Chapter 14
Quality Management in Prefabricated Vertical Drain Projects

14.1 SELECTION OF VERTICAL DRAIN
14.2 PLANNING AND SELECTION OF VERTICAL DRAIN RIGS
14.3 TYPES OF MANDREL
14.4 TYPES OF ANCHOR
14.5 INSTALLATION
14.6 QUALITY CONTROL ON PREFABRICATED VERTICAL DRAIN MATERIAL
14.7 VERIFICATION OF FIELD PERFORMANCE

Quality management is one of the most important tasks in prefabricated vertical drain (PVD) project. The expected results can be ensured only if a good quality control system is established. The quality management of PVD works consists of several processes, starting from the selection of vertical drain material to performance verification in the field. The system includes controlling the quality at every stage of the process. The various stages are: (i) selection of material according to the specifications, (ii) planning and deployment of suitable types of vertical drain rigs, including special types of rigs required for difficult ground conditions, (iii) selection and quality control of accessories such as mandrel and anchor, (iv) quality control on PVD material, (v) quality control during installation, and (vi) assessing the performance of the PVD.

14.1 SELECTION OF VERTICAL DRAIN

There are several types of vertical drains. However, most vertical drains have basically the same dimensions of 100 mm width and 4 – 6 mm thickness. Vertical drains are fabricated with drainage cores and filter jackets. Most drains provide sufficient discharge capacity to drain out the pore water from compressible soil and an efficient filter which can retain the surrounding fine grain soil. The cross-section and shapes of various types of drains and cores are shown in Figures 14.1 and 14.2 respectively. Table 14.1 shows the various specifications called by different clients.

Even though some specifications are called to suit the actual site conditions, the environment, geotechnical parameters, and depth to install the vertical drain, it has never been possible to obtain particular types of vertical drains which fulfil client’s or designer’s requirements exactly. Thus, selection has to be made from the vertical drains available in the market which have specifications close to requirements.

14.2 PLANNING AND SELECTION OF VERTICAL DRAIN RIGS

The selection of a suitable type of installation rig is essential in implementing a vertical drain project.

The types of rig normally used in PVD projects are:

The size and weight of the rig should be suitable for the prepared platform with a certain load bearing capacity. The mobilization of an overweight rig would lead to instability of the equipment. On the other hand, a lightweight rig may not provide sufficient installation power and reaction against penetration resistance, and therefore may not have the desired penetration force.

Other factors to consider are the depth to install and the type of soil. Rigs of a suitable height are also important, depending upon the depth of installation. Installation rigs with heights of 20m – 54m can be found in the market. Based on the type of soil expected to be encountered, the capacity and type of rig selected are important. In addition, some special installation equipment should be made available in case there are difficulties during installation of the vertical drains. Table 14.2 shows a specialized rig used for troubleshooting in the Changi East reclamation projects. Figure 14.3 shows various types of drain installation rigs used in the Changi East reclamation projects, and Figure 14.4 shows various types of special equipment used in the same project.

14.3 TYPES OF MANDREL

Basically, there are four types of mandrel:

• Rhombic
• Rectangular
• Square
• Circular

The first two types are most commonly used and the last one is least commonly used. The shapes of the mandrels are shown in Figure 14.5. On the one hand, a mandrel must be strong enough to be able to penetrate the formation vertically, and on the other hand, the mandrel must not be so big that it disturbs the soil to a great extent. Normally, a smaller rhombic mandrel is preferable since soil disturbance and smear effect caused by such mandrel is minimal.

However, to penetrate firm to stiff soil, a rhombic mandrel may not have enough stiffness, and a rectangular mandrel made with stronger and thicker steel would be more suitable.

14.4 TYPES OF ANCHOR

An anchor serves two purposes. It must be strong enough to anchor the vertical drain on firm ground. It must also be capable of preventing the soil from ingressing into the mandrel. The available types of anchors are as follows:

• Steel bar
• Flexible metal plate
• Vertical drain material

Normally, small steel bars are preferable since disturbance to soil during penetration is less. To use the steel bar, the mandrel has to be equipped with a mandrel shoe with an opening capable of being adjusted to fit the steel bar dimension.

