Chapter 3 Experimental Apparatus
3.5 Earth-Pressure Data Acquisition System
A earth-pressure data acquisition system was used to collect and store the considerable amount of earth pressure data generated during the tests. In the Fig. 3.15 (a) and (b) the data acquisition system is composed of the following four parts: (1) dynamic strain amplifiers (Kyowa: DPM601A and DPM711B); (2) AD/DA card (NI BNC-2090); and (3) Personal Computer. The analog signals from the sensors were filtered and amplified by the dynamic strain amplifiers. Then, the analog experimental data were digitized by an A/D-D/A card. The digital signals were then transmitted to the personal computer for storage and analysis. The software LabView was used for data collection and recording.
24
Chapter 4
Soil Characteristics
The characteristics of the fill used for soil improvement experiments are introduced in this chapter. The reduction of friction between the soil and lubricated side wall is discussed. The control and measurement of soil density distribution in the fill are also introduced.
4.1 Soil Properties
Air-dry Ottawa sand (ASTM C-778) was used throughout this investigation.
Physical properties of the soil include Gs
= 2.65, e
max = 0.76, emin = 0.50, D60 = 0.39 mm, and D10 = 0.26 mm. Grain-size distribution of the backfill is shown in Fig. 4.1.Major factors considered in choosing Ottawa sand as the fill material are summarized as follows.
1. Its round shape, which avoids the effect of angularity of soil grains.
2. Its uniform distribution of grain size (coefficient of uniformity Cu = 1.5), which avoids the effects due to soil gradation.
3. High rigidity of solid grains, which reduces possible disintegration of soil particles under loading.
4. Its high permeability, which allows fast drainage and therefore reduces water pressure behind the wall.
4.2 Lubricated Side-wall Friction
To simulate the field condition of a infinite half space for the compaction constitute, the shear stress between the fill and the side walls of the soil bin should be minimized to nearly frictionless. To reduce the friction between side wall and fill Fang et al.
(2004) suggested to was a lubrication layer fabricated with plastic sheets. Two types of plastic sheeting, one thick and two thin plastic sheets, were adopted to reduce the interface friction. All plastic sheets were hung vertically on the side walls before the backfill was deposited as shown in Fig. 4.2.
In this study, two thin (0.009 mm-thick) and one thick (0.152 mm-thick) plastic sheets were adopted for the soil improvement experiments. Fig. 4.3 shows the variation of side-wall friction angle
sw as a function of the normal stress
v for the plastic sheet method (1 thick + 2 thin sheeting) reported by Fang et al. (2004). The measured side-wall friction angle with this method is about 7.5°. For all experiments in this paper, the lubrication layers were wall applied on four side walls of the soil bin.4.3 Control of Soil Density
4.3.1 Air-Pluviated of Loose Ottawa Sand
To achieve a uniform soil density in the backfill, Ottawa sand was deposited by air-pluviation method into the soil bin. The air-pluviation method had been widely used for a long period of time to reconstitute laboratory sand specimens. Rad and Tumay (1987) reported that pluviation is the method that provides reasonably homogeneous specimens with desired relative density. Lo Presti et al. (1992) reported
26
that the pluviation method could be performed for greater specimens in less time.
Das (2010) suggested that, for granular soil deposits, the relative density Dr of 15~50%, is defined as loose, Dr = 50~70% is defined as medium, and Dr = 70~85% is defined as dense. For the air-pluviation method, Fig. 4.4 shows the soil hopper let the sand flow through a calibrated slot opening at the lower end. A picture of the soil pluviating processes is shown in Fig. 4.5. To achieve a loose backfill, Chen (2003) adopted the drop height of 1.0 m and hopper slot opening of 15 mm. In this study, the drop height of 1.0 m and the hopper slot-opening of 15 mm were also selected to achieve the loose fill. In Fig. 4.6, the expected relative density of soil was about 35%
by Ho (1999).
4.3.2 Measurement of Soil Density
To observe the distribution of soil density in the soil bin, soil density cups were made. The soil density cup made of acrylic is illustrated in Fig. 4.7. The circular cup wall was only 10 mm-high, so that the shear deformation and volume reduction could occur in the cup during testing. A picture of the soil density cup is shown in Fig. 4.8.
During the preparation of the 0.6 m thick loose soil specimen, density cups were buried in the soil mass at different elevations and different locations in the backfill as shown in Fig. 4.9 and Fig. 4.10. After the loose soil had been filled up to 0.6 m from the bottom of the soil bin by air-pluviation, density cups were dug out from the soil mass carefully. Fig. 4.11 shows the mass of the cup and soil in the cap was measured with an electrical scale.
