Chapter 1 Introduction
1.3 Organization of Thesis
This paper thesis divided into the following parts:
1. Review of past investigations regarding cyclic torsional shear compaction of cohesionless soils. (Chapter 2)
2. Description of the National Chiao Tung University non-yielding soil bin, cyclic torsional shear compactor and cone penetration device. (Chapter 3) 3. Backfill characteristics and soil density control technique. (Chapter 4) 4. Description of testing procedure. (Chapter 5)
5. Experimental results of the surface settlement due to compaction, relative density distribution in soil mass, stresses induced by compaction and cone resistance in the soil due to cyclic torsional shear compaction. (Chapter 6) 6. Conclusions. (Chapter 7)
4
Chapter 2
Literature Review
Das (2010) stated that the soil at a construction site may not always be totally suitable for supporting structures such as buildings, bridges, highways, and dams. For example, in granular soil deposits, the in situ soil may be very loose and perform a large elastic settlement under loading. In such a case, the soil needs to be densified to increase its unit weight and thus its shear strength.
Sometimes the top layers of soil are undesirable and must be removed and replaced with better soils on which the structural foundation can be built. The soil used as fill should be well compacted to sustain the desired structural load. Compacted fills may also be required in low-lying areas to raise the ground elevation for construction of the foundation.
To improve its engineering properties, contractors are generally required to compact the loose soils to increase their unit weights and reducing settlements.
Previous studies associated with the compaction-induced effects such as the change of soil density, the change of stresses in the soil mass and mechanism of soils under compaction are discussed in this chapter.
2.1 Soil Improvement with Densification
Kramer (1996) defined the common soil improvement techniques to mitigate seismic hazards, soil improvement method were divided into four categories including (1) densification techniques (vibrofloatation, vibro rod, dynamic compaction, blasting,
and compaction grouting); (2) reinforcement techniques (stone columns, compaction piles, and drilled inclusions); (3) grouting and mixing techniques (permeation grouting, intrusion grouting, soil mixing, and jet grouting), and (4) drainage techniques. In this thesis, only the densification of cohesionless soil were discussed.
2.1.1 Densification Techniques
Fig. 2.1 shows two of the many possible ways that a system of equal-sized spheres can be packed. Looser systems than the simple cubic packing can be obtained by carefully constructing arches within the packing, but the simple cubic packing in Fig.2.1 (a) is the loosest of the stable arrangements. The dense packings in Fig.2.1 (b) represent the densest possible state for such a system. A dense packing of soil spheres can be reached by soil densification techniques.
2.1.2 Soil Densification with Vibratory Compactor
D’Appolonia et al. (1969) proposed the vibratory rollers are particularly useful for compacting granular soils. Fig. 2.2 shows the effects of compaction of a 8-ft lift dune sand after five passes by a vibratory roller. The low unit weight that remains in the uppermost zone is due to vibration and lack of confinement in sand. Fig. 2.3 shows the compacted unit-weight profiles for the same dune sand after 2, 5, 15, and 45 roller passes. For field compaction work, the specification requires that the granular soil be compacted to a certain minimum relative density at all depths. Determination of the height of each lift depends on the type of roller and the economic number of passes.
The method for determination of the lift height is shown in Fig. 2.4. For soils at all depths to reach a minimum relative density Dr = 75%, the lift thickness should be controlled to be less than 18 inch.
6
2.2 Earth Pressure At-Rest
2.2.1 Coefficient of Earth Pressure At-Rest
In Fig. 2.5(a), a soil element A located at depth z is compressed by the overburden pressure
=z. During the formation of the deposit, the element A is consolidated
under the pressure
v. The vertical stress induces a lateral deformation against surrounding soils due to the Poisson’s ratio effect. Over the geological period, the horizontal strain is kept to be zero and the surrounding soil would develop a lateral stress to counteract the lateral deformation. A stable stress state will develop that the principal stresses acts1 and 3 on the vertical and horizontal planes, as shown in Fig.2.5(b).
