Simulation Vacuum Preloading Method by TriAxial Apparatus

It is very important to control any risk of instability of embankment during vacuum construction, the simulation vacuum preloading method using tri-axial apparatus is proposed to predict the behavior of soft soil improvement in the laboratory, as well as to make this method become familiar and easier in the future. The tri-axial apparatus is used instead of the large-scale one, which has been performed by Bergado (1998) and Indaratna (2008). The tri-axial test on small size specimen can be carried out in one week compared to the large-scale apparatus takes one month for big specimen. In addition, the lateral deformation as well as the shear strength increase with time can determine accurately.


Introduction
Nowadays the vacuum preloading consolidation method becomes the popular method to improve soft soil.This method is an effective method of improving soft soil conditions as introduced by Kjellman (1952) [1] in early 1952.With the merging of new materials and technologies, this method has been further improved in recent years.
The modeling vacuum method to improve soft soil in the laboratory has been performed by Indaratna 2008 [2] using the large-scale apparatus and follows one-dimensional consolidation theory (Tezaghi).The results obtained from this modeling partially evaluated the behavior of soft soil reinforced by vacuum preloading method in the laboratory.Using the large specimen 45 cm × 90 cm in diameter and height respectively in the large-scale apparatus, the time used for this test was more than one month.
The horizontal deformation ε r during the tested time, which is the typical deformation of soft soil improvement by vacuum, and also the increasing of shear strength could not be measured.So far, the controlling surcharge processing during vacuum construction has not been discussed sufficiently.
The new method is proposed using tri-axial apparatus to simulate the comprehensive behavior of soil improvement by vacuum preloading method in the laboratory to support the engineering task quickly.In addition, it is desired to make the method become familiar in the fu-ture.
The finite element method (FEM) is used to analyze two cases of drainage condition at the boundary and center of the axisymmetric soil cell.
The study aspects to solve these matters are as follows: 1) Simulation the behavior of soft soil improved by vacuum method by tri-axial apparatus under axisymetric consolidation condition; 2) The lateral deformation and vertical settlement are concerned during soil improvement by vacuum preloading method; 3) Evaluating the degree of consolidation during soil improvement; 4) Combination surcharge and vacuum preloading to estimate the increasing shear strength).

Tri-Axial Apparatus
Tri-axial apparatus can clearly evaluate the failure mechanism as well as the capacity of increasing shear strength of soil in the laboratory.From the tri-axial test, the result parameters are used to predict the behavior of soil in the field during construction.
Under vacuum pressure condition alone, the soil mass at depth is subjected the isotropic stress status (K = 1).With flexible functions of the tri-axial machine as shown in Figure 1, the isotropic condition of the soil mass can generate the same vacuum condition by controlling the lateral earth pressure (K).
During vacuum condition, the surcharge loading can be generating as axial force by the loading rod at top of the machine.The deformations of soil specimen in vertical and horizontal direction are measured during testing to evaluate the behavior of soil specimen.
The specimen is covered by rubber membrane and placed in the water tank; therefore the friction between circular soil specimen and the cell is eliminate, which is different to the oedometer apparatus.
The steel plates at the ends of specimen are as the impervious layer, only radial drainage is induced during consolidation.The filter paper covers around the boundary of specimen and overlaps the porous discs at the both ends of specimen as the drainage layer.For the largescale oedometer apparatus, the drainage was established at the center of specimen.The FEM has been used to define the different between the drainage path conditions of specimen, and defines the correction factor for this simulation.The boundary conditions were illustrated in the Figure 2.
The assumptions were used for simulation of vacuum preloading method are as follow: 1) Soil mass as subjected vacuum pressure follows axisymmetric consolidation.
2) Under vacuum pressure only, soil mass will be subjected the Isotropic stress state, it mean that the coefficient of horizontal earth pressure (K) equal to one, while for the surcharge only (K) value can calculate from Equation (1).
3) For the soil mass, the vacuum pressure is distributed along to the specimen is uniform.  where φ : the friction angle

Drainage Boundary Condition Analysis
.