14.5 INSTALLATION

Vertical drains should be installed on a land platform that is graded as a flat plane to ensure a verticality of drain. The verticality of the rig can be checked with a spirit level. A vertical drain should normally be installed at platform level, that is, one meter above the groundwater table. Groundwater table is taken to be the high tide level after reclamation. However, in some vertical drain projects, vertical drains are installed offshore. In addition to the verticality of the rig, it is important to maintain the mandrel in good and straight condition. Any bending of the mandrel at the tip would cause the

vertical drain to deviate from the vertical. The operator should also be careful when installing vertical drains which have to penetrate through layers of different density. Density changes between two layers can cause the mandrel to slide along the boundary and the vertical drain can deviate from the vertical. Normally, vertical drain installation points are set out using anchors.

14.6 QUALITY CONTROL ON PREFABRICATED VERTICAL DRAIN MATERIAL

Generally, drains are delivered in rolls of 200–300 meters in length. A single delivery may consist of one million meters of drains in ten containers. Therefore, full-scale testing of the drain material by an accredited laboratory for every one million meter length of drains is usually carried out as they generally come from the same batch in the manufacturing process. An insitu specialized laboratory is also usually set up to check on the discharge capacity, tensile strength, AOS and permeability.

The consistency of quality of the drains should be monitored and controlled based on the results from the laboratory. Details of quality management in PVD works in land reclamation projects are widely discussed by Bo et al. (2000f). The types of measurements required and the factors affecting the measurement are discussed in the following sections.

14.6.1 Dimension of the drain

The dimensions of drains supplied by various PVD manufacturers are now standardized at 100 mm wide and 4 mm thick, except for some large dimension vertical drains produced for special purposes by some manufacturers.

The specifications of most projects, however, require the drain dimension to be 100 ± 2 mm wide and 3—4 mm thick. Most drains comply with the specified values although some have been found to be slightly

lower than the specified margin. The effect of drain dimension on the time required for consolidation is shown in Figure 14.6. It can be seen from the figure that variation in time required for 90% consolidation, with specified spacing caused by variation of dimension to a maximum 6 mm in width and 2 mm in thickness in certain types of soil, is found to be only 12 days. The maximum percentage difference in duration for one (1) year is only 3%. However, it may be necessary to check the thickness of the vertical drain under pressure. Since a reduction of drain thickness will lead to a reduction of discharge capacity of the drain, it can affect the performance of the drain. At the laboratory in the Changi East reclamation project, the measurement of the width was carried out with a vernia scale and variation of thickness under pressure was measured with simple shear equipment which was able to measure accurately the vertical displacements under different loads. Figure 14.7 shows the variation in the thickness of vertical drains caused by normal pressure. It can be seen in the figure that the variation of thickness of the Colbond drain is greater than the Mebra type.

14.6.2 Apparent opening size (AOS)

The apparent opening size (AOS) is specified for filter material. Since the filter of the drain is to prevent the fine grain soil from entering into the core and yet provide sufficient permeability, an AOS of O₉₅ less than or equal to 75 mm is required. For a woven or non-woven geotextile, if the AOS is smaller than D₈₅ of the surrounding soil, piping will not occur. Generally

D₈₅ of natural soft clay is greater than 75 mm. Therefore, specifying O₉₅ of 75 mm is sufficient to retain the surrounding soil. On the other hand, AOS of O₉₅ or O₁₅ should also be large enough to prevent clogging. The following criteria are suggested to prevent clogging:

The AOS tests are carried out using standard glass beads with diameter of 40–170 mm. The percentage of glass beads retained on the filter cloth is calculated from the following equation:

where B represents the beads passing through the specimen (%)

T & P are the total mass of glass beads used & mass of glass beads in the pan in grams respectively.

The AOS is obtained from the grain size distribution curve provided by the glass beads manufacturer (Figure 14.8). The test method follows ASTM D4751-87. Some manufacturers use the optical method to measure the AOS.

