This paper summarises the results of a full-scale trial soil nail wall reinforced with the PermaNail system. The trial wall was constructed in a slightly cohesive granular fill and was eventually 6.5m high, with a face angle of 20degrees to the vertical. After six months, and a surcharge load of 105t (equivalent to a surcharge pressure of 32.9kPa). The maximum recorded horizontal movement of the face was 3.5mm.
Soil nailing is an increasingly common technique for the retention of steep cut slopes. Soil nails are very commonly used, for instance, on road widening projects to support an oversteepened cut soil face, allowing the construction of an extra traffic lane without additional land take. To date, the most common form of soil nail utilises a high yield or mild steel bar, either driven directly into the soil or grouted into a pre- drilled hole.
The corrosion resistance of such steel nails is of constant concern, particularly where a long service life is required. Current techniques to counter this concern tend to encase the steel nail within a corrugated plastic sleeve filled with cement grout, in a similar manner to the method used for corrosion-protected ground anchorages (eg Barley, 1997). This is often used in association with a galvanised coating to the bar and fittings. Alternatively, a greater cross-sectional area of steel may be used to provide a sacrificial layer, in a similar way to procedures adopted for steel sheet piles (eg Murray, 1993). Other systems utilise the grout cover formed during the drilling of the borehole for the nail. Typical of this approach is the Ischebeck system, which uses a hollow bar as a drill rod and cement grout flush to surround the bar with passivating cement.
An alternative approach is to use a non-corrodible material for the load carrying tendon. Some success has been obtained with soil nails utilising the technology developed for grp rock bolts, as commonly used in the mining industry (Harper et al, 1995).
A further development, which is the subject of this paper, is the PermaNail system. This is a patented system based on the use of a non-corrodible Paraweb polyester webbing system similar to that used for many years in the construction of reinforced soil walls (Kempton, 1994).
The PermaNail system consists of one or more Paraweb geotextile straps, looped at the base of the nail tendon to form two, or multiples of two, tension legs (Figures 1 and 2). The Paraweb straps consist of high modulus polyester fibres within an outer polyethylene case. The use of these fibres as the load carrying elements results in a material with excellent long- term load carrying capabilities. Long-term, or secondary, creep of the polyester fibres forming the Paraweb strap would typically be less than 1%. The polyethylene skin both maintains the shape of the material and protects the polyester fibres from installation damage. Materials such as this are already in widespread use in the field of ground engineering, and include Paralink, Paraweb and Websol (now Freyssisol) reinforced soil walls.
The use of such a material as a soil nail provides an inexpensive, flexible and durable solution to the problem of providing a permanent tensile member to reinforce cut slopes.
As noted above, the use of Paraweb straps is well established for reinforced soil walls. In this technique the wall is formed by a filling process, and uses a selected fill material and controlled filling and compaction. However, their use in soil nailing and the PermaNail system was innovative, so that the details of their performance was not known with certainty. By contrast with reinforced fills, in soil nailing the wall is formed by an excavation process, and the nails are installed into materials whose geotechnical properties are uncontrolled compared with a reinforced fill. It was not known, for instance, how important the controlled compaction of the fill around the strapping was as a factor.
The PermaNail would also be installed into a cement grout filled borehole, giving different performance characteristics to a normal reinforced fill process. Because of this last point, as part of the conceptual design process it was perceived necessary to ensure that the Paraweb tendon could not pull out of its encasing grout body. For this reason, the tendon is looped around the base of the nail to avoid any possibility of pull-out failure of the tendon from the grout body.
As part of the development process, it was deemed necessary to examine the performance of a PermaNail wall at full scale. With the kind permission of BFI a site was made available to construct a trial PermaNail wall near Alton in Hampshire.
The site was in a working sand quarry near Kingsley, some 6km east of Alton, which was excavating materials from the Cretaceous Upper Greensand. The test site formed part of the southern rear wall in a worked-out section of the pit. The material forming the wall was generally a slightly cohesive granular fill, derived either from the ground stripping for the quarrying works themselves or imported and dumped onto the site from nearby road construction works.
A control site investigation borehole was constructed by cable tool methods adjacent to the test site, to confirm the nature of the slope materials.
