Luận văn Statnamic testing of piles in clay

viscoelastic and inertial effects are ignored. Plus, the soil was assumed to be incompressible and therefore based on the volume conservation of the soil under the penetration of an indenter, the velocity functions for soil particles are determined. Once the velocity functions are determined the soil deformations can be established by integration of the velocity functions along the streamlines. Likewise, other components of the soil deformation are determined, such as strain rate and stain paths. To calculate the stresses and pore water pressures soil models were proposed in which stresses can be calculated once the strains are determined. The main steps to determine stresses and pore water pressures are shown in Figure 2.27. In comparison, the two approaches give a similar pattern of field displacement except for the zone near the pile shaft (about 10% of the pile radius) and the zone near the pile tip (few diameters close to the pile tip) (Randolph, 2003). Chapter 2 Literature review 39 Several projects in which instrumented model piles with the capability of measuring shaft, tip resistances and pore water pressure were carried out to validate these theories. Coop and Wroth (1989) used an instrumented model pile, 80 mm in diameter and 1135 in length, with the capability of measuring pile shaft resistance and pore water pressures at the pile tip and shaft. Two test-bed sites, one with heavily overconsolidated clay and one with normally consolidated clay, were chosen for the experiments. The data showed that pore water pressures were far lower than the cavity expansion prediction and fluctuated much more widely than the theory predicted. Bond and Jardine (1991) carried out experiments using instrumented model piles in a clay where overconsolidation ratio varied from 50 to 20 with depth. Large negative pore pressures had occurred during pile installation and the cavity theory could not be applied. In comparison with the strain path method, they concluded that the method achieved a similar shape of shear strain distribution. However, the strain path method underpredicted the measured strains. It was suggested that the possible reason for this was the discrepancy of the roughness between the pile and the soil around it. In the strain path method there is an assumption of a perfectly smooth boundary between the pile and the soil. No comparison between the measured and strain path method prediction of pore water pressures was reported. Several attempts have been made to measure the pore water pressure changes during statnamic load tests. However, no theoretical analyses have been used to try to quantify these changes. Hajduk et al. (2000) published the results of a series of statnamic load tests which were carried out for two heavily instrumented piles. However, only the results of pore water pressures for one pile were reported. The pore water pressures along the piles were measured by four pore water transducers mounted at different levels (5 m, 10, 15, 20 m from the pile top). The first one was positioned in overconsolidated clay with an initial pore pressure value of 32 kPa, the second in soft normally consolidated clay with an initial pore pressure value of 80 kPa, the third in normally consolidated clay with an initial pore pressure value of 135 kPa, and the final in silty sand with an initial pore pressure value of 192 kPa. Six statnamic tests were carried out on this pile over 2 hours. All pore water transducers indicated the same trend. Pore pressures dropped during each test (2-10 kPa for the first transducer, 3-8 kPa for the second, 3-9 for the third, and 4-16 kPa for the final) Chapter 2 Literature review 40 and then recovered when the tests finished. However, due to six tests being carried in a two hour period the pore water pressures after one test did not recover to the prior test value before the next test. Maeda et al. (2000) published the results of a statnamic test for a cast-in-place concrete pile diameter of 1.2 m and length of 13.4 m with measurement of pore water pressure at the pile tip which was in gravelly sand. During the test excess pore water pressure increased simultaneously with applied statnamic load by up to 80 kPa then dissipated immediately after the test. Conversely, negative pore water pressures at the pile tip developed during statnamic load tests for steel pipe piles 216.3mm in diameter, 4.5 m in length which were installed in sand (Ishida et al. 2000). 