Case Studies

views updated

Chapter 8
Case Studies


Owing to the helerogeneity nature of soil, soil improvement is a complex and site specific process. Some degree of engineering judgement are often required. Past experiences contribute a great deal to good engineering judgement and experiences can be gained partially by learning from case histories. In this chapter, ve well documented case studies are presented. These include the Changi Airport project, the Changi East land reclamation projects in Singapore, the Chek Lap Kok reclamation project in Hong Kong, the Tianjin East pier project and an oil storage station project in China. In the last two projects, the vacuum preloading method was used together with the vertical drain techniques. The projects are described based on mainly


The construction of Changi Airport in the late 1970s necessitated the reclamation of about 700 ha of land along the northeastern shoreline of the Republic of Singapore (Fig. 8.1). The project called for dredging and hydraulic lling of 40million cubic meters of sand from the adjacent seabed. Seven cutter suction dredgers each with a capacity of 5000–10,000 HP were deployed, and the reclamation was completed in 1979.

Site investigation indicated that the proposed runway location is partly on hard ground of sandstone and shale and partly on thick deposits of soft marine clay. The marine clay exists in depth up to about 40m below the seabed and is deposited mainly in deep channels. The isopach of soft clay and soil prole along the centreline of runway II is shown in Fig. 8.2. Long-term settlements caused by6.5 m of hydraulic ll and the runway pavement load were estimated to be up to 2 m. To minimize the problem of uneven settlements and heavy maintenance work, soil improvement became necessary for the runway and its associated turnoffs and taxiways.

8.1.1 Soil Properties

Geophysical survey and boreholes indicated that the site was under-lain by a thin layer of sand gravel followed by a stiff-to-hard clayey silt and cemented sand. Numerous valleys resulting from erosion of this hard clayey silt and sand were subsequently infilled by soft

marine clay. The marine clay varies in thickness from 2 to 40m below the seabed. The deeper deposits of marine clay appear to be of two different ages separated by a rm-to-stiff silty clay stratum or a sandy, peaty clay stratum. It is believed that this transition zone of stiff clay was the result of a fall in the sea level sometime in the geological past, exposing the older marine clay to the atmosphere and forming a desiccated crust, which was then submerged again with the subsequent rise in the sea level, and further deposits of younger marine clay on top of this stiff clay stratum.

The marine clay had a plastic limit of around 40%, liquid limit between 90 and 100%, and a natural moisture content between 80 and 100%. The undrained shear strength of the marine clay increased from around 5 kPa at the seabed, to 20kPa at a depth of 8 m below the seabed. The compression index Cc was about 1.0. The marine clay showed a slight overconsolidation with an overconsolidation ratio p′c/p′o of about 1.2. The coefficient of consolidation with flow in the vertical direction cv was about 1–2 × 10–7 m2/s from laboratory tests. The coefficient of consolidation with ow in the horizontal direction ch was about 4 × 10–7m2/s from dissipation tests carried out using a BAT pore pressure sounding device.

8.1.2 Instrumentation

The soil instruments were installed in clusters consisting of a settlement plate on the seabed, deep settlement gauges at various depths below the seabed and a deep reference point embedded in the hard stratum to act as a datum. At the same location, pueumatic piezometers were installed at various depths in the marine clay to measure pore pressures. In some locations where lateral movement was anticipated, inclinometers were also installed. Borros settlement devices were installed in several locations where the marine clay thickness exceeded 8 m in order to measure the relative settlements of the clay at various depths.

The reclamation at Changi was divided into six zones. At least one cluster of instrument was installed in each zone before the reclamation began. After the first lift of the reclamation, when the ground level had reached +3.40mCD (Chart Datum), additional settlement plates were installed at the seabed in a 300 × 300 m grid to contour the settlement due to the reclamation fill.

The settlement plates installed before the reclamation consisted of 25 mm diameter steel pipes welded and braced to 600 mm × 600 mm steel plates. These plates were lowered directly on to the seabed. The settlement plates installed after the first lift of reclamation consisted of 25 mm diameter steel pipes welded to a 150mm diameter base plate. These settlement plates were installed in 200 mm diameter boreholes advanced to the original seabed level through the reclaimed fill. The deep settlement gauges consisted of helical augers welded to a steel pipe. These were installed in boreholes and screwed into the soil to the desired depths.

The deep reference points consisted of steel rods with pointed ends which were driven into the ground at the bottom of the bore-holes advanced at least 1.5 m into very hard stratum. This last 1.5 m was then grouted. All these settlement devices were encased in PVC pipe outer casings to eliminate friction on the central steel pipe due to the settling soil around the instrument.

Inclinometers were used to monitor the lateral deformations associated with the sand filling and rock bund construction as well as with the excavation works for the drainage channels. Closed type

piezometers such as pneumatic piezometers were required at various elevations in the clay layers to monitor the build-up of the pore pressures due to the ll operation and the subsequent dissipation of pore pressures as consolidation took place. The piezometer readings indicated the degree of consolidation that had taken place at each elevation and could be used for settlement predictions as well as assist in the stability analyses for staged construction of the ll and rock bunds. Water standpipes and open type piezometers consisting of pvc pipes attached to porous Casagrande type filter tips were installed to measure the water table in both the hydraulic fill as well as in the underlying permeable sand layers. During hydraulic filling, the water table in the fill fluctuated considerably both in plan as well as with respect to time.

The intensity of the discharge of the hydraulic fill from cutter suction dredgers was very great. The instruments installed during reclamation had therefore to be adequately protected. One fairly successful means was to form a sand mound around the instruments and to brace the instruments to a steel scaffolding. The scaffolding also served as a platform from which the instruments could be monitored. Even after the rst lift of reclamation, sand mounds had to be constructed around the instruments to protect them against the second lift of hydraulic ll. These mounds also served as a protection against damage from earth-moving equipment such as dozers and bucket loaders which were often used to lay pipeline and re-grade the slopes on the reclamation site.

A typical layout of the soil instrumentation used to monitor the soil improvement works is shown in Fig. 8.3. It should be noted that the instrumentation scheme enables the monitoring not only of the pore pressures, settlements, and horizontal deformation at various elevations of the different soil strata, but also the variation of these measurements in the lateral direction. To fully understand the consolidation process of a site undergoing improvement by vertical drains and surcharge, it is necessary to appreciate the three-dimensional nature of the problem. The consolidation settlements and pore pressures are very different at the edge of the drain treatment zone, under the surcharge over-width, where there are no drains, and beyond the surcharge over-width, where the reclamation

is still undergoing consolidation and where excess pore pressure has not yet completely dissipated. Figure 8.4 illustrates the variation of pore pressures at various distances from the centreline of the treated taxiway. Figure 8.5 shows a typical variation of surface settlements with distance from the centreline of the runway. In this figure is also shown the settlements predicted by a numerical method of analysis using finite difference approximation and based on the Biot three-dimensional consolidation theory (Karunaratne et al., 1989). The details of case study on reclamation and soil improvement works for runway II at Changi Airport can be found in Choa (1981, 1991, 1994) and Choa et al. (1979a–c 1981).