The soil in a state of static equilibrium condition is commonly termed as the Ko
condition. Donath (1891) defined the ratio of the horizontal stress h to vertical stress
v as the coefficient of earth pressure at-rest, Ko, orwhere E is the elastic modulus and is the Poisson’s ratio of the soil.
Base on the definition of the at-rest condition, the lateral strain would be zero (εh= materials only. However, the behavior of soil element is more complex and far from these assumptions. It is evident that the relationship between Ko and elastic parameter,
Eq. 2.5 is not practical for predicting in-situ horizontal stress.
2.2.2 Jaky’s Formula
Several scholars attempted to set up a theoretical relationship between the strength properties of a soil and Ko. The empirical relationship to estimate Ko of coarse-grained soil is discussed in the following section.
Mesri and Hayat (1993) reported that Jaky (1944) established a relationship between Ko and maximum effective angle of internal friction by analyzing a talus of granular soil freestanding at the angle of repose. Jaky (1944) supposed that the angle of repose is analogous to the angle of internal friction
. This is reasonable for a
sedimentary, normally consolidated material. Jaky (1944) reasoned that the sand cone OAD in Fig. 2.7 is in a state of equilibrium and its surface and inner points are motionless. The horizontal pressure acting on the vertical plane OC is the earth pressure at-rest. Slide planes exist in the inclined sand mass. However, as OC is a line of symmetry, shear stresses can not develop on it. Hence OC is a principal stress trajectory. Based on the equations of equilibrium, Jaky expressed the coefficient of earth pressure at-rest Ko with the angle of internal friction,8
(2.6) In 1948, Jaky presented a modified simple expression given by Eq. 2.7.
(2.7) Mayne and Kulhawy (1982) reported that, the approximate theoretical relationship
for Ko for normally consolidated soils supposed by Jaky appears valid for cohesionless soils. Using Jaky’s equation to estimate the in-situ lateral earth pressure is reliable for most engineering purposes.
2.2.3 Study of Mayne and Kulhawy
Mayne and Kulhawy (1982) investigated the relationship between Ko and (over-consolidation ratio)OCR for the soil with primary loading–unloading–reloading conditions. They considered the simplified stress history depicted in Fig. 2.8 for a homogeneous soil deposit with horizontal ground surface. Stress path OA represents virgin loading of the soil deposit, associated with sedimentation and normally–consolidated conditions. As represented by Fig. 2.8, the at-rest coefficient remains constant during virgin compression (Konc). Any reduction in the effective overburden stress results in overconsolidation of the soil, represented by path ABC.
As shown in Fig. 2.8, it is obviously that the overconsolidation ratio, OCR which is defined as
v,max /
v has a pronounced effect on the value of Ko. If loading is reapplied, the reload relationship subsequently will follow a path similar to CD in Fig.2.8. Subsequent unloading and reloading is likely to cause stress path to occur within the loop ABCDA.
To evaluate the behavior of horizontal stresses during vertical loading–unloading–reloading conditions. Mayne and Kulhawy (1982) reviewed laboratory data of 171 different soils tested and reported by many researchers from
statistical analysis as indicated in Fig. 2.9, Mayne and Kulhawy (1982) concluded the approximate theoretical relationship for Konc of normal consolidated soil introduced by Jaky (1944) is in good agreement with these data. Numerous investigators have suggested that Konc may correlate with liquid limits, plasticity index, clay fraction, uniformity coefficient, void ratio, or other index properties of the soil. However, on the basis of findings, the data collected did not confirm any of these relationships.
Mayne and Kulhawy (1982) deduced that only the effective stress friction angle, and prior stress history (OCR and OCRmax) are needed to predict approximate values of Ko.
2.3 Effects of Soil Compaction on Earth Pressure
Compaction of a loose soil can produce a stiff, low settlement-free and less permeable mass. It is usually accomplished by mechanical means that cause the density of soil to increase. At the same time the air voids were reduced.