2.9 Summary From the literature review, the main points related to this study are as follow: ♦ Theoretical approaches for predicting the pile capacity have advanced significantly. However, pile load testing methods still play an important role in pile design and for research purposes. ♦ The development of the statnamic pile testing method shows the innovative and promising potential of the method. However, further attention needs to be given to the rate effects and the pore pressure regime around the pile during testing and after a statnamic test. ♦ A large number of studies have examined rate effects. In general, an exponential increase of shear capacity with respect to shearing velocity has been proposed. However, different studies have the relationship in different forms. Attention has usually been given to rate effect on the ultimate shear strength whereas rate effects on the shearing resistance prior to failure have received less attention. ♦ The exponent value in the rate effect equation for clays, i.e. damping parameter β, can be taken as 0.2 as this value has been proposed by several researchers who have investigated a wide range of clays (reconstituted clay with w = 17%-20%, wP = 17%, wL = 37% and undisturbed glacial clay at a site near Grimsby, UK with w = 13% - 25%, wP = 12% - 18%, wL = 20 – 36% by Balderas-Meca, 2004; London clay wP = 27, wL = 70, Magnus clay with wP = 17%, wL = 31%, and Forties clay from the North sea with wP = 20, wL = 38% by Litkouhi and Poskitt, 1980; overconsolidated clay Chapter 2 Literature review 41 from Heather and Claymore in the NorthSea and Kontich clay, Belgium by Heerema, 1979). ♦ Quake values for pile shaft and base have been studied by several researchers. Nevertheless, no clear mechanism of quake values for dynamic or statnamic tests has been reported. ♦ Several models have been proposed to predict the pore water pressures around the pile during pile installation. Among them, cavity expansion theory and the strain path method are the most widely used. The cavity theory is easy to apply due to its simplicity. However, it seems not to provide a good match with the measured data (Coop & Wroth, 1989; Bond & Jardine 1991). In particular, it fails to predict the changing of pore water pressure for heavily overconsolidated clays when negative pore water pressures occur due to shearing. On the other hand, the strain path method seems to provide a better prediction. However, a sound model for soil with many parameters are required for the method and due to the difficulty in specifying this model, the method is less popular. Chapter 2 Literature review 42 Soil type in bearing strata Suggested range of Jc Correlation value of Jc Sand Silty sand/sandy silt Silt Silty clay/clayey silt Clay 0.05 – 0.20 0.15 - 0.30 0.20 – 0.45 0.40 – 0.70 0.60 – 1.10 0.05 0.15 0.30 0.55 1.10 Table 2.2 Case damping coefficient for different soil types (Fleming et al. 1992) Table 2.1 Damping parameter in Dayal and Allen study (Dayal and Allen, 1975) Shear strength Low velocity High velocity (kPa) less than152 (mm/s) above 305 (mm/s) 3 0.38 0.93 45.09 0.31 0.75 49.97 0.24 0.66 78.33 0.17 0.38 Damping coefficient (k1) Chapter 2 Literature Review 43 Figure 2.1 O-Cell Figure 2.2 Schematic arrangement of a Osterberg test (Osterberg & Pepper, 1984) Transducer feedback amplifier Load cell Actuator LVDT Electro- pneumatic servo valve Comparator Error amplifier Command signal CDAS Sample Error signal Pressure supply Current Figure 2.3 Balderas-Meca’s test apparatus arrangement. (Balderas-Meca, 2004) Chapter 2 Literature Review 44 0 0.2 0.4 0.6 0.8 1 1.2 0 1 2 3 4 5 6 7 8 Axial strain % D am pi ng c oe ffi ci en t α Figure 2.4 Damping coefficient, α, versus axial strain for monotonic consolidated undrained triaxial tests at different rates. (β=0.20; OCR=1) (Balderas-Meca, 2004) Figure 2.5 Half steel tube with semi-circular soil sample (Heerema, 1979) Chapter 2 Literature Review 45 Figure 2.6 The shear device for the study of pile-soil interfaces (Chin, 2004) Shear box Pile on carriage Actuator Data acquisition console PC Pump Oil reserve Servo-valve A/D Load cell LVDT LVDT LVDT High speed actuator Controller (High-speed actuator) Accelerometer Dynamic load cell Controller (Low-speed actuator) Chapter 2 Literature Review 46 Figure 2.7 Penetrometer and soil container in experimental set-up (Dayal and Allen, 1975) Figure 2.8 (a) Schematic of the test arrangement (b) geometry of penetrometer for side friction tests (Poskitt and Leonard, 1980) Chapter 2 Literature Review 47 Figure 2.9 Undrained peak strength measured from vane tests (Biscontin & Pestana, 2001) Chapter 2 Literature Review 48 Figure 2.11 Shaft resistances and pile movements (King et al. 2000) Figure 2.10 Slow and quick-penetration tests (Jaime et al. 1990) Chapter 2 Literature Review 49 Figure 2.12. Wave propagation in a bar produced by an impact load Figure 2.13. Idealization of a pile as an elastic rod with soil interaction at discrete nodes dσ -A(σ + dz ) dz z A A’ B B’ dz u