For a 0.6 m thick air-pluviated Ottawa sand layer, the distribution of soil density with depth is shown in Fig. 4.12. For the loose sand, the mean unit weight
is 15.6
kN/m2, the mean relative density is Dr = 35.5 % with the standard deviation of 0.8%.Das (2010) suggested that for the granular soil deposit with a relative density 15% Dr
50% is defined as loose sand. The relative density achieved in Fig. 4.12 is quite
loose and uniform with depth.
28
Chapter 5
Testing Procedure
The procedure to conduct the cyclic torsional shear tests are introduced in this chapter. The testing procedure can be divided into three parts: (1) specimen preparation; (2) application of vertical static load; (3) application of cyclic torsional shear; and (4) cone penetration. These parts will be introduced in the following sections with pictures. The “plastic-sheets” lubrication layers were hung on the sidewalls of the soil bin before testing.
5.1 Specimen Preparation
Fig. 5.1 shows air-dry Ottawa sand in the soil storage container. Fig. 5.2 shows sand was shoveled from the soil storage to the sand hopper, and the mass of the fill was measured with an electrical scale. Fig. 5.3 shows the sand hopper was lifted by overhead crane in the laboratory. Fig. 5.4 shows Ottawa sand was deposited by air-pluviation method into the soil bin. The drop height was controlled to be 1.0 m and the hopper slot-opening of 15 mm were selected to achieve a loose fill, Fig. 5.5 (a) and (b) show portable hanging ladders were placed on top of the sidewalls, and a bridge board was placed between the ladders. Throughout the test, the operator will stay on the bridge board to avoid any unexpected surcharge on the soil specimen.
Leveling of the pluvuated soil surface by the graduate student with a brush is shown in Fig. 5.6. Four density cups were placed on each 50 mm-thick soil layer. A
total of 44 density cups were buried in the fill. Fig. 5.7 shows how check the density cup horizontal with a bubble level. Placement of a soil density cup and soil-pressure transducer on the soil surface is shown in Fig. 5.8. Fig. 5.9 shows density cups and soil-pressure transducers were buried in the soil mass at different elevations in the fill.
Eight soil pressure transducers were placed at the depths of 100, 250, 400 and 550 mm The soil pulviation and density cup placement operations were repeated unit a backfill thickness T = 0.6 m was reached.
5.2 Application of Vertical Static Load
The procedure to apply the vertical static load q on top of the air-pluviated loose sand is introduced. The cyclic torsional shear compactor (24.3 kg) and the loading discs (41.7 kg) used to apply static load has a mass of 66 kg. Diameter of the circular footing is 0.3 m and the vertical static load q = 9.2 kPa. Fig. 5.10 illustrates the grid points for the vertical load application. For the first row of static load, the center of circular load was applied at 1A, 1C, 1E, 1G and 1I.
Fig. 5.11 shows the CTSC was hoisted with overhead crane into the soil bin. Fig.
5.12 (a) shows the vertical static load q was applied on the loose sand with 5x5 formation. Fig. 5.13 shows the circular static vertical load was applied on the surface of the fill with the 5x5 loading formation.
5.3 Application of Cyclic Torsional Shearing
In this study, the cyclic torsional shear was applied on the soil surface for = ±1o,
±3o, ±5o, ±7o and ±10o. Fig. 5.14 showed a light dot from the laser distance meter on the angle steel bar was used as a fixed point to the soil surface. Fig. 5.15 shows the cyclic torsional shear was applied by the operator on the loose fill to increase its
30
density. In Fig. 5.16, 5.17 and 5.18, with the guidance of the fixed light dot, the circular disc shears the soil from 0∘to +5°and -5°. The application of cyclic torsional shear to loose sand is shown in Fig. 5.19.
For the test for with 20 shearing cycles (N=20),after the application of vertical static load, the torsional shear was first applied on the 4x4 loading formation (Fig.
5.12(b)) for the first 5 cycles, as shown in Fig. 5.20. To prevent disc penetration due to continuous shearing at the same spots, the shearing was moved to the 5x5 formation Fig. 5.12(a) from N = 6 to 10 as shown in Fig. 5.21. For N = 11 to 20, the shearing was applied on the 4x4 formation, as shown in Fig. 5.22. Fig. 5.23 shows, after compaction the soil density cup was carefully dug out of the soil mass. Fig. 5.24 (a) to (c) shows the density cup with a spatula. Fig. 5.25 (a) to (d) shows the brush away soil particles from base plate of the density cup. Soil mass in the cup was measured with an electrical scale and the density of the compacted soil could be determined.
5.4 Cone Penetration Test
In Fig.5.26, the points of penetration in the soil bin were labeled as C1, C2, C3 and C4. The steel beam on the top of soil bin to support the CPT facility is shown in Fig.
5.27 (a). Fig. 5.27 (b) showed the beam was fixed on the soil bin with steel c-clamps.