Several theories and analytical methods had been proposed to analyze the residual lateral earth pressures induced by soil compaction. Most of these theories introduced the idea that compaction represented a form of over consolidation, where stresses resulting from a temporary or transient loading condition were retained following the removal of this load.
2.3.1 Study of Duncan and Seed
Duncan and Seed (1986) presented an analytical procedure for evaluation of peak and residual compaction-induced stresses either in the free field or adjacent to vertical, non-deflecting soil-structure interfaces. This procedure employs a hysteretic Ko
-loading model shown in Fig. 2.10. The model is adapted to incremental analytical methods for the evaluation of peak and residual earth pressures resulting from the
10
placement and compaction of soil. When the surcharge is applied on the soil surface, it will increase the vertical stress and the horizontal stress. In Fig. 2.10, as the virgin loading is applied on the soil, both v andh increase along the Ko-line (Ko = 1-sin).
Nevertheless, when the surcharge is removed,
v and
h would decrease along the virgin unloading path. As virgin reloading was applied again, the increment of earth pressure is less than that induced by the first virgin loading.Seed and Duncan (1983) brought up a simple hand calculation procedure which results in good agreement with the incremental procedure described above. In Fig.
2.11, it is apparent the simple hand solution has a good agreement with the incremental procedure.
2.3.2 Study of Chen and Fang
The distribution of horizontal earth pressure against the nonyielding wall after the compaction of soil Lift 1 to Lift 5 is shown in Fig. 2.12 (a) to (e). Each compacted lift was 0.3 m thick after compaction. The variation of lateral earth pressure was monitored by the soil pressure transducers mounted on the wall. After vibratory compaction, it is clear in Fig. 2.12 (a) to (e) that an extra horizontal normal stress
△h,ci was induced by compaction. The lateral stress distribution measured near the top of the backfill was almost identical to the passive earth pressure estimated with Rankine theory. From Fig. 2.12 (a) to (e), it is interesting to note that the compaction-influenced zone rose with the rising compaction surface. It was also interesting to note in Fig. 2.12(c) to (e) that, below the compaction-influenced zone, the measured horizontal stresses converged to the earth pressure at rest based on Jaky’s equation. In Fig. 2.12, data points obtained from Tests C0903 and C1141 indicated that the experimental results were quite reproducible.
The stress paths of
v versus
h for soil elements adjacent to the surface of thenonyielding wall are displayed in Fig. 2.13. Test data shown in Fig. 2.13(a) were measured by SPT 2 and SPT 102. In the figure, the path F1 represents the stress variation due to the “filling” of the loose Lift 1. It is clear that the stress path F1 is in good agreement with Jaky’s prediction. The filling of sand Lifts 1-5 stress (paths F1-F5) caused an obvious increase in vertical pressure.
Stress path C1 represents the stress variation due to the “compaction” on the surface of soil Lift 1. During the compaction of soil Lift 1 (stress path C1), the lateral earth pressure
h measured by SPT2 on the nonyielding wall increased significantly, but the vertical normal stress in soil mass was not affected by compaction. The compaction on Lift 2 (stress path C2) caused the h to increase further. However, the compaction on the surface of Lift 3 resulted in a lateral pressure reduction at SPT2 as indicated by the stress path C3. The compaction on the surface of Lifts 3 and 4 gradually brought the soil element located in front of SPT2 back to an at-rest stress condition. The horizontal earth pressure change was mainly caused by the compaction process, not soil filling. Similar trends can also be observed in Fig. 2.13 (b) and (c).In Fig. 2.14, the experimental test results are compared with the design recommendations proposed by Broms (1971), NAVFAC DM-7.2 (US Navy 1982), Duncan and Seed (1986), Peck and Mesri (1987), and Duncan et al. (1991). Parameter values used in the pressure calculation such as the unit weight
, relative density Dr,
internal friction angle, wall friction angle , and cyclic compaction stress
cyc are shown in Fig. 2.14. The horizontal pressure distribution suggested by the Navy Design Manual DM-7.2 was based on the analytical method proposed by Ingold (1979). The pressure distribution determined with the method proposed by Duncan et al. (1991) was obtained from the design chart for vibratory plates with a cyclic compaction stress q = 34.9 kN/m2 (5 psi).12
2.4 Cyclic Simple Shear Test
The cyclic simple shear test is a convenient method for determining the shear modulus and damping ratio of soils. It is also a convenient device for studying the liquefaction behavior of saturated cohesion less soils. In fig.2.15, Airey and Wood (1987) showed the NGI cyclic simple shear apparatus. In the cyclic simple shear test, a soil specimen, usually 20-30 mm high with a diameter of 60-80 mm, is subjected to a vertical effective stress v and a cyclic shear stress τcyc, as shown in Fig. 2.16. The horizontal load necessary to deform the specimen is measured by the horizontal load cell, and the shear deformation of the specimen is measured by the linear variable differential transformer LVDT.