8.1.3 In-Situ Testing

In-situ vane shear tests and thin wall piston sampling were carried out in boreholes advanced by percussion drilling rigs and rotary wash boring rigs mounted either on specially designed drilling platforms or on jack-up barges. Drilling platforms or jack-up barges are essential to ensure minimum disturbance during the testing

and sampling. Drilling rigs mounted on floating pontoons should not be used because the disturbance from wave action makes it almost impossible to carry out undisturbed vane test and to obtain undisturbed samples.

Dutch cone tests were also carried out to determine the quality of the reclaimed ll and the thickness of the underlying marine clay. The occurrence of sand lenses and silt partings in the marine clay was determined by examining continuous soil samples from boreholes and substantiated by using the BAT type peizocone. The pore pressure results clearly show the occurrence of the sand lenses and silt partings. The BAT pore pressure sounding device was also used to carry out dissipation test to determine the coefficient of consolidation with flow in the horizontal direction ch.

All the data from the various in-situ tests, laboratory tests, and soil instruments were used to evaluate the performance of the soil improvement works. However, eventually most reliance was

placed on the surface and deep settlement observations and the preconsolidation pressures obtained from oedometer tests.

8.1.4 Improvement of Soft Clay with Prefabricated Vertical Drain and Preloading

At Changi Airport the second runway was treated with prefabricated vertical drain and preloading. A pilot test embankment was constructed to compare the effectiveness of various types of vertical drains in combination with surcharge for accelerating the consolidation of the soft marine clay. After careful study of continuous bore-hole samples together with field and laboratory tests, three drain spacings 2.1, 2.6, and 3.2 m on square grids were selected for testing. A single type of prefabricated vertical drain was used. Machines that can install these drains to depths of 60m at speeds of 1 m/s were available at that time. At Changi Airport where the drains were installed to depths of 43 m, an average production of 2200 m of drains per rig per 10-hour day was achieved. The pore pressure dissipation, settlement, and laboratory test results from the pilot tests were then used to determine ch, the coefficient of consolidation with flow in the horizontal direction.

From the value of ch thus determined, the selection of drain spacing and surcharge for the main works were as follows:

(a) 3.2 m square grid in areas where the marine clay is between 5 and 15 m.

(b) 2.5 m square grid in areas where the marine clay is greater than 15 m.

In both cases a surcharge of 4 m of sand was used. For areas with less than 5 m of marine clay, a surcharge of 2.5 m only was used. The design was reviewed after six months, and a further 2–3 m of surcharge was found to be necessary for the deeper clay areas.

Figure 8.6 shows a typical result of the change in preconsolidation pressure after soil improvement. A stagnation of pore pressure with continuing surface settlement was recorded. This phenomenon was attributed partly to the high pore pressure built up in the clay under

the surcharge overwidth and in the adjacent untreated area of recent reclamation (Choa et al.,1981).

The soil improvement works including the pilot test took about three and a half years and was successfully completed by the end of 1980.


The reclamations at Changi East multi-phases, for the rst three phases called Phase 1A, B, and C, were carried out between 1991 and 2002. Another phase called Area A was carried out under two concurrent projects, namely Area A (North) and Area A (South). Of the five phases, only Phase 1A did not require soil improvement

work with prefabricated vertical drain. The other phases required a significant amount of prefabricated vertical drain works. The location map of Changi East Reclamation Project is shown in Fig. 8.7. The projects included a few hundred million cubic meters of sand in water of up to 15 m depth to reclaim about 2100 ha of land for extending the Changi Airport and facilitating infrastructure development. The created land area, and prefabricated vertical drain used are shown in Table 8.1.

The reclamation was carried out in four phases, because large quantities of ll material and prefabricated vertical drains were required. As the reclamation covered an extensive area, the soil profiles and characteristics of soils varied from one area to another. Besides, the future loading and timing of land usage were different within the area. Therefore, the design of soil improvement works was considered accordingly, on the basis of the existing soil profile, future land use, and allowable duration for soil improvement works.

Table 8.1 Project specifications (after Choa et al., 2001).
ProjectArea (ha)Length of vertical drain (Mm)
Phase 1A501
Phase 1B5228
Phase 1C52449
Area A – North9113
Area A – South4550

The acceptance criteria of the soil improvement works also varied because the design criteria and future land use were different. Details of the design concept are discussed by Bo et al. (2000b) and Choa et al. (2001) and are shown in Table 8.2.

As most of the reclamation area was underlain by up to 50m thick of highly compressible marine clay, about 140million meters of prefabricated vertical drains in combination with up to 8 m thick of surcharge were used to improve the engineering properties of the seabed soils. The total area of soil improvement was approximately 1200 ha.

In addition to treating the natural thick clay, the project required a 180-hectare slurry pond having ultrasoft soil of up to 20 m thickness to be reclaimed. The soil improvement work for this slurry pond was difficult and challenging. The work done to improve the natural clay and the ultrasoft slurry is described in the following sections.

The reclamation was carried out with hydraulic sandfill except in a few deep areas where direct dumping of 4 m of clay was carried out before the hydraulic sand filling.

8.2.1 Site Conditions and Geological Profile

The project area covers the foreshore of 2000 ha in the eastern part of Singapore. The seabed at the site varies from –3to –15 mCD (Chart Datum is +1.6 m below mean sea level). The average seabed is about –3to –5 mCD for most of the areas. The seabed slopes

Table 8.2 Details of soil improvement design with PVD in Changi East Reclamation Project (after Bo et al., 2000).
AreaYear of designThickness of clay (m)Type of clayFuture land useDesign spacing sq. grid (m)Design surcharge el. mCDSurcharge period monthsSpecified acceptance criteria
A199220-40Marine clayRunway1.5+101890% consolidation of fill and surcharge load
B199210-35Marine clayTaxiway1.8+8.51890% consolidation of fill and surcharge load
C199520-30Marine clayInfrastructure area1.8+8.52490% consolidation of fill and surcharge load
D199530-45Marine clayInfrastructure area1.8+9.52490% consolidation of fill and surcharge load
E199530-45Marine clayOthers1.8+8.52490% consolidation of fill and surcharge load
F199520-40Marine clayRoads1.5+8.51290% consolidation of fill and surcharge load
G199840Marine clayFuture material stockpile area1.5+121290% consolidation of fill and surcharge load
H199840Marine clayInfrastructure1.8+101890% consolidation of fill to finished level to +5.5 mCD plus 20 kPa
I199840Marine clayInfrastructure1.5+101290% consolidation of fill to +5.5 mCD plus 20 kPa
J199810-35Marine clayInfrastructure1.8+91890% consolidation of fill to +5.5 mCD plus 20 kPa
K199210-20Soft slurryInfrastructure2.0 3 passes+93690% consolidation fill and surcharge load

northwards from –5to –15 mCD in the northern-most part of the area, whereas it was found sloping southwards from –5to –10mCD in the southern part of the area. Deep pockets of seabed varying from –10to –13 mCD were found at the eastern part of the area before the area was reclaimed. This large deep pocket of seabed up to –14 mCD at the northeastern area was formed by the dredging of seabed in this area for a re-handling pit used for Phase 1B and 1C projects. It was noted that the seabed at the southern side and the central area had been highly altered by human activities during three decades. In the late 1970s, about 40 million cubic meters of sand from the seabed of the western part of the central area was dredged out up to –22 mCD for reclaming 700 ha of land for constructing the Changi Airport with two runways.