14.6.3 Tensile strength

The drain should be able to withstand the tensile stress caused by the drain installation process. Elongation of the drain may also occur. Therefore, the vertical drain should have the required tensile strength at an allowable elongation which could more or less maintain the dimension of the drain without major deformation.

For the Changi East reclamation project, the tensile strength of the vertical drain has been specified as 100 N/cm, or 1 kN/10 cm, at 10% elongation for both dry and wet conditions. However, the actual elongation test carried out with a vertical drain installation rig shows that elongation of the Mebra MD7007 is as low as 1%. It indicates that the stress occurring as a result of the friction between the roll of the vertical drain during penetration is insignificant. However, Bergado (2000) has reported that stress on the PVD caused during extraction of the mandrel is much higher than that during installation.

Typical tensile strength test results of the Mebra MD7007, Colbond CX1000, and Flexi FD767 under wet conditions are shown in Figure 14.9. It can be seen from the figure that the strain of the drain at a specified tensile strength is normally lower than the specified strain of 10% for most drains. The tensile strength tests at the site laboratory were carried out with a modified triaxial compression machine. The vertical drain was gripped across the whole section at the two ends. The size of the jaw face was 140mm in width and 50mm in height. Vertical displacement was measured during extension with a linear vertical displacement transducer and stress incurred was measured using an extension proving ring. The tensile strength

tests were carried out under a strain rate of 7% strain/min on a PVD sample with a gauge length of 200 mm. Figure 14.10 shows tensile strength testing in progress at Changi East reclamation project. The method of testing followed ASTM, D4595D:1988 except that the test was carried out under high strain. However, the effect of the strain on the strength of geotextile was found to be insignificant.

14.6.4 Permeability

To meet the permeability requirement, Holtz et al. (1991) has suggested that the permeability of geotextile should be at least 10 times more than the surrounding soil. Permeability tests can be carried out with a simple constant head permeability apparatus. The apparatus used is similar to the apparatus specified in the ASTM D4491-85. The device consists of an upper and lower unit which can be fastened together. The sample can be positioned between the two units. There are manometers connected to the upper and lower units for water supply and head measurement. Permeability tests were carried out under various head differences. Figure 14.11 shows the permeability test apparatus used in the Changi East project. The Darcy permeability measured for different drains are shown in Table 14.3.

14.6.5 Discharge capacity of drain

Discharge capacity is an important parameter that controls the performance of prefabricated vertical drains. Only PVDs that have sufficient discharge capacity will function well. There are two major problems related to the discharge capacity of vertical drains. The first is the determination of the required discharge capacity (q_w) to be used in the design (Holtz et al. 1991). Another is the measurement of the discharge capacity of the drain in the

laboratory and in the field. The first problem is due to a lack of sufficient field data and the second is the variation of discharge capacity with lateral stress, and buckling of drain and siltation, etc.

14.6.5.1— Definition of discharge capacity

Discharge capacity is defined as the rate of water flow per unit of hydraulic gradient.

where q_w is the rate of water flow per unit hydraulic gradient in m³/s, Q is the average quantity of water discharged per unit time (m³/s), i is the hydraulic gradient and is dimensionless, l is the sample length and h is the head difference over the sample length of water in the meter.

Since the discharge capacity is dependent upon the water flow rate, it is measured under a temperature of 20 °C (68°F).

ASTM D4716-87 is usually adopted in the determination of constant head hydraulic transmisitivity (in-plane flow) by the laboratory for geotextile and geotextile products. A discharge capacity parameter can be obtained as by product. More specific and relevant methods have been proposed by Ali (1991), Miura et al. (1993), Kamon et al. (1994), Bergado (1994) and Chu and Choa (1995) with the use of different apparatus to measure the discharge capacity. Details on the NTU discharge capacity apparatus developed by Chu can be found in Bo et al. (2003a). Photographs of the NTU discharge capacity apparatus for both straight condition tests and buckled condition tests are shown in Figure 14.12. Different apparatus, and methods of testing often give different values of discharge capacity. This section discusses the determination of the required discharge capacity and factors that affect the test results.