2.4.1 Study of Youd
Youd (1972) reported the experimental results regarding the void-ratio reduction of sand due to cyclic simple shearing. Fig. 2.17 shows the gradual densification of sand by repeated shear displacement in a simple shear test. Each cycle of shear straining reduces the void ratio of the soil by a certain amount, although at a decreasing rate.
Decrease in volume of the sand, as shown in Fig. 2.17, can take place only if drainage occurs freely. In the figure, after 10,000 cycles, the void ratio of sand was reduced from 0.54 to 0.42. It is obvious from the figure that cyclic shearing is an effect measure to densify the cohesioless soil.
2.4.2 Study of Hsu and Vucetic
Hsu and Vucetic (2004) studied the volume decrease of dry or partially saturated sands subjected to several cycles of cyclic shear strain amplitudes
c. If the cyclicshear strain amplitudesc1 are smaller than a certain threshold value called the volumetric cyclic threshold shear straintv (c1 <tv), their volume will not change.
Such cyclic behavior is depicted schematically in Fig. 2.17. In the figure the results of three cyclic strain-controlled direct simple shear (DSS) tests conducted on dry or
represents the boundary between two fundamentally different types of volume change behavior. Belowtv , the soil particles are not displaced with respect to each other and the soil’s mineral skeleton and volume remain practically unchanged during cycling loading. When the soil is subjected toc >tv , the particles are displaced with respect to each other irreversibly, resulting in permanent changes of the soil’s volume and microstructure. It is clear in Fig. 2.18 that the cyclic shearing is an effective method to reduce the vertical strain of soil, and to densify the soil mass.
2.5 Cyclic Torsional Simple Shear Test
Fig. 2.19 shows the cyclic torsional simple shear device proposed by Ishibashi et al.
in 1985. In this device, a hollow cylindrical specimen 71.1 mm in outside diameter, 50.8 mm in inside diameter, and 142.2 mm in height, can be subjected to independent variations of axial stress, inner and outer confining pressure, and torsional shear stress
14
or strain. Thus the device closely simulates the ideal in situ condition and enable us to apply a known value of the lateral confining stress.
Static and cyclic torsional shear stresses can be applied by MTS closed-loop servo-hydraulic linear actuator by means of a torque loading rod, ball bearing spline, and axial loading piston. The movement of the MTS actuator is corrected by feedback signals either from the torque transducer for cyclic stress controlled tests, or from the rotational LVDT for cyclic strain controlled tests. The specimen is subjected to the programmed cyclic motion by the MTS commanding unit without any effect of the piston friction and torsional distortion of the loading piston and transducers.
2.5.1 Study of Ishibashi et al.
Ishibashi et al. (1985) studied the volume change of a hollow cylindrical Ottawa sand specimen subjected to cyclic torsional shearing in drained conditions. The experiments were conducted under uniform cyclic shear strains and the following conclusions were drawn. In Fig. 2.20, relationships between the cyclic volumetric strain the uniform cyclic shear straincyc for a given number of cycle is nearly linear.