Changes in the original seabed conditions at the eastern part of the area can be noticed, because some deep pockets of the seabed have been filled up to –4.0to –5.0mCD with either slurry or a mixture of clay and sand.

Generally, the project area is underlain by Singapore marine clay in the two deep valleys cut in the Old Alluvium. Figure 8.8 shows the typical soil profile along the North–South line of the area. The Old Alluvium consisting of cemented silty clayey sand is exposed to the seabed at the central area close to the southern side. This Old Alluvium sand is generally distributed throughout most of the eastern area of Singapore island where exposure can be found in some old sand quarries. Two erosional limbs were formed at the northern and southern side of the study area by a geological process. The northern valley is deeper than the southern one, and the surface of the Old Alluvium slopes towards the northern side. The thickness of marine clay found in this valley ranges from 5 to 55 m. Two marine members of the Kallang formation, locally known as upper marine clay and lower marine clay, are found to be separated by a layer of either stiff silty clay or medium dense silty sand of 2–5 m in thickness. This desiccated crust layer of silty clay, locally known as the ‘Intermediate layer’, was formed by desiccation when the sea level dropped by 20–25 m during two regressions between 10,000 and 20,000 years ago (Pitts, 1983). The desiccated silty clay and silty sand layers are found at elevations

varying from –10to –28 mCD. In some areas the alluvial sediment of silty sand was deposited on top of the lower marine clay where the lower marine clay was exposed at the river mouth in the past.

In some of the northern-most parts of the project area, the marine clay is directly underlain by granite rock at about 60m depth from the seabed level. The granite belongs to a member of the intrusive Pulau Ubin Granite. The southern valley cut, which is shallower than the northern one, is also filled up with marine clay of 4–20m in thickness. However, unlike the northern valley, no intermediate layer is found between the upper and lower marine clay. The marine clay in this southern valley is more overconsolidated than that in the northern valley owing probably to the removal of the top part of the upper marine clay and sand deposit in the past. Therefore, undrained shear strength of the marine clay in the southern part is much higher than that in the northern part. Some parts of the Old Alluvium in the project area are found to be overlain by the soft slurry clay or clay and sand mixture; these are the result of human activities.

8.2.2 Soil Properties Physical Properties

The moisture content of the upper marine clay ranges from 50to 85%, whereas that of the lower marine clay ranges from 40to 65%. The plastic limits of both clays are similar and are found to be about 26–28%. The liquid limit of the upper marine clay is 70–90%, whereas in the lower marine clay it ranges from 60to 90%. The plasticity index of the upper marine clay and lower marine clay are ranging between 44 to 64 and 34 to 64%, respectively. The characteristics of both clays are plotted on the plasticity chart, and both are above the A line. The upper marine clay is found to be highly plastic, whereas the lower marine clay is medium-to-high plastic clay (Fig. 8.9). The grain size distributions of both clays are shown in Figs. 8.10. The upper marine clay has 50% silt and 50% clay, whereas the lower marine clay has 60% silt and 40% clay.

As explained earlier, the intermediate clay is a dissicated clay of the lower marine clay; therefore the mineralogy of this clay is very similar to that of the lower marine clay, and the grain size

distribution is also similar to that of the lower marine clay, although it has slightly more silt content as shown in Fig. 8.10. The other physical properties of the three types of clay are shown in Fig. 8.11 and Table 8.3

Table 8.3 Properties of Singapore Marine Clay
Bulk density (kN/m3)14–1618–2116–18
Specific gravity2.58–2.652.6–2.752.55–2.7
Moisture content (%)50–8520–4040–65
Liquid limit (%)70–9030–7060–90
Plastic limit (%)26–2820–2826–28
Plastic index (%)44–6410–4234–64
Liquidity index (%)0.4–10.1–0.60.2–0.8
Natural void ratio1.5–2.30.5–11–1.7
Compression index0.6–1.20.2–0.30.4–1
Re-compression index0.1–0.20.02–0.10.05–02
Coefficient of permeability in vertical direction (kv) (m/s)10-9–10-10-10-9–10-10
Coefficient of permeability in horizontal direction (kh) (m/s)10-8–10-9 10-8–10-10
Coefficient of consolidation due to vertical flow (Cv) (m2/yr)0.5–1.21.2–60.6–22
Coefficient of consolidation due to vertical flow (Ch) (m2/yr)1–4-4–6
Overconsolidation ratio1.5–72–41.5–2 Compressibility and Consolidation Properties

The compressibility parameters of the three types of clay at Changi East are shown in Fig. 8.12. Both the upper and lower Singapore marine clays at Changi East are highly compressible, and the intermediate clay is significantly less compressible owing to its desiccation. The compression index of the upper marine clay ranges from 0.6 to 1.2 with an average of 1, whereas the lower marine clay has a compression index ranging between 0.4 and 1 with an average of 0.8. The intermediate clay has a low compressibility of only 0.2–0.3. Both layers are lightly overconsolidated with OCR of between 1.5 and 2 generally due to aging of the clay. The intermediate clay has high OCR of 2–4 owing to desiccation.

The coefficients of consolidation of the three types of clays are shown in Fig. 8.13. The coefficient of consolidation in vertical

direction, cv, was measured using oedometer tests. The cv in the upper marine clay ranges between 0.5 and 1.2m2/year, and in the lower marine clay between 0.6 and 2.2m2/year. The cv of the intermediate clay is highly variable and ranges between 1.2 and as high as 6 m2/year. kv was also interpreted indirectly from oedometer tests and found to range between 10–10 to 10–9 m/s (Fig. 8.14). The permeability reduction ratio of ckv was found to range between 0.3 and 0.87 (Fig. 8.15).The coefficients of consolidation in horizontal

direction, (cu), were obtained from Rowe cell tests. ch of the upper marine clay ranges between 1 and 4 m2/year, whereas in the lower marine clay it ranges between 4 and 6 m2/year. ch values were also estimated from in-situ tests using cone penetrometer (CPTU), dilatometer (DMT), self-boring pressuremeter (SBPT) and BAT permeameter. Some of those measurements are presented in chu et al. (2002). The permeability properties of the three types of clays can be found in Bo et al. (1998e).