Thetheory of Asixmetric Consolidation
The unit cell theory representing a single circular surrounded by a soil annulus in an axisymmetric condition has been used (e.g.Barron, 1948; Yoshikuni and cepted that under embankment loading, the single drain analysis could not provide an accurate prediction due to lateral yield and heave compared to plane strain multidrain analysis (Indraratna, et al., 1997) although the degree of consolidation (DOC) in this model is acceptable accuracy.
In this topic, the method is proposed to model the behavior of soft soil improvement by vacuum combine with su variation of the pe dar re Hooke's la ring application.o (1 rcharge preloading method while the lateral displacement is concerned during consolidation. The following assumptions are based on Hansbo solution (1981) about equal strains (ε), the rmeability (k) when void ratio (e) decrease during consolidation and the volume compressibility (m v ).
1) Soil is homogeneous and fully saturated; the Darcy's law is adopted.Depend on the purposes the outer boun y of unit cell the drainage path is occurred.
2) Soil strain is uniform at the boundary of the unit cell and the small strain theory is valid, therefo w should be applied for calculation.
3) For the mass soil, the vacuum pressure distribution along to the drain boundary is uniform du The accuracy of the FE analysis was checked against the analytical solutions of Barron (1948) and Hansb 981) [3].According to Barron (1948) [4], the degree of consolidation U for "equal-strain" consolidation is given in Equation (2).
where D e and D w are diameters lent of vertical drain respectively.Hansbo (1981) intro-of unit cell and equivaduced a circular smear zone (of diameter D s ) in the solution which resulted in a modified expression for μ For two dimensional consolida pressed in terms of integrals of the excess pore water pr tion U(%) can be exessure over the unit cell domain as Madhav et al. 1993.
where u(x,y,T) = excess pore water pressure at any point with x, y dimension at a time factor T, an excess pore water pressure.
Terzaghi-Rendulic proposed differential equation for where C h : horizontal coefficient of consolidation.
As the strains are small, if E' effective stress, n' is Poisson's ratio for effective stress material is isotropic, Hooke's law is is Young's modulus for and the r r  where: {r, q, z} is the principal axes set.{ε r , ε q , ε z } are the strain in radial, circumferential and vertical respectively.oefficient of consolidation C h in the horizontal di train deformation is sh The c rection for axisymmetric plane s owed as: where:  v : volume strain of specimen σ' a : is the axial effective stress of consolidation.Three cases of axisymetric unit cells under vacuum preloading only as the Figure 3 were conducted to verify the relationship about degree of consolidation in the different boundary and drainage condition.
The first case, the outer boundary is fixed and drainage occurs at the center (FC) of unit cell.This is the conventional model to estimate the consolidation of axisymetric unit cell, which has been conducted by se s before such as Indraratna (2005) [5], Chai (2006) [6], Tran (2007) [7].One-dimensional consolidation theory is used for this case.As we know that, this case is very suitable for both cases the surcharge loading only and the vacuum zone is infinite.Therefore, this method should not apply well for vacuum preloading.
The second and the third case are proposed to verify the consolidation of axisymetric unit c eloading in term of the lateral displacement as well as two-dimensional consolidation is concerned, the outer boundary of specimen is free or none (N) combines drainage condition at the center (NC) and the outer boundary of unit cell (NB).
The series of case studies for axisymetric unit cell were conducted b own in the Table 1.The cases from N1 to N12 are to check the accuracy of FEM analysis with vary ratio (n) in 10 and 20 with vacuum pressure applied of 50 kPa and 100 kPa respectively.Comparisons were made for the cases of fix and none fix the outer boundary with ratio n = 10, n = 20 and vacuum pressure only Va = 50 kPa, Va = 100 kPa.The results of DOC (U%) and Time factor T h from the FEM were found to compare well with the analytical solutions Baron (1948) [4] as show in the Figures 4-7.
The maximum difference in U, for U > 50%, was about 0.26%, occurring at T = 0.5 for the case n = 10 as sh 20 own in Figures 4 and 6.Forthe case n = , the maximum difference in U was 0.17%, occurring at T = 0.5 as shown in Figures 5 and 7.
Conclusion the degree of consolidation in both cases fixed boundary (FC) and free boundary (NC) are almost sam  For the cases none outer boundary and drainage at the outer boundary (NB), degree of consolidation e almost same in both case n = 10 and n = 20 as during vacuum The Figure 8 and Table 2 show the relationship between the modeling of axisymetric unit cell in free boundary condition, drainage at the center (NC) and at the boundary (NB) during vacuum stage.
As the same DOC, the average ratio of time factor for drainage at the center and outer boundary is T hNC /T hNB about 6.7 and 9.7 for n = 10, and n = 20 respectively as showed from case N01 to N12.
The consolidation procedure of unit cell in (NB) case is faster than (NC) case by ratio (T hNC /T hNB ).The different values relative to (n) ratio are shown in the Table 2.
The average coefficient of consolidation (C h ) in both case (NC) and (NB) were found the same value and indepe