dz -Aσ Soil surface Ti-1 Ti Ti+1 Nodes for soil resistance Δt Δz = c Impedance Z Base Node n Chapter 2 Literature Review 50 Figure 2.14. Model of downward and upward waves due to soil interaction (Fleming et al. 1992) Dashpot Pile node Spring Soil at distance from pile Figure 2.15. Smith Model for pile and soil Time t Time t+Δt Nodes Original velocities Updated velocities Pile Pile (vu)i+1 (vu)i (vd)i (vd)i-1 i -1 i i +1 Soil resistances Ti-1 Ti Ti-1 (vu)i-1 = (vu)i+ 2Z Ti (vu)i = (vu)i+1+ 2Z Ti-1 (vd)i = (vd)i-1 - 2Z Ti (vd)i+1 = (vd)i - 2Z Chapter 2 Literature Review 51 Figure 2.17 Randolph & Deeks model for pile tip and soil Soil adjacent to pile Pile node Viscous Dashpot Spring Plastic slider Inertial dashpot Soil at distance From the pile Lumped mass Pile node Viscous Dashpot Spring Plastic slider Inertial dashpot Soil at distance From the pile Figure 2.16. Randolph & Deeks model for pile shaft and soil Figure 2.18. A typical statnamic loading-time relationship Chapter 2 Literature Review 52 A, Pile B, Load Cell C, Cylinder D, Piston E, Platform F, Silencer G, Reaction Mass H, Gravel Container I, Gravel J, Laser K, Laser Beam L, Laser Sensor Figure 2.19. Statnamic device Figure 2.20 Forces acting on a pile during statnamic loading Fmax Static Force Displacement [mm] Rate effect Statnamic Unloading Point Figure 2.21 Unloading point method FSTS Fa Fd Fs FSTS : Statnamic force Fa : Inertia force Fd : Damping force (Fd = c.v, v- Velocity and c – Damping factor) Fs : Static force Fsoil = Fd + F s Chapter 2 Literature Review 53 Load Settlement Total Shaft Base Figure 2.22 Load-Settlement response (Tomlinson 2001) Figure 2.24 Ramberg-Osgood model for the relationship of shaft resistance and displacement (Armaleh & Desai, 1987) Figure 2.23 Shaft quake values compared with the pile diameter (Kraft et al. 1981a) Displacement, w S ha ft re si st an ce k os p m = 1s f k = 0.005k f os Chapter 2 Literature Review 54 Pile movement, w Sh ea r s tr e ss A B q s maxτ τresidual w r Figure 2.25 Idealised softening behaviour for a pile in clay (Kraft et al. 1981b) Figure 2.26 Chandler & Martins’ test apparatus (Chandler & Martins, 1982) Chapter 2 Literature Review 55 1-Initial Stresses, σoij, uo 2-Soil velocities Vi 4- Strain rates ε’ij 3- Deformation 5- Strain path εij along stream line 6.A Effective stress approach 6.B Total stress approach Model: Effective Stress vs. strain Model: Deviatoric Stress vs. strains Model: Shear induced pore pressure vs. strains Deviatoric Stresses sij Shear induced Pore pressures us Effective Stresses σij Equilibrium 7- Pore pressures u 8- Total stresses, σij Figure 2.27 Strain Path Method to deep penetration in clays (Baligh, 1988) Chapter 3 Testing procedures and Equipment 56 CHAPTER 3 TESTING EQUIPMENT AND PROCEDURES 3.1 Introduction To get a better understanding of clay response under dynamic loads, an instrumented pile with the capability of measuring the pile tip and shaft resistances separately was used. Clay beds with final dimensions of height approximately 1100mm and diameter 780mm were prepared by a two stage consolidation process in a calibration chamber for the pile installation and subsequent testing. The use of a servo hydraulic system enabled two types of test, which were the constant rate of penetration tests with the rates of from 0.01mm/s to 500mm/s, and statnamic tests with the maximum peak loads of up to 41 kN to be carried out. Maintained load tests were also carried out using another actuator with the ability to maintain a constant load for a long period of time. The servo hydraulic system was not suitable for this. The chamber for clay bed preparation was designed and manufactured at the University of Sheffield (Anderson et al., 1991). In this study some modifications were made to it which will be described later. The instrumented pile was designed by Brown (Brown, 2004). In this study, the pile tip and shaft pore water pressure transducer arrangements were modified to get better responses. Brown (2004) carried out two types of tests using the instrumented pile: constant rate of penetration tests and statnamic tests. Constant rate of penetration tests were carried out at different rates (0.01 mm/s, 1 mm/s, 10 mm/s, 50 mm/s, 100 mm/s, 200 mm/s, 350 mm/s, 500 mm/s) to give an insight into the rate effect of clays (Hyde et al. 2000). Statnamic load tests were carried out at different peak loads (10 kN, 15 kN, 20 kN, and 30 kN) for studying clay responses under the statnamic loading pulses. In this study, more constant rate of penetration and statnamic tests have been carried out. In addition, maintained load tests have been carried out as these eliminate totally the rate effects present in constant rate of penetration tests and thus give a relatively long- Chapter 3 Testing procedures and Equipment 57 term, drained pile capacity. The results of these tests were incorporated with those of the constant low rate (0.01 mm/s) of penetration tests to define the pile static capacity. 