The test wall was excavated in a loose to medium dense, slightly clayey, brown/grey sand, often containing rootlets, and with occasional gravel- sized fragments of flint, brick and ash. Occasional thin topsoil-like bands could also be identified within the fill. From approximately 3m to 4m below the crest level the clay content of the material increased. Occasional concrete blocks or lumps were also noted within the fill. At the foot of the deepened slope, at approximately 6.5m below crest level, the fill was underlain by a dense yellow-brown, slightly clayey sand, which appeared to be natural ground.
SPT N-values in the material varied between N=7 to N=13, except where concrete obstructions were encountered. A cross-section through the test wall is shown on Figure 3. The parameters assumed for design, based upon the testing and sampling at the site, were:
c' = 0
' = 30degrees
= 18 kN/m3
No groundwater was encountered in the slope materials and, from an appraisal of the layout of the site, the groundwater table was clearly well below the toe of the cut slope.
Details of the test wall
The test wall was initially excavated to a height of 5m, with a face angle of 70degrees to the horizontal. The face was supported by four rows of 5m long PermaNails (Figure 3). After completing the first sequence of loading and monitoring, the effective slope height was increased to 6.5m by excavating at the toe for a further 1.5m. The final cross-section of the test wall is shown in Figure 4.
The PermaNail reinforced trial slope was designed using a proprietary reinforced earth/soil nail analysis program in general accordance with the methods outlined in BS8006: 1995, the British Standard code of practice for strengthened/reinforced soils and other fills.
BS8006 is based upon a limit state partial factor design approach. With this design philosophy partial factors are applied to each of the disturbing and restraining forces acting within the system. For a 'safe' design, the overall design factor of safety, Fdes, should be equal to or greater than unity. In practice, Fdes may be taken up to 1.1 to provide an additional margin of safety.
The initial design case was based upon a 5m high slope with a nominal 5kN/m2 surcharge load. The analysis indicated that four rows of PermaNails 5m long and at 1m horizontal centres would be required for stability. Each nail required a design working load of 42.5kN to provide a Fdes of 1.1. On either side of the test section the nails were opened up to 1.5m centres, giving a Fdes of 1.0. In terms of a conventional overall or 'global' factor of safety (FoS) approach, the initial design was equivalent to an FoS of approximately 1.18.
The initial part of the test involved progressively applying a surcharge to the crest of the slope to a maximum load of just over 17kN/m2 (global FoS approximately 1.08) and monitoring the performance of the slope during this loading cycle.
Thereafter the surcharge was increased to 33kN/m2, at which it was calculated the global FoS would fall to approximately unity. The 1.5m excavation at the toe effectively reduced the global FoS still further, to approximately 0.8.
Details of PermaNails at the test site
The PermaNail configuration at the Kingsley Pit test site was based on a 50mm wide Paraweb geotextile composite, each strip having a breaking load of 1,700kg (17kN). The four-legged tendon thus had a breaking load of 68kN and a designated working load of 42.5kN.
The PermaNail tendons were fabricated on site and were attached to circular spacers at intervals along their length, to prevent the geotextile strips from collapsing onto one another. A typical view of fabricated PermaNails is shown in Figure 5.
The PermaNails were fabricated complete with circular spacers on to a former consisting of a 50mm diameter plastic pipe. This was also used to push the fabricated nail tendon down the grout-filled hole. Figure 6 shows a view of fabricated nails ready for installation. After installation, each PermaNail was fitted with a head plate assembly. This secures the Paraweb straps and allows a nominal tension to be applied to the nail.
Installation and construction
The 5m high slope was excavated by a mechanical excavator in a single operation. All the nails were then installed by a long-mast tracked rotary hydraulic anchor drilling rig, working at the toe level and standing away from the slope face. From this level the drill rig was able to reach up to the highest row of nails, so that all the nails were installed in a single pass.
The nail holes were drilled at 114mm diameter to their full depth without casing, using rotary methods and water or grout flush. Once the hole was down to depth, the flushing fluid was switched to a neat cement grout with a 0.5 to 0.6 w/c ratio. The drilling rods were steadily withdrawn once cement grout showed at the surface. A typical view of the drilling and installation operation can be seen in Figure 7.
The PermaNail on its plastic former was then easily pushed down the grout- filled drill hole to its required position. The plastic former was then withdrawn, leaving the PermaNail tendon down the hole.