Solution for Axisymetric Unit Cell under Vacuum Combine Surcharge Loading
In this solution, the surcharge is applied after some degree of consol  Normally, the prefabricated vertical drain (PVD) of dimensions 10 cm × 0. The surchage stages are applied in three cases of the degree of consolidation (U) reach to 40%, 70% and 100% as shown in the Table 1.The loading rate is 0.5 kPa/min for drainage at the center case (NC), the ration T hNC /T hNB are used from the Table 2, of 10.2; 8.16 and 7.31 as degree of consolidation 40%; 70% and 100% respectively to define the loading rate when the drainage at outer boundary (NB).
The results of relationship between DOC (U) and time are shown from the Figures 10-12    In the Figure 14, degree of consolidation at 70%, the maximum different in volume strain is 1.75% occuring at 420 min for case 40% is almost the same in the Figure 15.
From the analyzing above, the final volume strain in case outer boundary is almost same value as that in drainage at the center as surcharge applied at 40%; 70% and 100%.
Under vacuum 50 kPa and surcharge was applied as degree of consolidation is 100%, the Figure 16 showed the deformation of unit cell (volume strain ε v ) during time of vacuum and vacuum combine surcharge are same in two cases (NC) and (NB) after adjustment with the ratio of time factor T hNC /T hNB .
.98 during the surcharge was apply.These results are go ccu ar

4.
From the Figure 13 and Table 3 the The strain ration ε r /ε a are the same in both cases and vary from 4 to 1.3 during vacuum stage and reduce to 0 od agreement with vacuum preloading theory, the lateral deformation of embankment is internal during applying vacuum pressure.
The maximum different in volume strain is 2.13% o rring at 350 min while the final volume strain is the same in 13.24 (%).
These results also agree strongly with the solution in ticle 2. To make this solution more effectively, the laboratory test by Tri-axial apparatus should be carried out to support this research.

Soil Specimen
The serial tests were performed by tri-axial apparatus in    the laboratory in Hokkaido University to simulate the behavior of Akasaoka clay improved by vacuum preloading method.The specimens of clay in sizes 75 mm × 150 mm in diameter and height respectively were used for this research, the specimens is suitable to control consolidation time under vacuum condition by tri-axial apparatus, also the retrieved samples from the field.For dummy research specimens were remodel in the laboratory from the commercial Akasaoka clay powder.The reconstituted Akasaoka clay was pre-consolidated under a pressure of 100 kPa and OCR = 1.25, e physiard odometer test result.The pre-consolidated condition, the effective vertical stress and the effective horizontal stress are 80 kPa and 40 kPa respectively (the coefficient of horizontal earth pressure at rest K 0 = 0.5).
Series conventional tri-axial test were carried out to verify the failure line (K f ) and relationship between the coefficient K and the ratio s u /σ v ' as show in Figure 18.

Test Procedure
Before application vacuum pressure, the specimen is saturated with B value more than 0.98 and reach to the initial pre-compression stress with effective vertical and horizontal stress are 80 kPa and 40 kPa respectively after ted by applying the efctive stress target with the lateral earth ratio equal to k th cal properties of soil are listed in the Table 4 and Figure 17.The permeability coefficient (k) and compression index (Cc) deduced from the stand 24 hours (Step loading and recompression step).
The vacuum pressure is simula fe one (K = 1), the soil specimen is subjected the same condition under vacuum pressure as the period research, the behavior of soil mass under surcharge and vacuum pressure is shown in the Figure 19.
While the effective stress is applied, the drainage vale is clocked, under the undrained condition the excess pore water pressure has been increasing up to the vacuum   Copyright © 2012 SciRes.IJG pressure desired.In the dummy research vacuum pressure, desired are 50 kPa and 100 kPa to concern the depths of specimen there for the total vertical effective stress target of 130 kPa and 180 kPa were generating.(Vacuum pressure supplying step).The parameters in vacuum proceed by tri-axial apparatus are shown in the Table 5.
The coefficient of horizontal earth pressure (K) is the ratio of the effective horizontal earth pressure due to the confinement from the surrounding soil mass to the vertical effective stress.Then from the Figure 20, K value can be calculated as follows: During ation has een occurred due to the excess pore water is dissipated, the effective stress of soil as well as the shear strength will be increase, this behavior is suitable with the vacuum mechanism, and the tri-axial apparatus's gauges record the behavior data and deformation of soil specimen automatically.The soil is consolidated completely when the effective stress increase to the target and the excess pore water dissipates completely.However, from data from FEM and test the times of primary consolidation reach to as the excess pore water pressure dissipates about 95%.
Where: CP: Cell pressure Pore water pressure σ v ': Vertical effective stress target σ h ': Horizontal effective stress target K: Coefficient of horizontal earth pressure The shearing steps are carried out to define the undraind shear strength of soil improved.Depend on the goals these step could be performed at the time after vacuum loading completely or with the increase of degree of consolidation reach to 40%, 70%, 100% combine with surcharge to estimate the capacity of soil embankment each construction stage respectively.This step is performed drained progress stage, the consolid b BP: Back pressure PWP:    some sections during simulate vacuum preloading method, pre-consolidation (AB), vacuum application (BC), surcharge loading (CDEF).Under vacuum condition, the stress path of soil moves from B to C far from the failure line, while it changes from D to E close to the failure line when surcharge was applied.The behavior of soil specimen under vacuum preloading method simulated by Tri-axial apparatus is well matched the before studies.

