Recent Norwegian research has shown for the first time that the capacity of anchors fitted with inflatable expander bodies in soft clay can be substantially increased by controlled preloading. Consolidation of soil at the base of the anchors installed on the Sentrumsring Drammen highway resulted in a 100% capacity increase over eight weeks.
Swedish manufacturer Soilex's Expander Body soil anchors have already been shown to achieve higher load bearing than conventional anchor systems since they were first introduced 15 years ago. In soft clay especially, conventional anchors only achieve modest bearing capacities. This means that these systems are often difficult and costly to install if the soft clay is deep lying or thick because anchors have to be installed through these deposits and into denser soil or rock to achieve the required capacity.
The expander body, which is the key component of both Soilex's anchors and piles, consists of a 0.15m diameter package of folded sheet steel, which is inflated by injecting grout after the body is in position. Bodies can be made to give a range of expanded diameters from 0.4m to 0.8m, and can be driven or vibrated into place, or inserted in predrilled holes.
The main advantage of the Soilex anchor system is the way applied load is resisted. In conventional anchors, the majority of the load is taken by skin friction, but with the expander body capacity it is derived by the soil resisting the movement of the expander body.
In conventional anchors, maximum capacity is reached at relatively small displacements - typically a few millimetres - and drops once the peak value is exceeded. The soil's residual strength then governs the capacity at large displacements. In these anchors, with predominant shaft resistance, excessive pretensioning has a negative influence on the ability to carry load and this strain softening effect is an important part of the factor of safety calculation.
In the Soilex anchor, expander body side resistance is also mobilised at relatively small deformations and falls when the peak value is exceeded. However, the base resistance increases gradually and relatively large deformations are required to reach the peak value. This leads to strain hardening behaviour and implies that at even larger strains the bearing capacity will increase until the ultimate soil strength is reached. For design purposes, Soilex takes the maximum base capacity to occur at deformations of 5% of the expander body diameter (about 20mm to 40mm).
During inflation, soil surrounding the expander body is displaced laterally and the expander body length, increasing the pressure against the sides. In clay soils, this increase results in a build-up of excess pore water pressure. Immediately after the injection phase the clay starts to reconsolidate, water content drops and the undrained shear strength increases. This effect is largest close to the body side and is negligible at the base.
Soilex says this reconsolidation effect is similar to that of conventional anchors or driven piles in clay but affects a larger volume of soil, up to three times the inflated diameter. And although this effect only happens once in conventional systems, reconsolidation of the clay in the zone at the expander body base can be controlled and repeated by several preloading steps.
Preloading, typically to 1.5 times the working load, sets up a tension force in the soil in this zone - again three times the inflated diameter - at the base of the body. Consolidation again results in an increase in stiffness and undrained shear strength of the clay and can also eliminate any disturbance effects along the shaft that occur during installation. Repeated loading causes further increases in soil strength.
The Swedish commission on pile research recommends two alternative design methods for Soilex piles and anchors based either on field penetration tests and vane tests or by measuring the injection pressure. However, Soilex considers it unwise to use the latter as the injection pressure is relatively low (around 10 times the soil's undrained shear strength). The company recommends using methods based on field measurements of undrained shear strength which are proven to be more reliable.
However neither method accounts for the reconsolidation effect and thus, says Soilex, underpredicts capacity for practical design.
At the Norwegian test site at Drammen, ground conditions were almost homogeneous, with a 2m thick surface crust of dense sandy silt overlying a 1m thick layer of silty clay. This was underlain by soft, normally consolidated clay down to 9m. Below this was a stiff till on bedrock. Groundwater level was at the top of the clay layer.
Five 16m to 20m long anchors were installed as part of the field trial for the Norwegian Geotechnical Institute. Soilex EB825 anchors were used, which have an expanded size of 0.8m diameter and are 2.1m long. A 115mm diameter hole was predrilled through the stiff surface crust and the anchors pushed and driven at an angle of 20degrees to 30degrees into the soft clay. Finally, four 150mm diameter steel strands were installed in each shaft. Prior to installation, an expander body was injected with water to calibrate the resistance of the folded sheet steel and allow the required expansion pressure to be calculated.
Two of the anchors were fitted with restriction rings to increase the strength of the body and prevent leakage of grout when using high injection pressures. However, because the diameter of the rings is greater than the shaft, some disturbance of the clay can occur during installation, which is counter-productive. The remaining three expander bodies did not have rings, allowing 'over-injection', which is theoretically advantageous in soft clays, provided the process is carefully controlled, says Soilex.
Injection was carried out within two days of installation, monitored by Soilex's Plogg system, data from which forms the basis for site verification of the design assumptions and checks the reliability of the installation. Injection stopped when there was a big increase in expansion pressure, indicating higher resistance from the soil.
After installation and injection the five anchors were each tested four times over an eight week period in which piles were preloaded on two occasions. Initial estimated failure load was 139kN, assuming an average undrained shear strength of 22kPa; the expander body achieved a base (end) resistance of 99kN and side resistance of 64kN.
Load tests after one or two weeks showed capacity ranging from 150kN to 250kN at failure for the five anchors. Anchors were then preloaded in two stages with re-testing after each.
Capacities of the five anchors after eight weeks ranged from 336kN to 464kN, showing a more than doubling in capacity following prestressing.
The testing showed that 'over-injected' anchors without restriction rings achieved higher loads at smaller deformations than those with rings, which can be explained by shaft disturbance by the rings during installation and no over-injection affecting the base resistance of the body.
Soilex now proposes to investigate, with further field trials, the extent to which the capacity can be increased earlier and when preloading should take place in other soils such as stiff, overconsolidated clays.