Required discharge capacity was proposed by Mesri and Lo (1991) as five times the discharge factor (D) based on previous analysis data for three major embankment projects. The discharge factor is defined as:

14.6.5.2— Determination of required discharge capacity

The required discharge capacity is q_w = 5K_h l²_m, where Kh is horizontal permeability of soil, l²_m is maximum drainage length. For most clay, the required discharge capacity varies from 2 to 80 m³/yr. Kamon et al. (1994) has defined the required discharge capacity as follows:

where q_W(req) is the required discharge capacity in cm³/d, T_h is thedimensionless time factor for radial drainage, C_h is the horizontal coefficient of consolidation is cm²/day, and l is the length of the PVD in cm.

Since consolidation involves the dissipation of water, the total amount of water dissipated is dependent on the compressibility and thickness of the soil. Den Hoedt (1981) has stated that the required discharge capacity for a 100 mm wide drain should be at least 3 x 10^-6 m³/s, based on an allowable settlement of 40 mm/day for a 30 m long drain. Dutch has recommended a minimum q_w of 5 x 10^-6 m³/s under a hydraulic gradient of 0.6 and a confining pressure of 100 kN/m². Kremer et al. (1982) has proposed q_w of 25 x 10^-6 m³/s at 10°C under a hydraulic gradient of unity and confining pressure of 15 kN/m². Holtz et al. (1991) recommended that the q_w be between 3 x 10^-6 m³/s and 9 x 10^-6 m³/s under 300-500 kN/m² pressure.

A summary of the discharge capacity specified in a number of soil improvement projects is presented in Table 14.4. It can be seen that the specified discharge capacity ranges from 5 to 100 x 10^-6 m³/s under straight conditions and from 6.3 to 32.5 x 10^-6 m³/s under buckling conditions.

As a rule of thumb, the total volume of water to be discharged can be estimated using the following equation:

where Q_v is the volume of water drained out from the soil in m³, l is the drain length in meter, ε_v is the volumetric strain, and D_e is the equivalent diameter of the drain in meters. If the time to complete the primary consolidation is known, the average flow rate can be calculated as:

The initial hydraulic gradient can be estimated using the initial excess pore pressure and the vertical drain length. Then the required discharge capacity becomes:

where L is the prefabricated vertical drain length in meter, ∆σis additional effective load (kN/m²), and y_w is unit weight of water (kN/m³).

Equation 14.9 gives the required average discharge capacity. Since the rate of settlement varies during the consolidation process, the rate of flow will be faster in the earlier and slower in the later stage. Therefore, the discharge capacity required in the early stage may be greater than the average discharge capacity estimated from Equation 14.9. In the Changi East

reclamation project, the required average discharge capacity for each plot of PVD with various spacings were estimated by applying Equation 14.9 and were shown together with the manufacturer’s specifications of discharge capacity and the project specified discharge capacity (Table 14.5). It can be seen that the project specified discharge capacity is usually lower than the manufacturer’s discharge capacity and generally much higher than the estimated required average discharge capacity. However, it should be noted that the average discharge capacity obtained from Equation 14.9 is based on the average rate of flow throughout the preloading period. In reality, the rate of flow or the rate of settlement is much faster in the early stage of settlement and much slower in the later stage. Therefore, the actual required discharge capacity may be one or two orders higher than that suggested from Equation 14.9.

14.6.5.3— Factors affecting the measurement of discharge capacity

• Type of apparatus
• As explained in the earlier section, there are several types of discharge capacity testing equipment. Based on Darcy’s law, the rate of discharge is calculated as:
  
• As long as the hydraulic gradient is constant, the rate of discharge is constant if the flow media has a constant permeability and cross-section

regardless of the magnitude of the head (dh). However, as observed in some tests, the rate of discharge through vertical drains increases with head differences under the same hydraulic gradient. Therefore, the discharge capacity of the vertical drains measured at the same hydraulic gradient may increase with the length of the vertical drain tested. In other words, the discharge capacity measurement is affected by the dimension of the apparatus.