Once the nail installation had been completed and grout had cured sufficiently, the face was cleaned down as necessary. A load distribution plate consisting of a 600mm by 600 by 25mm thick plywood board was fitted over each nail and secured by the nail head assembly and the assemblies tightened to a nominal tension. A completed assembly is shown in Figure 8.
Construction and loading sequence
The test face was excavated in February 1997 and the soil nails were installed over a two day period. Nail heads and load distribution plates were fitted prior to leaving site.
Nail load tests were undertaken at the beginning of April 1997, and Stage 1 loading began on 8 April. Stage 1 loading was maintained until July 1997, with movement monitoring at daily and then fortnightly intervals.
Stage 2 loading began on 1 July 1997 and was maintained till 23 July. Slope movements were monitored at daily intervals for the first three days, and then weekly. Stage 3 loading commenced on the morning of 23 July 1997, and was terminated at the end of that day.
Surcharge loading was achieved by placing large steel water tanks on the crest of the slope. These could be filled progressively and the slope behaviour monitored. Additional loading was achieved by concrete kentledge blocks in addition to the water tanks.
Loading of the slope was undertaken in the following three stages:
Stage 1: Slope height 5m
Two water tanks placed on the slope crest.
Maximum applied load = 555kN
Bearing pressure = 17.2kPa
Stage 2: Slope height 5m
Two water tanks placed on top of concrete kentledge
blocks on the slope crest.
Maximum applied load = 503 + 555 = 1058kN
Bearing Pressure = 32.9kPa
Stage 3: Slope height increased to 6.5m by excavation of a trench
along the toe of the trial section. Two water tanks
placed on top of concrete kentledge blocks on the slope crest.
Maximum applied load = 503 + 555 = 1058kN
Bearing pressure = 32.9kPa
Each stage was commenced with the water tanks empty. Monitoring of the slope was undertaken as the tanks were filled with water. The loading stages are illustrated in Figure 9.
The behaviour of the cut face was monitored in two ways:
By optical surveying of survey targets established on the slope face and on the kentledge stack.
Horizontal displacements of the face were measured at two locations by horizontal extensometers.
Each extensometer consisted of a fibreglass rod 6m long, with its upper 5m sheathed with a loose fit UPVC tube. The distal 1m of the rod was grouted into the drill hole, so that it was effectively anchored beyond the distal ends of the soil nails. It should be noted that these were not sensitive instruments, but were installed as a monitor of gross soil movements and to serve as a check of the optical survey data. An elevation of the nailed face is shown on Figure 10, with the positions of the survey targets and extensometers indicated.
Optical survey markers
The optical survey allowed the movement of the survey markers to be monitored in three dimensions. Lateral movements horizontally parallel to the plane of the slope were found to be negligible and not discernible within the scatter of the measurements. Vertical movements and horizontal movements into and out of the slope face have been plotted on Figure 11 for survey line B, one of two survey lines through the area of greatest surcharge load. The other survey line, Line C, gave similar results. Figure 11 indicates the measured displacements of the survey markers on Line B during the test loading stages 1, 2 and 3.
Also shown on Figure 11 are the horizontal displacements measured by Extensometers 1 and 2, which straddled the line of Line B. The displacements recorded are also summarised on Table 1.
Summary of results
As might be expected, the survey marker located on the kentledge stack itself (Point 1, Table 1) recorded the greatest horizontal and vertical displacements, and exhibited a total vertical settlement of 79mm and a total horizontal displacement of 35mm outwards.
Movements of the slope face were between one and two orders of magnitude less than the kentledge stack. Maximum horizontal displacements were 3.5mm outwards at the end of the Stage 3 loading. These displacements were measured by Extensometers 1 and 2 located some 2m and 3m below the crest of the slope.
Maximum displacements measured by the survey markers at the Stage 3 loading were a vertical settlement of 4mm together with an outward movement of 1mm at Point 5, 2m below crest level. Prior to Stage 3, Points 7 and 9 on the lower portion of the face had recorded slight upward movements.
Load testing of individual nails
Bond stress at grout/ground interface
The design of individual PermaNails nails is undertaken in a similar manner to conventional steel nails in terms of the grout-to-ground interface and the capacity of the tendon.
