The characteristics of the backfill and the method to reduce the wall friction are introduced in this chapter.
4.1 Backfill Properties
Air-dry Ottawa silica sand (ASTM C-778) was used as the backfill material in all experiments. Physical properties of the soil are summarized in Table 4.1.
Grain-size distribution of the backfill is shown in Fig. 4.1. The major reasons to select Ottawa sand as the backfill material are listed below.
1. Its round shape, which avoids effect of angularity of soil grains.
2. Its uniform distribution of grain size (coefficient of uniformity Cu = 1.78), 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.
To establish the relationship between unit weight of backfill γ and its internal friction angle φ, direct shear tests have been conducted. The shear box used has a square (60 mm ×60 mm) cross-section, and its arrangement is shown in Fig. 4.2. Before shearing, Ottawa sand was air-pluviated into the shear box and then compacted to the desired density. Details of the technique to control soil density are discussed in section 5.1.
Chang (2000) established the relationship between the internal friction angle φ and unit weight γ of Ottawa sand as shown in Fig. 4.3. It is obvious from the figure that soil strength increases with increasing soil density. For the air-pluviated backfill, the empirical relationship between soil unit weight γ and φ angle can be formulated as follows
4.2 Reduction of Wall Friction
To constitute a plane strain condition for model wall tests, the shear stress between the backfill and wall should be minimized to nearly frictionless. To reduce the friction between wall and backfill, a lubrication layer fabricated with plastic sheets was furnished for all experiments. 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 four walls before the backfill was deposited as shown in Fig. 4.4.
Multiple layers of thin plastic sheets (without any lubricant) were used by McElroy (1997) for shaking table tests of geosynthetic reinforced soil (GRS) slopes. Burgess (1999) used three thin plastic sheets to reduce side wall friction in full-scale GRS wall tests. The wall friction angle was approximately 15° as determined by the shear box tests. In this study, two thin (0.009 mm-thick) and one thick (0.152 mm-thick) plastic sheet were adopted for the earth pressure experiments. The friction angle δw developed between the plastic sheets and steel sidewall could be determined by the sliding block test. A schematic diagram and a photograph of the sliding block test proposed by Fang
et al. (2004) are illustrated in Fig. 4.5 and Fig. 4.6. The wall friction angle δw for the sliding block test was determined using the basic principles of physics. Fig. 4.7 shows the variation of friction angle δw as a function of the normal stress σn for the plastic sheet method (1 thick + 2 thin sheeting) used in this study. The measured friction angle with this method is about 7.5°. It is clear in Fig.4.7 that the interface friction angle δw is nearly independent of the applied normal stress σn. This constancy is an important advantage in establishing the input soil properties for analytical models that might be used to analyze the experimental results. For all experiments in this paper, the lubrication layer wall applied on four walls as indicated in Fig. 4.4. The plastic sheets not only can help to reduce the friction angle between the wall and the backfill. The plastic sheets can also help to reduce the reflection of elastic waves transmitted to the soil-wall boundaries during compaction.
Chapter 5
Test Results for Loose Sand
This chapter introduces the distribution of soil density, horizontal and vertical stresses in the loose sand backfill before compaction. The sections discussed included: (1) the method to prepare the loose backfill; (2) the method to control soil density; (3) the measured distribution of soil density in loose sand; and (4) the distribution of earth pressure in loose sand.
5.1 Testing Procedure
The testing procedures adopted for this study for the measurement of relative density and stresses in the loose backfill are briefly described below.
For the measurement of the relative density:
(1) Sand pluviated into the soil bin by controlling the drop height of soil and slot opening of the sand hopper.
(2) Density cups placed at the different elevations and locations.
(3) After the soil had been filled up to 1.5 m from the bottom of the soil bin, soil density cups were dug out from the soil mass carefully.
(4) The weight of the cup and soil was measured and recorded.
For the measurement of the earth pressures:
(1) Sand pluviated into the soil bin by controlling the drop height of soil and slot opening of the sand hopper.
(2) Placed the SPTs at the desired locations.
(3) When the backfill was filled up to 1.5 m, the earth pressures were recorded and stored.
5.2 Distribution of Soil Density
5.2.1 Air-Pluviation of 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 that the pluviation method could be performed for greater specimens in less time.
Das (1994) suggested that relative densities of 15~50%, and 70~85% are defined as loose and dense condition, respectively. To achieve loose backfill (Dr = 32%), Chen (2002) adopted the drop height of 1.0 m and hopper slot opening of 15 mm. According to the test results, Ho (1999) established the relationship among slot opening, drop height, and density as shown in Fig. 5.1. As a result, the drop height of 1.0 m and hopper slot-opening of 15 mm are selected to achieve the loose backfill for testing in this study. Fig. 5.2 shows the method to control the drop height = 1.0 m. In the picture, a 1.0 m-long rope was hung from the hopper to the surface of the soil to control the drop distance of soil. In Fig. 5.3, the soil hopper that lets the sand pass through a calibrated slot opening (15 mm) at the lower end was used for the spreading of sand.
The raining of the Ottawa sand into soil bin is shown in Fig. 5.4.
5.2.2 Uniformity of Soil Density
To observe the distribution of soil density in the soil bin, the soil density cups were made. The soil density control cup made of acrylic is illustrated in Fig. 5.5 and Fig. 5.6.
During the preparation of soil specimen, density cups were buried in the soil mass at different elevations and different locations in the backfill as shown in Fig. 5.7 and Fig.
5.8. After the soil had been filled up to 1.5 m from the bottom of the soil bin, soil density cups were dug out from the soil mass carefully. Fig. 5.9(a) shows the density cup was placed in the soil bin and Fig. 5.9(b) shows the weight of the cup and soil was measured with an electrical scale. The distribution of soil density with depth for loose sand is shown in Fig. 5.10. The mean relative density is Dr = 34.1 % with the standard deviation of 2.4%. The soil density distribution was reported by Chen (2002). The test results are in fairly good agreement with data. The backfill achieved with the air-pluviation method was loose, Dr = 15%~50% as suggested by Das (1994).
5.3 Distribution of Earth Pressure
For comparison purposes, at the beginning of this study, experiments were conducted to investigate the stresses in an uncompacted backfill. Fig. 5.11 shows the location of soil pressure transducers to measure the distribution vertical earth pressure σv with depth. The method to confirm the location and depth of the SPT in the soil mass is shown in Fig. 5.12. Fig. 5.13 shows the photograph of SPT used to measure vertical stress in the soil mass. After the backfill had been filled up to 1.5 m thick, the vertical earth pressure σv measured in the soil mass was illustrated in Fig. 5.14.
Obviously, the vertical pressure increased with increasing depth and the test data were in good agreement with the equation σv = γz, where γ is the unit weight of the backfill.
The locations of soil pressure transducers to measure the distribution of horizontal
earth pressure σh were shown in Fig. 5.15. Fig. 5.16 shows the photograph of SPT used to measure horizontal stress in the soil mass. The distribution of horizontal earth pressure σh with depth was illustrated in Fig. 5.17. In the figure, the earth pressure profile induced by the 1500 mm-thick loose backfill was approximately linear and was in good agreement with the Jaky’s equation. Mayne and Kulhawy (1982), Mesri and Hayat (1993) reported the Jaky’s equation is suitable for backfill in its loosest state.
From a practical point of view, it may be concluded that for a loose backfill, the vertical and horizontal earth pressure in the soil mass can be properly estimated with the equation σv = and Jaky’s equation, respectively. γz