The variation of discharge capacity with different dimension apparatus is shown in Figure 14.13. Various configurations of the core for some types of vertical drains are shown in Table 14.6. It can be seen in Figure 14.13 that type B drain is most affected by the dimension of the testing apparatus. Type A drains are also tested under straight and buckled conditions in different laboratories using different types of apparatus such as ASTM 4716 and that of Chu and Choa (1995). It was found that the Chu and Choa’s 100 mm by 100 mm tester gave a lower discharge capacity than the ASTM 4716 apparatus in both straight and buckled conditions (Figure 14.14).

• Duration of test
• When PVD is compressed under pressure, the cross sectional area of the drain becomes smaller over time due to creep. The filter is also squeezed into the channel of the core. Furthermore, during the testing some fine materials enter into the drain and hence leads to clogging of the drainage channel. The variation of discharge capacity with the duration of the test measured for type A drain is shown in Figure 14.15.

• Reduction of discharge capacity with hydraulic gradient
• In Darcy’s law, the permeability of porous media is assumed to be constant. Therefore, the discharge capacity of porous media for certain dimensions of area is constant although the hydraulic gradient may vary.
• However, the permeability of the vertical drain core is not constant and varies with the hydraulic gradient. It can be seen in Figure 14.16 that the permeability of the drain core reduces with hydraulic gradient. Hence, the discharge capacity of the drains is reduced with hydraulic gradient. The flow through the vertical drain may not follow Darcy’s law which usually applies for flow through porous media. This matter was also discussed by Kamon et al. (1994) after obtaining the critical flow rate through vertical drain material using a transition number of 600. They commented that most vertical drain flow rates are greater than the critical flow and the type of flow is not laminar but found to be in transition.

• Types of surrounding material
• It was observed that different discharge capacity values were obtained when the PVDs were tested using different types of surrounding soil with the same loading and hydraulic gradient, as shown in Figures 14.17 and 14.18. It can be seen that the softer the soil, the lower is the discharge capacity. This can be due to the following factors: (1) the amount of deformation of the filters in the channels of the core is affected by the hardness of the

surrounding soil; and (2) the filter may be clogged when fine soils are used. Thus, the flow through the filter is reduced. The discharge capacity tests were also carried out with synthetic surrounding materials such as geomembrane and it was found that the greater the modulus of the material, the higher is the discharge capacity.

• Confining pressure
• The thickness of the prefabricated vertical drain is reduced under pressure, as shown in Figure 14.17. Therefore, the discharge capacity will reduce with increasing confining pressure. Kamon et al. (1994) has reported that when PVD was confined at a cell pressure of 320 kPa, it could reduce to 55 – 90% of that measured at a cell pressure of 5 kPa. However, this reduction of discharge capacity due to confining pressure varies widely, depending upon the type of drain.
• Several types of vertical drains were tested with straight 100 mm by 100 mm discharge capacity test equipment and it was found that all types of drains show reduced discharge capacity with confining pressure (Figure 14.19). It was also noted that the vertical drain with wire mesh core reduces the discharge capacity under confining pressure more significantly than the corrugated core.
• Discharge capacity of deformed drain
• Since PVD deforms with the consolidation of soil, the discharge capacity of the drains should be measured under buckled conditions, as often requested in land reclamation projects. However, the configuration of buckling is different from apparatus to apparatus and from test to test. Some types of discharge capacity tests are carried out with artificially deformed drains without soil, force kinking, folding and twisting whereas some are carried out on PVDs which have been compressed together with the soil.
• Miura et al. (1993) carried out a discharge capacity test on five different types of drains in a modified triaxial cell under five different configurations and reported that in the most extreme case of a sharp bend, the discharge

capacity decreased to only 26% of the discharge capacity under straight conditions. A reduction of discharge capacity with an increase in strain was observed. Kamon et al. (1994) reported a reduction in discharge capacity to 35–70% when the axial strain reaches 50%. Bergado (1994) also reported that the discharge capacity of drains with a sharp bend was reduced to 10–20% of straight condition values. Twisting of the drain also reduces the discharge capacity to 50% of straight conditions.