Load tests were undertaken on both PermaNails and steel nails using a hollow ram hydraulic jack to confirm that their performance was in accordance with the design requirements. Both forms of nail developed the design bond stresses without failure or distress at the grout/ground interface. The maximum bond stress developed was 41kN/m2 at this interface, without distress or significant displacement at the grout/ground interface.
Load/extension characteristics of PermaNail tendons
The load-extension characteristics of the Paraweb tendon are somewhat different to conventional steel nails. The elastic extension of a high yield steel nail is approximately 0.5% to its yield point. By contrast, the load-extension characteristics of a Paraweb polyester strap are linear to failure at 10% extension. Hence the extension of a Paraweb tendon is approximately 20 times that of an equivalent steel nail.
This difference in load-extension characteristics could be expected to result in a more flexible and resilient structure than a wall constructed of stiffer, steel elements (Jones, 1995). In practice, in terms of serviceability, there appears to be little recorded difference in the performance characteristics of reinforced soil structures constructed with geosynthetic reinforcement members as opposed to steel.
Discussion and conclusions
The behaviour of the reinforced soil slope under the imposed surcharge loadings was very encouraging.
Movements of the kentledge stack and the face of the nailed slope itself were monitored by both optical survey techniques and by extensometers anchored into the soil beyond the soil nails. The maximum recorded horizontal movement of the slope face at 2m below the crest was around 3.5mm at the maximum surcharge load of 32.9kPa (105t of kentledge) with an effective slope height of 6.5m.
Published data on nailed slope displacements (Bruce and Jewell, 1987) suggests that one guide to performance may be given by the ratio of the horizontal displacement (say, e) of the crest of the slope divided by the slope height (say, h): termed the 'performance ratio' by Bruce and Jewell.
Typical values obtained from a review of published case histories gave typical values for e/h of between 0.0004 and 0.002, for a range of materials. Taking the increased surcharge loading on the slope as equivalent to an increase in slope height of, say, 1.8m, and including the overdig of 1.5m at the toe, the equivalent slope height at the end of the Stage 3 loading was some 8.3m. With a measured displacement of 3.5mm, this gives a performance ratio, e/h, for the PermaNail wall of 0.0004, which is well in accord with other published values.
It has been noted above that the more extensible Paraweb reinforcing elements would be expected to provide a more resilient structure than a slope reinforced with steel elements (Jones, 1995).
It should be noted that, in common with many cut slopes in apparently cohesionless materials, the test face stood unsupported to a height of 5m in the short term. This indicates that the material exhibited, at least in the short term, an element of effective cohesion due to its slightly clayey nature, to negative pore pressures generated within the partially saturated soil, or, possibly, to a degree of secondary cementation between the mineral grains.
The cut face also performed satisfactorily over the period of the testing, from February to July 1997, with face support provided only by the 600mm square distribution pads. Over the testing period the face generally became covered by a luxuriant over-growth of weeds. These evidently helped to shield the face from erosion by wind and rain, and served to bind the face by root action. Parts of the face were also covered by grout spillage from the nail installation and this also evidently assisted in maintaining face stability.
The PermaNail trials were supported by Soletanche (now Bachy Soletanche), who also undertook the site works. Messrs BFI kindly made a site available for the trials. Without the help and support of both these parties, the trials would not have been possible.
Barley AD, Maddison JD and Jones DB (1997). The use of soil nails for the stabilisation of a new cutting for the realignment of the A4059 at Lletty Turner Bends. Proc. third int. conference on ground improvement geosystems. London, 3-5 June. Thomas Telford.
Bruce DA and Jewell RA (1986/87). Soil nailing: Application and practice, parts 1 and 2. Ground Engineering, Nov pp10-15 and Jan pp21-38.
BS8006 (1995). Code of practice for strengthened/reinforced soils and other fills. BSI, London.
Harper JS, Smart BGD, Sommerville JM, Davies ML and Spencer IM (1995). Recent applications of soil nails and cables in the UK. Proc. int. symp. on anchors in theory and practice, Salzburg, Austria, 9-10 October 1995.
Jones CJFP (1995). The development and use of polymeric reinforcements in reinforced soil. The practice of soil reinforcing in Europe. Thomas Telford, London, 1995.
Kempton G (1994). Geosynthetic soil reinforcement materials. Global Construction 1994. Sterling Publications.
Murray RT (1993). The development of specifications for soil nailing. TRL research report 380. TRL, Crowthorne, Berks