It is clear in Fig. 2.17 that the volume reduction of the soil specimen is significantly influenced by the cyclic shear strain load
cyc and the number of cyclic shear stress application N.2.6 Densification with Cyclic Torsional Shearing
2.6.1 Study of Yang
Yang (2002) used the disc-shearing instrument at Chung-Yuan University to study
the soil settlement due to cyclic torsional shearing. The diameter of the shearing disc was 198 mm. The diameter of the cylindrical sandy specimen was 200 mm, and the height of the soil specimen was 105 mm. The cyclic shear tests were carried out with initial relative densities from 30 % to 50 %, and normal stresses applied from 7 kPa to 150 kPa. One-way and cyclic (N=1) shear stresses were applied on Mai Liao sand, Vietnam sand, and Ottawa sand. Fig. 2.21 shows the relative density increase due to cyclic shearing (N=1) was about twice that due to one-way shearing.
2.6.2 Study of Ren
Ren (2006) studied the soil densification due to cyclic torsional shearing. The diameter of the sandy specimen was 200 mm and the height was 105 mm. The diameter of the shear disc was 198 mm. Mailiao sand, Ottawa sand and Vietnam sand were tested with an initial relative density of 30 %. Normal stresses of 20, 60 and 100 kPa, and the shear angle 10˚, 20˚, 30˚, 60˚ and 90˚ were used for testing.
Fig. 2.22 showed the relative density of sand increased with increasing number of cyclic shear stress application N. The first 6 cycles of τcyc application was most change due to cyclic torsional shearing compaction. A new cyclic torsional shearing compactor was designed and constructed at NCTU. The thickness (T) of the soil after compaction was 0.6 m. The initial relative density was 34.5% ± 2.3% and the vertical
16
pressure on the surface loading was 9.2 kPa. Fig.2.24 showed the soil surface settlements after 1, 2, 5, 10, 20, 30 and 40 cycles of cyclic torsional shearing. It was obvious that the soil settlement increased with increasing number of cycles (N) of torsional shearing.
In the first 2 cycles of torque application, surface settlement increased significantly.
However, after N = 20, the major part of settlement has accomplished, soil particles were sheared and reached a densely-packed condition. Therefore, it was difficult to increase the settlement any further with more cyclic shear application.
Fig. 2.25 showed the relative density distributions of the compacted specimen for N
= 1, 2, 5, 10, 20 and 40. Test results showed that the density distribution increased with increasing number of cycles of torsional shearing.
In fig. 2.26, cyclic torsional shearing was applied on the surface of each 0.15m-thick lift, and the distribution of relative density in Lifts 1 to 4. Test results revealed that the trend of pressure distribution in each 0.15 m-thick lift was similar.
The average relative density achieved in each lift was greater than the required value of 70 %.
2.7 Assessment of Relative Density
ASTM Test Designation D-4253 (2007) provide a procedure for determining the minimum and maximum dry unit weights of granular soils. These unit weights can be used to determine the relative density of soil compacted in the field. The term relative density is commonly used to indicate the in situ denseness or looseness of a granular soil. Relative density is defined as
x 100% ( 2.10 )
Where e = in situ void ratio of the soil, emax = void ratio of the soil in the loosest state, emin = void ratio of the soil in the densest state.
Das (2010) reported that the value of Dr may vary from a minimum of 0 % for very loose soils to a maximum of 100 % for very dense soils. Soils engineers qualitatively describe the granular soil deposits according to their relative densities. In-place soils seldom have relative densities less than 20 to 30 %. Compacting a granular soil to a
Das (2010) reported that the value of Dr may vary from a minimum of 0 % for very loose soils to a maximum of 100 % for very dense soils. Soils engineers qualitatively describe the granular soil deposits according to their relative densities. In-place soils seldom have relative densities less than 20 to 30 %. Compacting a granular soil to a