The discharge capacities of particular types of drain under various types of buckling configurations are shown in the Table 14.7. It can be seen that the twisted condition reduces the discharge capacity significantly. The folded condition does not reduce the discharge capacity as much as buckling with the soil column. Configurations of the core also affect the discharge capacity of the drain under buckled conditions. A corrugated core does not reduce the discharge capacity significantly under buckled conditions whereas a wire-mesh type of core reduces the discharge capacity significantly under buckled conditions.

14.6.5.4— Field measurement of discharge capacities

Field measurements of discharge capacities were carried out using settlement and pore pressure data at various pilot test areas. It was found that field mobilized discharge capacities ranged from 8.5 x 10^-7 to 1.3 x 10^-6 m/s. Field measurement of discharge capacities from various test areas are shown in Table 14.8.

14.7 VERIFICATION OF FIELD PERFORMANCE

Verification of the performance of a vertical drain system is usually determined by a pilot test. A pilot test generally consists of a few test plots with different spacings and a control area with no vertical drain. The general performance of the vertical drain can obviously be seen by comparing the area with a vertical drain and without a drain. By comparing the performance among various different spacings, suitable spacing can be determined. Figure 14.20 shows a pilot embankment test carried out at one of the Changi East reclamation projects. The results from a pilot test area carried out with two different types of drains and three different drain spacings are shown in Figure 14.21 where “M” denotes a Mebra manufactured in Korea, “C” denotes a Colbond drain, and “MH” denotes a Mebra drain manufactured in Holland. The degree of consolidation achieved in terms of settlement can be assessed by applying hyperbolic and Asaoka (1978) methods as shown in Figure 14.22. Figure 14.23 shows an assessment of the degree of consolidation from settlement monitoring data using the Asaoka and hyperbolic methods.

The average degree of consolidation ()is given by:

where the ultimate settlement (S_ult) is determined by applying either the hyperbolic or Asaoka method using settlement monitoring data. Settlement at time “t” (S_t) can also be obtained from the settlement monitoring data.

On the other hand, the degree of consolidation (U) can be assessed using pore pressure monitoring data. The degree of consolidation of soil element is given by:

where u_i is the initial excess pore pressure, and u_t is pore pressure at time “t“.

Figure 14.24 shows the degree of consolidation assessed from settlement gauges and pore pressure data.

Alternatively, in-situ testing methods can be used for assessing improvement of soil. With in-situ testing, both improvement in terms of the degree of consolidation, which is usually calculated from the over-consolidation ratio (OCR), and of undrained shear strength can be determined. Useful in-situ testing methods are:

• Cone penetration test (CPT).
• Dissipation test with CPTu.
• Field vane shear test (FVT).
• Dilatometer test (DMT).
• Dissipation test with dilatometer.
• Self-boring pressuremeter test (SBPT).
• Dissipation test with self-boring pressuremeter.
• Dissipation test with BAT permeameter.

Details of in-situ tests can be found in Bo et al. (1997a and 2002a) and also explained in Chapter 4. Figure 14.25 shows determination of OCR using in-situ methods, and Figure 14.26 shows determination of undrained shear strength using various in-situ methods.

Another way of assessing improvement is by laboratory testing. Undisturbed samples can be collected from the boreholes and tested at the laboratory. Water content, undrained shear strength determined from a triaxial compression test, and preconsolidation pressure determined from oedometer tests will indicate the improvement of soils.

Figure 14.27 shows a comparison of laboratory test results before and after improvement at the soil improvement area, and Figure 14.28 shows the degree of consolidation determined from laboratory test results.

Figure 14.23 Settlement monitoring data and ultimate settlement prediction by the hyperbolic and Asaoka method at a vertical drain location (after Bo et al. 1997a).