3.2 The calibration chamber Model testing is a popular method to study geotechnical problems. Model tests using a calibration chamber have advantages over field tests in that soil stress history is known and the tests are carried out under strictly controlled conditions. Also, laboratory tests are relatively cheap compared with field tests, and the field test-to-test differences are minimized in the calibration chamber tests. However, calibration chambers have disadvantages with boundary and scale effects present. To get useable results from the calibration chamber tests it is necessary to either eradicate the boundary and scale effects or control them. The second choice is usually adopted because of the space limitations in a laboratory and costs. There are two types of calibration chambers: centrifugal chambers and 1-g chambers. A centrifugal chamber test tries to replicate a field test at a predetermined scale for as many parameters of the test as possible. However, with pile tests centrifuge scaling laws are difficult to apply for pile shaft resistance because many factors influence shaft friction such as the pile shaft roughness and the intrinsic soil properties (for instance soil cohesion). They cannot be applied by using the conventional scaling laws. A 1-G calibration chamber only studies pile tests at a certain ambient stress value by applying a predetermined pressure to the top and side chamber. Although the field stress variation is not replicated in the 1-G calibration chambers they still provide an excellent means of studying the fundamentals behind soil-probe interaction, and provide reasonable data points to compare with those obtained from the field. The uses of a 1-g calibration chamber for modeling field tests in sands are quite numerous (Parkin et al. 1980; Bonita, 2000; White, 2002). The 1-G calibration chambers for sands are characterized by a short time for bed preparation, small volume changes, and simple preparatory bed procedure. Conversely, 1-D calibration chambers for clays need to be designed to accommodate the relatively large volumetric change during consolidation and to allow the change over from the 1-G Chapter 3 Testing procedures and Equipment 58 consolidation to the triaxial consolidation. For this reason, small 1-G calibration chambers for clays (Anderson et al., 1985; Houlsby & Hichman, 1988; Huang et al., 1988) are more popular than the big ones (Anderson et al., 1985; Smith, 1993). 3.3 Boundary effects Calibration chamber test results can only be used with confidence to analyse the relevant field events if the boundary effects are either eliminated or controlled. Normally it is not practical to produce samples of sufficient size to satisfy the first condition. For this reason, the second option is usually adopted. Boundary effects need to be studied and the chamber was designed in such a way that its boundary effects are negligible. The boundary effects are mainly influenced by the probe dimension to the specimen dimension ratio, soil properties, test types (both static tests and dynamic tests were carried out for this study), and the boundary rigidity. In this study the soil bed diameter is 780 mm, giving a soil bed to probe diameter ratio of 11 using a instrumented pile with the diameter of 70 mm. The soil beds were consolidated under pressures of 280 to 400 kPa and there was no reduction of consolidation pressure during the tests so the soil was normally consolidated if extra transient loads exerted on the soils surrounding the pile shaft and the pile tip were ignored. In practice, these transient loads caused local overconsolidation for these thin zones. The use of a flexible side boundary for this chamber was an advantage in terms of the boundary effects. It gave a better uniformity for clay bed during the triaxial consolidation. However, the use of a rigid bottom plate was disadvantage during pile loading tests when the pile advanced near to the bottom plate. For this reason, the testing programme for each bed finished when the pile tip was about 140mm (four pile radii) above the bottom plate. This was considered a reasonable distance to avoid an unacceptable pile tip boundary effect as discussed at the end of this section. Additionally, the friction between the clay sample and the bottom rigid plate caused a gradual reduction of the horizontal pressure during the triaxial consolidation from the bed periphery to the centre. Chapter 3 Testing procedures and Equipment 59 A few attempts have been made for assessing the