8.2.3 Phase 1A

Phase 1 projects started in 1992 with Phase 1A which is located at the southern part of Changi East. The project was completed in 1997. The project, meant to create 501 ha of land with 65 million cubic meters of sand, did not require much soil improvement works, because the foundation seabed clay was rm-to-stiff and overconsolidated. It was felt that placing of additional sandfill of 8–10m may not contribute to significant settlements. However, the project called for dredging of firm-to-stiff clay of 467, 000 m3,which was eventually dumped inside the deep pockets found in the reclamation areas. These areas were treated by preloading with additional sandfill of 1.5–4 m thickness. Owing to the overconsolidated nature of seabed soil, the settlement varied between 0.01 and 0.6 m during the contract period of five years. Since dumped clay inside the deep pockets were firm-to-stiff and had slightly higher permeability, the consolidation process was somewhat faster. Besides, the ll sand migrated into the spaces of the voids, creating shorter drainage paths that helped to accelerate the consolidation process.

A typical time rate of settlement in the surcharge areas with thick clay is shown in Fig. 8.16. The contour showing the distribution of the settlement over 1A is shown in Fig. 8.18.

8.2.4 Phase 1B

The changi East Reclamation Project Phase 1B began in 1993 and ended in 1999. The project required the placement of 75 million m3 of sandfill and the installation of a 28-million-meter long prefabricated vertical drain at the runway, taxiways, and associated areas. The area of land reclaimed was about 520 ha. The prefabricated vertical drains for the runway were installed at 1.5 m by 1.5 m-square spacing and for taxiways at 1.7 m by1.7 m-square spacing. A 4.5 and 3.0m high surcharge were placed at the runway and taxiways, respectively. The duration of surcharge at the runway and taxiways was kept to a minimum of 18 months. Additional soil improvement work was carried out for an adjacent taxiway within 26 months, including soil improvement works for both soft clay and granular soil. For those locations, prefabricated vertical drains were installed with 1.5 m by 1.5 m drain spacing, and the duration of surcharge was limited to 12 months for most of the areas. However, to shorten the period of consolidation, a slightly higher surcharge of 4.5 m in height was placed in areas with clay thickness of less than 20m, and a 6.5 m high surcharge was placed in areas with clay thickness of more than 30m. At two linkways, where the surcharge period was limited to only six months, a higher surcharge loading of a 8.5 m high sandfill was deployed. In addition to the treatment of natural clay with prefabricated vertical drains, the work in Phase 1B involved the reclamation and improvement of clay slurry in an enclosed bund commonly referred to as the siltpond.

The total settlement after the surcharge period along the runway and taxiway is shown in Fig. 8.17. It was found that the thicker the clay, the greater was the total settlement. It was also found that the upper marine clay contributed most of the settlement, and the magnitude of settlement was controlled by the thickness of the upper marine clay. The contour of settlement showing Phase 1B area is shown in Fig. 8.18. The typical time settlement and pore pressure behavior of the thick clay location are shown in Fig. 8.19. Pilot Tests

The first pilot test in Phase 1B was conducted to verify the exhibited design spacings in the contract with a select type of vertical

drain (Colbond CX1000) installed by a typical rig, which was going to be used in the main works. Details of the pilot tests are shown in Table 8.4. It was found that in the first pilot test areas with three different drain spacings, 1.5, 1.7, and 2.0m, a 90% degree of consolidation was achieved in all areas within the specified 18 months of surcharge period. Owing to a favourable variation in the soil profile, the 1.7m × 1.7 m spacing area settled faster than the 1.5m × 1.5m spacingarea. The 2 m by 2 m areasettled the least owing to the occurrence of a thinner layer of compressible soil (Fig. 8.20). Details of the first test area can be found in Choa (1995) and Bo et al. (2000b).

The comparison of prior to reclamation and improved soil parameters are also shown in Fig. 8.21.

The second pilot test was carried out to study whether a drain spacing greater than 2 m could be used without compromising on efficiency. The spacings of 2.5 and 3.0m were tested in two test plots. It was found that spacings wider than 2 m were not so effective and took much longer to achieve the 90% degree of consolidation. Furthermore, the performance of the PVD with 2.5–3.0m spacing was also drastically reduced after two and half years of settlement. Details of the second pilot test area can be found in Bo et al. (1999b, 2000b), and Choa et al. (2001). The settlement and pore pressure monitoring data are also shown together with the construction stages in Fig. 8.22. Soil Improvement in Siltpond

The siltpond is 2000 m long and 900 m wide covering an area of 180ha. The very soft slurry in the pond had an average thickness of 10–15 m with a high water content ranging from 140 to 180%. The water level in the pond was about +3 mCD, and the top elevation of the slurry was about –3 to –4 mCD. The soft slurry had almost no strength and was highly compressible (Bo et al., 1997c, d).

Owing to the negligible undrained shear strength of the slurry, the sand-ll had to be spread in thin layers using a specially designed sand spreader. To ensure the stability of the ll, small lifts of 20cm were used in the first phase of the spreading. The first phase of

Table 8.4 Details of pilot test areas (courtesy of Choa et al., 2001).
AreaNo. of test plotsType of tests plots spacing (m)Type of vertical drainObjective
Phase IB (1st Pilot)4(a) 1.5Cofbond CX 1000To check the performance of vertical drain with typical installation rig and to verily soil and smear parameter.
 (b) 1.7Cofbond CX 1000
 (c) 2.0Cofbond CX 1000
 (d) No drain
Phase IB (2nd Pilot)4(a) 2.0Cofbond CX 1000To check the possible maximum spacing.
 (b) 2.5CofbondCXIOOOTo check the long-term performance.
 (c) 3.0CofbondCXIOOO
 (d) No drain
Phase 1C11(a) 1.5CofbondCXIOOOComparative study on material
 (b) 1.8Cofbond CX1000To check the boundary effect.
 (c) 2.0Cofbond CX1000To check the partial penetration effect.
 (d)1.5Mebra MD7007(Hofland)
 (e) 1.8Mebra MD7007(Hofland)
 (1) 2.0Mebra MD7007(Hofland)
 (g) 1.5Mebra MD7007(korea)
 (h) 1.8Mebra MD7007(korea)
 (i) No drain
Area A (North) (a) 1.5ColbondCX 1000To check the performance of dredged material
 (b) 1.8ColbondCX 1000
 (c) 1.5Mebra MD7007
 (d) 1.8Mebra MD7007
 (e) No drain
Area A (South) (a) 1.5Colbond CX1000To check the performance of exhibited drain spacing and Flexi drain.
 (b) 1.8Colbond CX1000
 (c) 1.5Mebra MD7007
 (d) 1.8Mebra MD7007
 (e) 1.5Flexi FD747
 (f) 1.8Flexi FD747
 (g) No drain

spreading was carried out to +2 mCD. Subsequently, a geofabric of size 700 m by 900 m was laid in the centre of the siltpond. The second phase of spreading was carried out to +4 mCD in lifts of approximately 50cm. The water level was lowered to +2 mCD. After the ll was exposed above the water level, PVD were installed with 2 m by 2 m square spacing as a first pass. The surcharge at the first stage was placed to +6 mCD. The settlement of the ll was monitored. After an appreciable amount of settlement had occurred, a second pass of vertical drains with the same 2 m by 2 m spacing was installed to further improve the drainage. This was necessary because the performance of the vertical drains installed in the first round was greatly reduced by buckling caused by large settlement.