Conclusions
The study has been developed based on the combination of finite element analysis and the results of laboratory experiments to simulate a new appropriate method.This method can be widely applied for soft ground improvement by vacuum preloading method.The results om the etric unit cell used to model the behav- fr study can be summarized as follows: 1) The axisym ior of soil treatment by vacuum preloading as none boundary conditions are considered, the behavior of soil is more close to real soil state in the field.
2) Two cases of drainages boundary showed that time for consolidation in cases outside drainage of unit cell is faster than at the center by T hNC /T hNB ratio.However, the deformations of the specimens in all cases are in same shape and value with the same applied condition.
3) Results of experiments by tri-axial apparatus entirely agree with FEM model.It is suggesting that the theories are given full compliance, highly compelling to predict the behavior of soil improvement by vacuum preloading method.

igure 2 .
The drainage condition of axisymetric cell.

3 . 1 drain
Nakanodo, 1974;Hansbo, 1981).Most researchers ac-of soil. with horizontal hydraulic conductivity  w : unit weight of water Dummy research, for the axisymetric unit cell the cot of volume compressibility (m v ) varies during co ression: efficien nsolidation; m v value calculate by exp cm × 15.0 cm in diameter κ = 0.06, the horizontal and vertical at time reach to degree .Solution for Axisymetric U Vacuum Pressure An axisymetric sy the behavior of soil specimen improved by vacuum p loading method.he unit cell with size (D e ) and height (H) respectively (H = 2D), the effective Young's modulus E' = 500 kN/m 2 , Poisson's ration v' = 0.33, λ = 0.55, hy veral researcher ell under vacuum pr y FEM with Camclay Model, which sh draulic coefficient k h = k v = 4.66E-10 m/sec and the ratio n = D e /D w from 10, 20 are used to analyze.D w is equivalent diameter of vertical drain as in the Figure 3.

Figure 3 .
Figure 3.The modeling of axisymetric cell.Ta ll.N0 Case Bou Vacuum pressure (kPa) DOC (U%) ble 1.The case studies for axisymetric unit ce ndary condition Drainage condition Ratio D e /D w N1 FC-1-50 Fixed Center N2 N
4 cm is installed by rectangular shape.For this research the equivalent diameter of vertical drain D w = 5 cm and diameter of cell D e = 100 cm were used.The example with D e = 7.5 cm and D w = 0.3875 cm, and n = 20, were used in this analysis and shown in Fig- ure 9.This unit cell also used for modeling the vacuum preloading in the laboratory test.
rease vertical effective stress to prevent the failure state, o factor T There are two stages for vacuum preloading method, e first stage is vacuum preloading only until the degree th of c ion reach to the target, and the seco case the loadings are 19.8 kPa, 29.6 kPa and 42.4 kPa respectively when 50 kPa vacuum pressure is applying. :

Figure 12 .
Figure 12.Case n = 20, vacuum combine surchage loading; Va = 50 kPa, U = 100%.average of ratio of time factor are equal to 8.71 & 8.65 for vacuum pressure are 100 kPa and 50 kPa respectively, when U = 100% the ratio of time factor is nearly same value 7.4.In the Figure14, degree of consolidation at 70%, the maximum different in volume strain is 1.75% occuring at 420 min for case 40% is almost the same in the Figure15.From the analyzing above, the final volume strain in case outer boundary is almost same value as that in drainage at the center as surcharge applied at 40%; 70% and 100%.Under vacuum 50 kPa and surcharge was applied as degree of consolidation is 100%, the Figure16showed the deformation of unit cell (volume strain ε v ) during time of vacuum and vacuum combine surcharge are same in two cases (NC) and (NB) after adjustment with the ratio of time factor T hNC /T hNB .

Figure 19 .
Figure 19.Behavior of soil mass under vacuum and surcharge preloading.