In Fig. 5.28, the electric motor and the movable plate was fixed to the steel beam by the screw. Fig. 5.29 shows the connection of cone penetrometer to the electric motor.
Fig.5.30 shows the cone was lowered to the surface of the soil mass. During penetrating, the speed of downward penetration of the mini-cone was controlled at 5 mm/s. After reaching the penetration depth of 400 mm, the testing was terminated.
then, moved the electric motor on the movable plate to the next point and repeat the
penetration procedure. The sequence of testing would be C1, C2, C3 and C4. Test results measured by the load cell on the mini-cone were collected, stored and processed with the CPT data acquisition system.
32
Chapter 6
Test Results
This chapter shows experimental results regarding soil densification due to static vertical load and cyclic torsional shearing. The static vertical load applied of the fill was q = 9.2 kPa. The cyclic torque and shearing applied on the soil surface was measured and reported. The surface settlement, distribution relative density, vertical and horizontal stresses and cone resistance of the compacted fill due to the static vertical loading and cyclic torsional shearing were investigated. The rotation angle
±1∘, ±3∘, ±5∘, ±7∘ and ±10∘, and the number of loading cycle were set to be N = 20. To obtain a soil mass with a relative density greater than 70%, compaction was applied on the fill surface for 0.15 m-thick lifts.
6.1 Static Load Test
To separate the densification effects due to static and cyclic loadings, in this section, the surface of four 0.15 m-thick soil lifts was compressed with the static vertical loading only. Effects of soil densification such as the surface settlement, change of relative density, vertical and horizontal stresses and cone resistance in the compressed fill are investigated.
6.1.1 Surface Settlement Due to Static Load
The surface settlements of the four 0.15 m-thick compressed soil lifts due the static
weight of the compactor were discussed. The initial relative density of the loose fill was 35.5% (see Fig. 4.12). The applied static normal stress was q = 9.2 kPa. To achieve a uniform settlement, the vertical static loading was applied on the surface with the 5×5 formation (see Fig. 5.12 (a)). Fig. 6.1 showed the settlement measurement was carried out with the laser distance meter placed on top of the steel beam. The surface settlement measured at the centers of disc loading disc was shown in Fig. 6.2. For the four 150 mm-thick soil lifts, the accumulated minimum and maximum settlements were 15.0 and 22.3 mm. The average settlement was 19.0 mm, which was about 3.2% (volumetric strain εv = 3.2%) of the soil thickness. It is obvious that static vertical loading is an effective method to compact the loose fill. To limit the scope of this thesis, only q = 9.2 kPa was used throughout this study. It should be mentioned that the vertical strain distribution in the soil lift may not be uniform.
6.1.2 Density Change Due to Static Load
To investigate the density distribution in the compressed fill, density cups were buried in the soil mass at different elevations and locations in the four 0.15 m-thick soil lifts (see Fig 5.9). For the un-compacted loose soil, the average relative density was about 35.5%. In Fig. 6.3, Chen (2011) reported that, for a 0.6m-thick fill, after the application of static vertical load q, the density increase more near the surface and the density increase less at greater depths. The induced density change decreased with increasing depth. Fig. 6.4 shows, after applying the static vertical load 9.2 kPa on each lift, the relative density of fill increased. This static vertical loading represents the weight of the cyclic torsional shear compactor. On the average, the relative density increase was about 26.4% from 35.5% to 61.9%. It should be mentioned that the distribution of density is not uniform with depth. However, the relative density
34
achieved Dr = 61.9% is not enough to achieve the target of dense sand (Dr = 70%-85%).
6.1.3 Stress Change Due to Static Load
For comparison purposes, at the beginning of this study, experiments were conducted to investigate the stresses in an uncompacted loose fill. Air-pluviation method was adopted to prepare the fill and the relative density achieved for the loose sand was 35.5 %. Fig. 6.5 shows the location of soil pressure transducers to measure the distribution of vertical and horizontal stress with depth. SPT2, SPT3, SPT6 and SPT7 were buried in the soil mass to measure
v and SPT1, SPT4, SPT5 and SPT8 were used to measure h. The vertical stress v measured in the soil mass was shown in Fig. 6.6. In this figure, the vertical stress v increased with increasing depth z. Test data were in fairly good agreement with the traditional equation v = z. In this study, unit weight was 15.6 kN/m3 for the loose sand. The distributions of horizontal earth pressure h with depth were shown in Fig. 6.7. In the figure, the earth pressure profile induced by the 0.6 m-thick loose fill was approximately linear and was in good agreement with the Jaky’s equation. Mayne and Kulhawy (1982), Mesri and Hayat (1993) reported that Jaky’s equation was suitable for backfill in its loosest state. From a practical point of view, it was concluded that for a loose fill, the vertical and horizontal earth pressure in the soil mass can be properly estimated with the equation
v = z and Jaky’s equation, respectively.To investigate the change of stresses due to static load, the loose fill was placed in four 0.15m-thick lifts as shown in Fig. 6.5. Static load q was applied each lift on the surface with the 5×5 formation (see Fig. 5.12) and then removed. Fig. 6.6 shows the vertical stress profile after the static vertical loading. It is clear in the figure that the vertical overburden pressure v can be properly estimated with the equation v = z.