The deterioration in effectiveness of the drains, was indicated by the reduction in the rate of settlement to 20cm/month, although the excess pore water pressure remained very high in the soil. The final surcharge to the level of +9 mCD was placed after the installation of the second round of vertical drains. Details of the reclamation on the slurry-like soil is given in Bo et al. (1998d), and the use of geotextile for such reclamation is given by Na et al. (1998).

Owing to the unexpected behavior of large settlement with little or no pore pressure dissipation and low permeability of the slurry, it took more than three years to complete the 90% of primary consolidation despite the use of PVD at an effective spacing of 1.4m × 1.4 m. During this considerably long period, there was a significant magnitude of settlement although little pore pressure dissipation and effective stress gain (Bo et al., 1997c, 1997d, 2001). The settlement and the pore water pressure dissipation behavior is shown in Fig. 8.23. Recent monitoring data showed that a settlement

of 3 to 7 m had occurred in the siltpond, and about 85% degree of consolidation had been achieved. The pore pressure and settlement behavior at the area where the largest settlement had occurred up to 7 m is shown in Fig. 8.24. The primary consolidation is still in progress in the year 2002. The eld vane shear strength of the soil had also increased from almost zero to 25–30 kPa. The total settlement up to year 2001 is show in Fig. 8.25.

8.2.5 Phase 1C

The Phase 1C project began in 1996 and ended in September 2002. The project involved sandfill of 68 million cubic meters to reclaim 520ha and the construction of a 5.2 km length of shore protection profile at the northern boundary. There was an enormous amount of soil improvement work both for the foundation compressible soil and the granular ll. A total of 49 million linear meters of vertical drains were installed to improve an area of about 400 ha. Settlement contours and settlement along the western part of Phase 1C area are shown in Figs. 8.18 and 8.26 together with the thickness of the clay layers. A typical time rate of settlement and pore pressure measurement of the thick clay layer location are shown in Fig. 8.27. Pilot Area

Work on the pilot area in Phase 1C project was carried out to check the boundary effect and to do a comparative study of different types of PVD. The layout of the third pilot area is shown in Fig. 8.28. It was found that the boundary conditions were the same and that, all drain types caused more or less the same magnitude and rate of settlements, (Fig. 8.29) (Bo et al., 1999a). However, even the same spacing provided different rates of settlement where the boundary conditions were different, as shown in Fig. 8.30(Bo et al., 1999a). Faster rates of settlement were registered near the radial drainage boundary. Several settlement plates and piezometers installed across the pilot area showed that the boundary effect extended up to 30m from the last row of PVD (Fig. 10.31). The pilot area was reloaded again in 1999 to +10mCD, and it was found that additional settlements occurred at a faster rate. This indicates that the

vertical drains are still functioning years after installation. When drains at spacing of 1.2 × 1.2 m were installed in the no-drain area, the rate of settlement increased, and the settlement reached a magnitude equivalent to the earlier drain area in six months, (Fig. 8.32). The complete time rate of settlement and pore pressure dissipation curves together with construction stages are shown in Fig. 8.33. Comparison prior to and after improvement soil parameters is shown in Figs. 8.34(a) to (c).

8.2.6 AREA A (NORTH)

The reclamation for Area A (North) project is confined to the northeast of the Phase 1C project. The project started in 1999 and is

due to be completed in early 2004. The area reclaimed was 70 ha. Since the area will be used for low-rise development, the land was improved to have 90% degree of consolidation equivalent to the fill and an additional 20kPa load. A sand volume of 12 M m3 was used, and 13 million linear meters of prefabricated vertical drains were installed.

The project area was underlain by a thick layer of clay up to 50m thickness, and the seabed was as deep as –15 to –17 mCD. Surcharge of up to +10mCD was placed at the C2 area. In addition, the higher surcharge of +12 mCD was placed at the C1 area owing to the higher future load requirement. Thus as much as 29 m of fill load was placed on the thick layer of compressible clay. Therefore, settlement as high as four meters was registered in the C1 area as shown in Fig. 8.18. Time versus settlement and pore pressure versus settlement curve for a typical area are shown in Fig. 8.35. Pilot Area

This pilot test was carried out to assess the performance not only of the exhibited drain spacings but also of the dredged material

improved with the prefabricated vertical drain. The layout of the pilot area is shown in Fig. 8.36. The Pilot area was installed with Mebra MD7007 manufactured in Holland and Malaysia. Two different drain spacings of 1.5m × 1.5mand 1.8m × 1.8 m were installed together with a control area with no vertical drain. In addition, the dredged spoils from sand key of the Phase 1C area were dumped in the deep pocket inside the future Area ‘A’ North area and capped by placing a sand blanket of 1–2 m thickness.

Since dredged materials are deposited in a thin layer bounded by double drainage layers, primary consolidation was expected to be completed in a reasonably short time. However, large settlements from such layer could be expected owing to greater stress ratio at the top part of the foundation soil. Details of reclamation using dredged material can be found in Bo et al., (2001). The pore pressure and settlement behavior of dredged material under fill and surcharge load is shown in Fig. 8.37. From the characteristics of the deposited material at the borrow source or after disposal, it was estimated that it would contribute a settlement of about 1 m. Although there were slight changes of soil characteristics together with formation of inter-lump voids, the measured settlement was found to be also about 1 m. This confirms that the settlement contributed from the closing up of the inter-lump voids is minimal in the study on this type of lumpy clay used for reclamation. Comparison of soil parameters before and after improvement is shown in Fig. 8.38. Complete time settlement curves and pore pressure curves at the Area A (North)

Pilot area are shown in Fig. 8.39. It was also found that both areas with Mebra Malaysia and Holland showed a similar settlement. The 1.5 m × 1.5 m drain spacing locations settled faster than the 1.8 m× 1.8 m drain spacing areas.

8.2.7 AREA A (SOUTH)

The Area A (South) Project was carried out at the southeastern part of Phase 1C. The project started in 1999 and is due to be completed in 2005. The area reclaimed was about 450 ha. The specification for the degree of consolidation was the same as that for the C2 area in the Area A (North) project. A sand volume of 52 M m3 was used, and 50millio n linear meters of prefabricated vertical drains were installed.

The project area covers almost the whole stretch from south to north of the Changi East project area. Therefore, the thickness of the soil profile varied along the area. The thickest layer of clay was found at the northern part of the site and the thinnest layer at the southern part. The profile of the soil along the north-south line is shown in Fig. 8.40. In this project, in addition to the naturally deposited marine clay, thin layers of 2–6 m thickness of hydraulically dumped clay was found inside the ‘containment bund’ area from a previous project. The material in this area has high moisture content and is in an ultrasoft condition. However, the material was deposited on the underlying sand layer and was also capped with 2–3 m thickness of sand. Therefore, direct hydraulic filling over the area was possible.