boundary effects of a calibration chamber (Smith, 1993). Based on undrained expansion theory, Smith (1993) calculated the plastic zone occurring around the probe during its installation using the following equation: r p I R R = (3.1) where Rp = radius of plastic zone around the testing probe R = effective radius of the probe Ir = rigidity index of the soils = G/cu G = soil shear modulus cu = undrained shear strength Smith (1993) supposed that if the outer bed boundary is beyond the above plastic zone the boundary effects will be minimized. A difficulty occurs in the choice of exact value for G/cu. Typical ratios of the initial shear modulus to limiting shaft friction are in the range of 400 to 1000 (Randolph & Deeks, 1992). However, the shear modulus is not constant and drops dramatically when the shear strain increases. It is, therefore, recommended that the shear modulus should be taken from typical stress-strain curves over the stress range zero up to half the ultimate shear stress. Thus, in this study it could be conservative if the range of G/cu is taken from 70 to 120 which was predicted by Randolph, et al (1979) for London clay with OCR from 1 to 32. Applying these values in Equation 3.1, the plastic zone varied from Rp = 293 mm to 384 mm. Comparing with the clay bed radius, 390 mm, the plastic zone was still safely inside the bed boundary. Baligh (1985, 1986a) pioneered in the use of the strain path method to predict the stresses and strains around a pile due to the pile installation. Figure 3.1 shows the deviatoric strain, E, and its rate, E’, around a pile with a radius of 17.8 mm penetrating at a velocity of 20 mm/s, which were obtained using this method (Baligh, 1986a). The deviatoric strain, E, and its rate, E’, were defined as follows: 23 2 2 2 12 1 EEEE ++= (3.2) Chapter 3 Testing procedures and Equipment 60 2 3 2 2 2 1 '''2 1' EEEE ++= (3.3) where zzE ε=1 (3.4) ( )θθεε −= rrE 3 1 2 (3.5) rzE ε3 2 3 = (3.6) Assuming that the soil clay is obeying Von Mises criterion for yielding, the plastic zone was bounded by the contour line for deviatoric strain Ey = 0.41% for clay with G/cu = 100. The plastic zone around the pile only expands to about 10R, which for this research is still inside the bed boundary. The plastic zone at the pile tip expands to about 6R (Figure 3.1) whereas the testing programmes in this study were completed when the pile tip advanced to about 4R above the bottom chamber plate. To reduce the boundary effects only low rate pile load tests were carried out for each bed at the end of the testing programmes. 3.4 Bed preparation 3.4.1 Clay slurry preparation A clay slurry was formed by thorough mixing of dry Speswhite kaolin powder with sand, silt, and deaired deionised water in a concrete pan mixer. The component proportions of materials (by weight) were 50% kaolin, 25% sand, and 25% silt. The Speswhite kaolin powder was supplied by Whitchem, Staffordshire and its properties are shown in Table 3.1. The sand was Buckland P30 Silica Sand and supplied by Hanson Aggregates, Kent and its properties are shown in Table 3.2. For this study, the first two beds used Oakamoor HPF4 Silica Flour silt supplied by Hepworth Minerals and Chemical, Cheshire but the company discontinued supplying this silt so the final beds used HPF3 and HPF5 silts mixed 50-50% to make HPF4 instead. Its properties are shown in Table 3.3. Chapter 3 Testing procedures and Equipment 61 De-aired water was used to produce slurry with a moisture content of about 55%, which was equivalent to 1.5 times its liquid limit. This was recommended by Sheeran & Krizek (1971) for the preparation of uniform clay beds using slurry consolidation. Prior to mixing, the water was deionised and then de-aired for at least 2 hours. Eight batches were required for producing one bed and the amount of each component for one batch is shown in Table 3.4. To make the slurry for one batch, firstly 50 litres of de-aired, deionised water followed by 45.5 kg kaolin and 22.75 kg silt were added to the pan. The mixture was mixed for about 5 minutes to form a slurry and then the sand was added gradually for about 5 minutes. After that the slurry was mixed for at least 30 minutes. For each batch, at least four samples were taken from different locations to determine the moisture content for assessment of the batch uniformity. The consistent moisture contents of 55% ± 1% found for the five beds demonstrated that the method of mixing the slurry was reliable. To be able to compare these results with previous study results the bed production procedure was kept similar to that used by Brown (2004). The material properties are shown in Table 3.5. Once the slurry was formed it then was transferred to a consolidometer (Figure 3.2) using a M