bull reports.
Geogrid reinforcement is a well-established technique for improving the performance of granular capping layers and sub-bases over soft ground.A principal benefit is improved deformation resistance with the corresponding opportunity to reduce sub-base thickness.
The use of three-dimensional reinforcement, commonly referred to as geocells, has interested researchers as an alternative to a two-dimensional geogrid.
Geocells typically consist of a mattress of interlocking cells formed from strips of polyethylene that are either stitched or welded together and filled with granular material. Not only does this give a stiffer foundation, but uses recycled or site-won fill materials that would otherwise not be suitable for pavement construction.
Field trials on unmade roads by the US Army Engineer Waterways Research Station in the late 1970s, for example Mitchell et al (1979), established that geocells can provide significantly improved load capacity compared with unreinforced soil.
Since then, geocells have found applications in embankment stabilisation, rail track ballast improvement, retaining walls and bearing capacity improvement under footings. But despite the possible potential, research work specific to paved roads appears to be limited, until now.
Bringing together the analytical and research capabilities of the Advanced Pavement Technology Centre at Heriot-Watt with the commercial expertise in pavement condition assessment and performance testing of Bureau Veritas Consulting, Pavement Technology is carrying out laboratory and field trials.
Work is concentrating on reinforcement for pavement sub-bases using secondary or recycled aggregates.The approach is to consider the available performance from aggregates in their 'as-won' condition to minimise the cost of extra screening, crushing and other processing.
Geocell behaviour In their work on unmade roads with geosynthetic-reinforcement, Dawson et al (1994) concluded there was no simple relationship between resilient deformation, permanent deformation and static strength.
Lekarp and Isaacson's (2000) state of the art review of permanent deformation in unreinforced granular sub-bases listed a number of factors influencing sub-base performance.
Adding geocell reinforcement increases the number of variables further:
l Applied stress level l Number of load applications l Fill moisture content l Fill compacted density l Fill grading and aggregate type l Geocell opening size to height ratio l Fill grading in relation to geocell size l Geocell wall roughness l Geocell material stiffness l Geocell shape l Subgrade strength and stiffness l Load transfer between geocell panels.
As well as permanent deformation, there are also a number of failure mechanisms specific to geocells:
l Fill penetration into the subgrade l Geocell seam or wall bursting l Geocell wall buckling l Geocell wall tearing Published studies are available on the performance of geocell-reinforced layers compared with equivalent unreinforced layers. Kazerani and Jamnejad (1987) obtained a 100% improvement in stiffness using well graded fill in laboratory tests. Using aluminium geocells of varying dimensions filled with sand, Mitchell et al (1979) reported stiffness improvements of between 40% and 173%.
Mitchell et al (1979) draw on previous site trials to suggest that geocell confinement of granular fill also provides a structural slab effect.
Geocells are delivered in collapsed form and need to be pinned out to open up the cells. Once the cells are filled and the pins removed, a unidirectional prestress may enhance fill containment and structural continuity between adjacent cells. This means that obtaining a more realistic picture of geocell behaviour requires testing over full-size panels rather than simply comparing stiffness values measured locally. As far as Pavement Technology is aware, there are no detailed studies of these macro-effects.
A geocell reinforced sub-base is a complex composite structure with a number of interdependent variables that influence performance.
Clearly, there are design and construction implications arising from both internal behaviour within individual cells and external composite behaviour.
As result, it is considered more productive initially to examine geocell performance experimentally rather than seek a mathematical solution.The future option of finite element analysis is clearly available but experimental data would still be a prerequisite in calibrating such modelling.
Geocell trials Investigations so far have concentrated on resilient stiffness obtained from full-size panels tested with a PRIMA 100 light falling weight deflectometer (LFWD). No published test data from falling weight methods has been found in the literature. Initial tests used fill consisting of reclaimed clay-contaminated trench arisings treated with pulverised fuel ash (PFA) to control moisture content. LFWD deflection profiles from initial laboratory trial panels indicated compaction below optimum, despite compacting cells individually.
A second series of field trials have been carried out with two trial panels using 240mm by 240mm cell openings with a 200mm thickness installed.
In both cases, reclaimed crushed concrete fill was used at different gradings.The first of these was a more finely graded fill within the Class 6F1 grading envelope in accordance with Specification for Highway Works (SHW) Table 6/1. A coarser material was also available that complied with the SHW requirements for Class 6F2.
The trial site was a former landfill where the subgrade was had an approximate California Bearing Ratio of 1% to 2%.
Although both fills complied with the SHW grading envelope, both were gap-graded. As well as resulting in a less dense compacted fill, gap-grading also carries the risk of segregation during placement. In the context of this trial, the fills used were readily available without secondary crushing and screening. The additional on-site treatment to achieve a well-graded fill would clearly have led to additional processing costs.
The average increase in stiffness between unreinforced and reinforced bases was 84% and 46% for Class 6F1 and 6F2 fills respectively.
The 84% stiffness improvement for the Class 6F1 fill is considered credible for an uncontrolled gap-graded material placed under site conditions.
Reduced improvement using the Class 6F2 fill was considered, due to segregation of large particles during placing, because the grading was too coarse in relation to geocell size. It is tentatively recommended that maximum particle size should not exceed 20% of the geocell opening size.
Conclusions In terms of increased resilient stiffness, geocells appear to be a credible means of unbound layer improvement using recycled materials.
It is clear that more work is needed to understand structural performance, particularly since there are a large number of variables. In particular, effects beyond resilient stiffness have yet to be fully investigated.
Acknowledgement Geocell samples for the tests reported in this paper were kindly provided by ABG as part of a collaborative programme.
Jack Bull is an associate at Pavement Technology.
References Dawson AR, Little PH and Brown SF (1994). Rutting behaviour in geosynthetic-reinforced unsurfaced pavements. Geotextiles and Geomembranes, 19, 235-256.
Kazerani B and Jamnejad GH (1987). Polymer grid cell reinforcement in construction of pavement structures. Geosynthetics '87 Conference, New Orleans, 58-68.
Lekarp F and Isaacson U (2000). Permanent deformation in granular materials - state of the art. Proc 5th International Conference on the bearing capacity of roads and airfields, 1247-1256.
Mitchell JK, Kao TC and Kavazanjian E (1979). Analysis of grid cell reinforced pavement bases. Technical Report GL-79-8, US Army Engineer Waterways Experiment Station, Vicksburg.