Even for an ultrasoft material, the settlement process may be faster because the material was deposited in thin layers. The consolidation process may start earlier, and a single pass of vertical drain will be working properly due to a smaller magnitude of settlement. The pore pressure and settlement behavior of thin layers of ultrasoft clay under fill and surcharge load are shown in Fig. 8.41. It can be seen in the figure that owing to the short drainage path of the deposited slurry, the pore pressure dissipated during filling. The situation is similar to the constant rate of loading consolidation test. Pilot Area

A pilot test was carried out with two types of drain spacings using two types of prefabricated vertical drain, namely Colband CX1000 and Flexi FD747. The layout of the pilot area is shown in Fig. 8.42. The time settlement curve and pore pressure measurement at the drain location are shown, together with the no-drain location (Fig. 8.43). It was found that both drains perform satisfactorily. Comparison of performance of drains is shown in Fig. 8.44.


The Hong Kong government constructed a new airport at Chek Lap Kok. A comprehensive site investigation was carried out, and a test embankment was built in 1981. The purpose of the test embankment

was to determine the feasibility of the techniques to be used to accelerate settlement of the soft clay layer while at the same time minimizing the formation of mud waves. The test embankment was also to provide information on the quarrying and handling of excavated rock for filling, seawall construction and the stability of the dredged seawall trenches, general construction techniques, and the behavior of the soil strata. Owing to the delay in constructing the airport, a unique opportunity was presented for acquiring continuous monitoring data over about nine years in the test embankment for soil improvement by vertical drains and surcharge.

In 1990, the Geotechnical Control Office (GCO) of the Civil Engineering Services Department (CESD) carried out further field work at Chek Lap Kok. The investigation focussed on an extension area, necessitated by the possibility of the airport having a 1300-meter runway separation as an alternative to the previously planned 900- meter separation. This work was described in the GCO report titled ‘New Airport at Chek Lap Kok, Geotechnical Investigation (1990) Vol. 1’ (Premchitt et al., 1990).

In the report, the GCO updated the monitoring data at the test embankment and reviewed all the previous geotechnical investigations done at Chek Lap Kok including the test embankment. A limited programme of soil sampling, laboratory and field tests at test embankment was also carried out in 1990b y the CESD

to update the soil information under the embankment. There havebeen numerous reviews of the Chek Lap Kok test embankment inconferences and journals by various authors since its completion in 1983. Many of these publications have been referred to by Premchitt et al., (1990).

Choa, Wong and Low, in collaboration with Maunsell Consultants (Singapore) Pte Ltd and Maunsell Geotechnical Services Ltd (Hong Kong), reviewed the available data from the site investigations as well as the test embankment (Choa et al., 1990). Prior predictions and back analyses of the settlement data from the test embankment and haul road were carried out, and recommendations for design parameters and drain spacings were made. Correlations were also obtained between various soil parameters by collecting all the soil test results from the past site investigations and entering them into a common database. Piezocone data were found to be a very useful supplement to borehole data for establishing soil profiles and drainage boundaries so that site specific settlement predictions can be made.

Two computer programs were used for predicting settlements. A finite difference program CONSOL 99 was used for one-dimensional consolidation of the layered soil deposit. This program was used in cases where there were no vertical drains. Where vertical drains were installed, program VDRAIN was used. VDRAIN is an extension of the conventional method for analyzing vertical and horizontal consolidation due to the presence of vertical drains. The method of analysis in this program is based on Barron’s solution for equal strain consolidation due to radial flow of water and Carrillo’s equation for combining consolidation due to vertical and horizontal drainage. The program takes into account:

(i) Stage loading;

(ii) Load reduction due to fill submergence;

(iii) Delayed vertical drain installation;

(iv) Changes in length of vertical drainage path with time; and

(v) Variations of soil stress history with depth.

Details of the program are explained in Chap. 2. ‘Prior predictions’ (i.e., predictions of settlements based on soil parameters

derived from laboratory tests and idealized soil profiles from piezocone data) were carried out with data from the test embankment using CONSOL and VDRAIN. The purpose of this exercise was to test whether the proposed method of analyses using conventional design practice was sufficiently accurate in predicting settlements.

The predicted settlements derived from field and laboratory test data at the test embankment, assuming no knowledge of the monitored performance, were then compared to the field measurements. If the agreement was reasonable, it would be a good indication that the approach and method of analyses adopted were appropriate for this study and might be used to predict both the rate and magnitude of settlements of the entire airport site under varying thickness and duration of fill loads.

The prior predictions were expected to show some disagreement with the actual performance of the test embankment. The reasons for the discrepancies were then explored by back analyses. In the back-analyses, the additional information from monitored performance was used to calibrate the input parameters so as to match the measured field behavior. Design parameters were then established from this exercise. Results of the prior predictions and back analyses are discussed by Choa et al., (1990).

A conventional approach for predicting the reclamation settlements and designing vertical drains was adopted because the accuracy of the predictions depended on the input parameters which, even in the relatively simple approach proposed, required 13 or more values of individual parameters. The uncertainties in determining these parameters and the probable variation of the parameters from location to location did not justify a more sophisticated and complicated analytical approach for design purposes where eventually some degree of engineering judgement based on experience is required. A pragmatic approach was therefore advocated using the time-tested conventional method of analyses with some improvements. To understand how the variation of the various parameters would effect the settlement predictions, some parametric and sensitivity studies were carried out. These sensitivity studies indicated that it was advisable to estimate the upper and lower bound time rate of settlement by considering the uncertainties associated with the input parameters.

This would provide estimates of the likely range of time required for preloading or surcharging for various drain spacings.


The Tianjin port is situated on the western side of the Bohai Gulf in the People’s Republic of China. The port was rapidly expanding its facilities to cope with the substantial increase in international trade. The East Pier Project was a major port development project and part of the first phase of port expansion.

The first phase comprising four timber berths, one building materials berth, and the back-up area were to be constructed on the south side of the reclaimed land. These berths were designed for conversion to container berths in the future. The second phase comprised the construction of six general cargo berths at the northern side of the reclaimed land. The berths would be constructed with narrow peripheral concrete decks with retaining structures at the edge of the reclaimed area.

Reclamation work for the project was carried out by dredging the soft silt from the adjacent harbour basin and depositing it within earth bunds constructed to retain the fill. The reclamation was carried out within the first cofferdam to an elevation of +3.5 mCD between April 1982 and June 1984. After construction of the second cofferdam and diaphragm dam, the reclamation was raised to about +6.5 mCD between November 1984 and October 1986 (Figs. 8.45 and 8.46). The reclamation fill was deposited on the existing tidal.ats. These tidal.ats were underlain by layers of soft silty/clayey material interbeded with thin lenses of sand of varying thickness. Below the soft material was a good bearing layer of dense fine sand. The thickness of the soft material below the original seabed varied from 17 to 30m depending on the locality. The reclaimed land was topped up with reed mats followed by 400 mm of hill cut and 300 mm–500 mm of sand to form a working platform. The reclamation works including the working platform were completed in April 1987.