As compared with the v for loose sand the measured,v increased slightly, probably because the compressed fill had a slightly higher density (see Fig. 6.4). It is clear in the figure that the static vertical load did not result insignificant residual stress in the vertical direction. It may be concluded that the effect of static loading on the vertical pressure v was insignificantly. Fig. 6.7 shows the horizontal stress was also increased slightly after the application of the static vertical loading.
6.1.4 Cone Resistance Change Due to Static Load
Cone penetration tests were conducted to investigate the change of soil properties due to static loading on the loose fill. The fill is 600 mm-thick as shown Fig. 6.5. The bottom of the soil bin is a solid steel plate. Due to boundary effects, the cone resistance may suddenly increase, when the penetrometer approached the bottom of the soil bin. For this study, cone penetration was conducted for z = 0 to 400 mm. For Fig. 6.8 shows the cone resistance for the loose soil varied from 0 to 300 kPa. After static vertical loading, the cone resistance varied from 0 to 1681.2 kPa. It is obvious that static vertical loading significantly increase the tip resistance of the compressed fill.
Jamiolkowski et al. (1985) were reported that could make cone resistance converted to relative density. Fig 6.9 shows the cone resistance of static load transformed to the relative density by Jamiolkowski theory. It could be found the test result and the theory were agreement in the loose fill. However, it was not agreement after compressed. It was probability that the soil mass not big enough.
6.2 Applied Cyclic Torsional Shearing
Fig. 6.10 shows the torque applied on the soil surface was measured with a digital torque meter. For the number of loading cycle N = 20, the torque measured at = ±1∘,
36
±3∘, ±5∘, ±7∘and ±10∘was shown in Fig. 6.11. In figure, for different
angles, the
applied torque varied between -53 to 47.8 N-m. Fig. 6.12 shows the applied torque T as a function of shearing angle. Test results indicated that the applied torque does
not change with increasing shearing angle. On the average, for = ±1∘, ±3∘, ±5∘, ±7∘ and ±10∘, the applied torque was about ±47.7 N-m. The variation of torque for the
angle between 0° and ±1° should be studied in the future.Fig. 6.13 indicates how to determine the maximum torsional shear stress
max at the edge of the shearing disc due to the applied torque. A linear distribution of shear stress from the center to the edge of the disc was assumed. Fig. 6.14 shows the6.3.1 Surface Settlement Due to Cyclic Torsional Shearing
After the application of the static loading, cyclic torsional shearing was applied on the surface of each lift. The cyclic shearing was applied on the circular areas of the 4×4 formation for N = 1 to 5, and 11 to 20. The shearing was applied on the 5×5 formation for N = 6 to 10 (see Fig. 5.12). Fig. 6.15 shows the surface settlements of the 0.6 m-thick fill after the disc rotation angle at ±1∘, ±3∘, ±5∘, ±7∘ and ±10∘. In the figure, the surface settlement increased with increasing shearing angle of torsional shearing. In the figure, the average settlement due to the static vertical loading q = 9.2
kPa was about 19 mm. After the cyclic disc rotation
= ±10° and N = 20 cycles of
torque application, the average settlement was 38.2 mm. The extra settlement due to the torsional shearing cycles was about 19.2 mm, which was more than the settlement due to static vertical loading. It was obvious that cyclic torsional shearing (static and cyclic load) is an effective method to compact loose soil. Fig. 6.16 shows the variation of surface settlement from 19 to 38.2 mm with the disc rotation angle from 0° to ±10°6.3.2 Density Change Due to Cyclic Torsional Shearing
Fig. 6.17 to Fig. 6.21 show the distributions of relative density due to cyclic torsional shearing angle of
= ±1
∘, ±3∘, ±5∘, ±7∘ and ±10∘. Fig. 6.19 shows the compaction could effectively increase the relative density up to 70% at all depths at the shearing angle = ±5°. However, the soil densification was effective only for the 150 mm-thick soil lift. To increase the effective compaction depth from 0.15 m to 0.30 m, it is possible that the radius of the compaction disc has to increase from 0.15 m to 0.30 m. .Fig. 6.22 shows the density distribution for = 0° to ±10°. In the figure, the relative density of compacted fill increased with increasing disc rotation angle