Two contracts were awarded for the first phase. The ‘Wharf’ contract comprised the construction of a 950m piled deck, dredging

works, construction of an L-shaped retaining wall over soil stabilized by deep cement mixing, back filling behind the wall, and improving the soft soil beneath the back fill with vertical drains in combination with surcharge preloading. The ‘Soil Improvement’ contract was for the improvement of the soft soil with vertical drains and preloading of all the back-up and stacking yard areas extending from 120m behind the seaward edge of the wharf deck to the northern boundary of the second cofferdam.

The soil within and under the reclaimed land was to be improved to enhance its bearing capacity to sustain the heavy loading from the stacking yard and to accelerate the time-dependent consolidation settlement of the underlying soft soil strata. The soft soil improvement works comprised the treatment of approximately 600,000 sq. m. of reclaimed land with approximately 6000 km of prefabricated bandshaped drains. The soil improvement works began in June 1987. The contract period was 38 months with partial handover of the site zones, starting 15 months from the start of the works.

In the proposed method of soil improvement, band drains were installed on a square grid of 1.3 m×1.3 m. The depth of the drains

was about 20m below the surface of the reclaimed land. The specified surcharge load intensity varied from 70to 100 kPa. The contractor proposed to provide the surcharge load intensity by either vacuum preloading alone or by vacuum preloading in combination with surcharge fill preloading. As this was an alternative design proposed by the contractor, he was required to demonstrate in a pilot test the feasibility of his design. The performance of the vacuum preloading method was compared with that of the conventional earth fill surcharge in a pilot test.

The vacuum preloading method required the placement of an airtight membrane over a sand blanket. The membrane was sealed in slurry trenches or impermeable soil around the edges. Suction tubes in the sand blanket were put through the membrane, sealed, and connected to vacuum pumps (Fig. 8.47). To enhance the effectiveness of the vacuum method, prefabricated vertical drains were installed

prior to the placement of the air-tight membrane. These vertical drains enabled the vacuum to be felt at depth in the soil mass. The ‘negative’ pressure created by the pumps caused the water in the pores of the soil to move towards the surface and towards the drains because of the hydraulic gradient setup. The principle behind the vacuum method is illustrated in Fig. 8.48. Instead of increasing the effective stress in the soil by a surcharge, the effective stress is increased by reducing the pore pressure in the soil. The initial and final pore pressures are designated uo and uoo, respectively, in Fig. 8.48. The pore pressure at time t is shown as ut.

In the pilot test area the soil profile comprised a 4 m thick hydraulic fill underlain by 15 m of silty clay followed by 7 m of clayey sandy loams which in turn was overlain by sand. The hydraulic fill was a dredged silty clay with clayey silt laminations. The particle size distribution, frequency, and thickness of the clayey silt laminations depended greatly on the distance from the discharge points of

the hydraulic fill. As these discharge points were situated mainly on the south side of the reclamation and the outlet points in the northern side, the particle size distribution was finer in the northern half of the site. Soil properties determined before soil improvement can be found on Fig. 8.49. The results of field vane shear tests carried out before and after preloading are shown in Fig. 8.50.

The pilot test zones of a minimum area 50m ×50m each. Four of the zones were for testing surcharge preloading. Three of the four zones were installed with prefabricated band drains at 1.3, 1.5, and 1.8 m square spacings. The fourth zone was left without drains to act as a control zone. The.fth zone was for testing vacuum preloading with drains at 1.3 m spacing. The depth of all the drains

was at elevations –14 mCD. The specified preloading intensity was 97 kPa. A hill cut gravelly soil was used for providing the earth fill surcharge. The maximum preload that could be obtained by vacuum was between 80and 90k Pa. The contractor therefore proposed to make up the difference by combining vacuum with a small earth fill surcharge.

The prefabricated vertical drains were manufactured in China and came mainly from factories in Nanjing and Tianjin. Their dimensions were 100 mm × 4 mm with a continuous plastic core wrapped in a nonwoven geotextile material.

The progress of the treatment works was monitored through extensive instrumentation, sampling, field and laboratory testing. Typical settlement and pore pressure results are shown in Figs. 8.51 and 8.52. A more detailed account of the project can be found in Choa (1990).

The pilot test demonstrated that vertical drains used in combination with vacuum preloading was suitable for soil improvement of the very soft silty clay in the Tianjin East Pier Project.

After some initial difficulties, a very good seal was obtained and a vacuum of up to 90k Pa was successfully maintained. The effect of the vacuum preloading was left within the depth of the vertical drains which were installed to about 20m below ground level. The vacuum preloading appeared to be less effective at depths greater than 14 m. This could be caused by a loss of vacuum in the relatively permeable clayey silt lenses or due to a relatively impermeable clayey stratum below 14 m. The effectiveness of the vacuum decreased at the perimeter of the vacuumed area. Particular attention was therefore paid to the boundary between various treated areas to ensure adequate preloading. The separation between treated areas was kept as small as possible. Additional surcharge was placed on the separation bunds to make up for the shortfall in preload.

In the extremely soft soil conditions found in this project, the vacuum method enabled the preload to be applied very quickly without any soil instability, whereas an earth fill surcharge had to be applied in stages to prevent slope failure. The surcharge method was comparatively slower in the pilot test. As there was a general shortage of surcharge fill material at this site, the conservation of fill material became an important consideration in choosing the preloading method. The decision was therefore made to treat only the area up to 120m behind the wharf front, where the settlement criteria was more stringent, with surcharge preload, and to use a combination of vacuum and surcharge preload for the remaining areas.

Eventually, even this area adjacent to the wharf was converted to vacuum preloading because of the extremely time-consuming rest period required for the staged construction of the earth fill preload.

Verification of the performance of the soil improvement works area by area indicated satisfactory improvements in both compressibility and shear strength of the soil. The project was successfully completed on schedule. The wharf was put into use at the end of 1990. Another vacuum preloading project is presented in the next section.


An oil storage station was constructed in 1996 near the coast of Tianjin, China, on a site recently reclaimed using clay slurry dredged from the seabed. The soil had a high water content, was very soft, and was still undergoing consolidation. The seabed soil on which the dredged slurry was deposited was also soft. The site soil conditions needed to be improved before any construction work could be carried out.

Several soil improvement schemes were considered. preloading using a fill surcharge was not feasible because it was difficult to build a fill embankment several meters high on soft clay. Vacuum preloading was adopted because it was considered the most suitable and cost-effective method for the project.

8.5.1 Site Conditions

The site for the oil storage station is shown in Fig. 8.53. It covered a total area of approximately 50, 000 m2. For the purpose of soil improvement, the site was divided into two sections: Section I of 30, 000 m2 and Section II of 20, 000 m2, as shown in Fig. 8.53.

The soil profile included two layers that required improvement. The first layer was soft clay consolidated from dredged slurry. It was about 4–5 m thick. The second layer below this was seabed marine clay. It was about 10–16 m thick and was underlain by a sti. sandy silt layer. The marine clay layer was further divided into three sublayers in accordance with the USCS classiffication system: a low plasticity silty clay (ML) layer of 2–4 m thickness, a low-to-medium plasticity clay (CL) layer of 7–8 m thickness, and a low plasticity silty clay (ML) layer which was relatively stiff. The basic engineering properties of the soils are shown in Fig. 8.54(a) and (b) for Sections I and II, respectively. It can be seen from Fig. 8.54(a) and (b) that except at the bottom of the marine clay layer, the water content (W) of the soil was generally at or above the liquid limit (LL), and the undrained shear strength (cu) of the soil was generally low.Thecoe.cient of consolidation in the horizontal direction (ch) was in the range of 1.1−4.7 × 10−3 cm2/s.

8.5.2 Soil Improvement Works

Treatment of the soft soil was required to meet the following specifications:

(a) a minimum bearing capacity of 80k Pa,

(b) an average degree of consolidation greater than 80% under a minimum surcharge of 100 kPa,

(c) an average settlement for a consecutive 10-day period less than 2 mm/day.

The soil improvement work was carried out as follows. The ground surface was very soft. To overcome this, a 2 m thick partially dried clayey fill with a 0.3 m sand blanket on top was first placed. PVDs were then installed on a square grid at a spacing of 1.0m to a depth of 20m. Corrugated.exible pipes (100m m diameter) were laid horizontally in the sand blanket to link the PVDs to the main vacuum pressure line. The pipes were perforated and wrapped with a permeable fabric textile to act as a filter layer. Three layers of thin PVC membrane were laid to seal each section. Vacuum pressure was then applied using vacuum pumps.

The schematic arrangement of the vacuum preloading arrangement used is shown in Fig. 8.55. Vacuum pressure was applied continuously for four months until the required degree of consolidation was achieved and the settlement rate for a consecutive 10-day period became less than 2 mm/day. The average vacuum pressure was 80k Pa. The total surcharge applied was about 120 kPa including the 2 m of fill and 0.3 m sand blanket.

8.5.3 Results

After the PVDs were installed, instruments such as pore water pressure gauges, surface settlement plates, multilevel settlement gauges,

piezometers, and inclinometers were installed in both sections to monitor the system performance. The instrumentation scheme is shown schematically in Fig. 8.53 (plan view) and Fig. 8.56 (elevation view). Undisturbed soil samples were taken, and field vane shear tests were conducted both before and after the soil improvement. Settlements

During PVD installation. It was observed that the ground settled 14.9 cm in Section I and 26.8 cm in Section II during the installation of the PVDs. This was due to further consolidation of the slurry clay under the in.uence of the surcharge effectof the 2 m of fill and 0.3 m sand blanket.

During vacuum preloading. The settlement at different depths was monitored during vacuum preloading. The monitored settlement versus time curves for both sections are presented in Figs. 8.57(a) and (b). The surface settlement at the end of preloading was 85.6 and 92.5 cm for Sections I and II, respectively. Pore water pressures

Under vacuum load, the pore water pressure in the soil decreased. The reductions in the pore water pressure at different depths are plotted versus time in Fig. 8.58(a) and (b) for Sections I and II, respectively. Generally, the pore water pressures at various depths approached constant after 1.5–2.5 months of vacuum treatment. The initial and final pore water pressures together with hydrostatic pore water pressures are plotted in Fig. 8.59(a) and (b). It can be seen that the final pore water pressures matched the amount of suction applied throughout the full depth, indicating that the vacuum preloading was very effective. Degree of consolidation

The consolidation of soils under vacuum preloading was very effective in both Section I and II. Nearly 1 m of settlement was induced, and this occurred throughout the thickness of the soft soils (see Fig. 8.58(a) and (b)). The rate of settlement decreased towards the end of the vacuum preloading period, indicating that a substantial part of the ultimate consolidation should have taken place within the four-month loading period. The ultimate settlements for both Section I and II were estimated using the hyperbolic method, a procedure described by Sridharan and Rao (1981) and Tan (1993). From the monitored settlements and the ultimate settlements calculated,

the average degree of consolidation for the whole layer and each sublayer was calculated, and the values are given in Tables 8.5 and 8.6 for Section I and II, respectively. It can be seen that the average degree of consolidation calculated was greater than 80% for all layers.

The suction line, which is the line given by the hydrostatic pore water pressure line minus the suction of 80k Pa, is also plotted in Fig. 8.59(a) and (b). It can be seen that the final pore water pressures were close to the suction line. The degree of consolidation at

Table 8.5 Average degree of consolidation calculated for different layers at Section I
Elevation (m)+6.4 to +4.2+4.2 to -1.1-1.1 to -3.6-3.6 to -11.3-11.3 to -13.5+6.4 to -15.0
Degree of consolidation %8710010086.38084
Table 8.6 Average degree of consolidation calculated for different layers at Section II
Elevation (m)+4.7 to -0.4-0.4 to -4.3-4.3 to -7.1-7.1 to -9.7-9.7 to -12.6+6.4 to -14.0
Degree of consolidation %9710086.382.881.087

a given elevation, U(z), can be estimated as (1 − Δuf/80)%, in which .Δuf is the excess pore water pressure at the end of vacuum preloading and is measured as the di.erence between the final pore water pressure and suction line as shown in Fig. 8.59(a) and (b). Increase in undrained shear strength

Field vane shear tests were conducted before and after vacuum preloading in both Section I and II, and the results are presented in Fig. 8.60(a) and (b). It can be seen that the undrained shear strength after vacuum preloading increased 2–3 fold for both sections. Lateral displacement

The vacuum load caused an inward lateral movement in the soil. The lateral displacements monitored at various depths and at different times are presented in Fig. 8.61 for both Sections I and II. The lateral displacement at the south boundary of Section II was not recorded owing to instrument failure.

As shown in Fig. 8.61, the lateral displacement was greatest at the ground level, and it reduced sharply with depth. The lateral

displacement at the ground level is plotted versus time in Fig. 8.62. It can be seen that the ground lateral displacements measured for Sections I and II on the north boundary were quite consistent. The displacement measured on the south boundary of Section I was higher compared with that on the north. Unlike the surface settlement, the rate of the lateral displacement did not reduce rapidly with time. This indicated that the ground lateral displacement was effected not only by the primary consolidation but also by the shearing and secondary consolidation processes. At the end of the vacuum preloading period, the lateral displacements were up to 48.3 cm in Section I and 31.7 cm in Section II. Cracks had developed on the ground surface at about 10m away from the edge of the preloaded area. As there were no adjacent buildings or facilities, the lateral displacement and cracks were negligible. However, at sites with adjacent structures, lateral displacements can cause problems. The effect

of lateral displacement and risk of damage to adjacent structures is often an important factor that needs to be considered when using the vacuum preloading method for soil improvement work.

8.5.4 Concluding Remarks

This case study has shown that the PVD method can be combined with the vacuum preloading method in consolidating soft clay. A vacuum distribution system comprising PVDs at a square grid of 1.0m together with horizontal 100 mm diameter corrugated.exible collector pipes was effective in distributing the vacuum pressure. A vacuum pressure of 80k Pa was maintained throughout the whole 20m depth of soft clay.

About this article

Case Studies

Updated About content Print Article