The effect of clay mineral content on the dynamic response of reconstituted fine grained soil

Original Research Paper
  • 30 Downloads

Abstract

Fine grained soils with considerable amount of silt may exhibit sand-like or clay-like behavior depending on several factors such as the amount of fines and clay content, as well as the consistency limits, other variables being kept unchanged. This unpredictable behavior makes silts highly problematic, especially under seismic conditions. This paper describes the laboratory behavior of low plasticity Adapazari silt, known to be highly sensitive to cyclic loading. In the first phase of the basic study reported herein, Adapazari silt was mixed with different percentages of bentonite and kaolin and the behavior of these reconstituted mixtures was investigated in cyclic triaxial and dynamic simple shear tests. The purpose was to identify basic index properties and their threshold values to delineate sand- and clay-like behavior. Such a distinction may make it possible to complement field penetration resistance with appropriate adjustment factors to evaluate the pore pressure development potential, thus the risk of ground failure during an earthquake. The results show that there is a range of liquid limit and plasticity index values above which cyclic failure is significantly mitigated. It can now be stated that silts of intermediate and high plasticity may be deemed of relatively low potential for ground failure during seismic loading.

Keywords

Reconstituted samples Dynamic response Fine grained soil Bentonite Kaolin Sand-like Clay-like 

1 Introduction

Characterization of the seismic response of fine grained soils is one of the challenges in geotechnical earthquake engineering. It is necessary to identify the underlying mechanisms of ground failure under seismic conditions for cases where such types of soils are encountered. There is a critical need to better understand the dynamic behavior of fine grained soils so that safer and more economical designs can be implemented. One specific challenge is to predict the development of pore water pressures under cyclic loading to evaluate the ground failure potential of fine grained soils during an earthquake and to estimate post-earthquake settlement of structures. This paper is based on an extensive laboratory testing program that focuses on the dynamic response of fine grained soils and the development of excess of porewater pressures with reference to consistency. Two series of laboratory tests were performed using the cyclic triaxial (CTX) and dynamic simple shear (DSS) devices. Soil samples were prepared from slurry that was consolidated under constant stress. Stress controlled cyclic triaxial and dynamic simple shear tests were performed with two-way sinusoidal wave loading. Stress–strain behaviors as well as pore pressure development were monitored during testing. This paper presents the data from this comparative laboratory study and summarizes the major findings that may be useful to diagnose seismically sensitive silts.

2 Background

The authors found that it may either be difficult, costly or sometimes practically impossible to procure UD samples of non-plastic silts to assess their seismic behavior through dynamic laboratory testing. The undisturbed samples of Adapazari silt usually slump upon extrusion from the sampler. To avoid this problem, several methods have been proposed in the past to assess the vulnerability of fine grained soils to seismic failure using criteria based on their physical and index properties (Wang 1979; Onalp et al. 2001). These methods mostly utilize gradation and index properties such as liquid limit, plasticity or the liquidity index, fines content, clay content and a critical grain size such as the mean size, to provide a simple criterion to describe the seismic response of fine grained soils.

The so called “Chinese Criteria” were based on observations of “liquefaction” in fine-grained soils (Wang 1979). Existing guidelines for identifying “potentially liquefiable fine- grained soils” have been developed around the consistency limits, grain size characteristics, and natural water content, in various combinations. Some researchers suggested that the use of the “unreliable” Chinese criteria should be discontinued (Kramer et al. 2011).

These criteria have been extended by the authors of this paper (Bol et al. 2010) based on field observations and voluminous laboratory testing in the past, stating that in order for a fine grained soil to fail during an earthquake, all of the following conditions should be satisfied:
  • Liquid limit, wL ≤ 33%.

  • Ratio of the natural water content to liquid limit (wn/wL) for NP soils, alternatively, the liquidity index in the case of soils where plastic limit can be measured, are higher than 0.9.

  • Clay content, C % < 10.

  • Mean grain size, D50 > 0.02 mm.

It is obvious that one may not expect to duplicate the in situ generation of excess porewater pressures or the resulting strains by testing laboratory samples of limited size. However, there is significant benefit to run a series of tests to establish the trends for determining the stage where effective stress drops to near-zero or identifying the limiting strains where ground failure modes for soils emerge.

There is general agreement that increasing plasticity of the fines that a soil contains contributes to its liquefaction resistance. However, there is limited information pertaining to the influence of the activity of the clay mineral on the cyclic failure of fine grained soils. Ishihara mentioned a study (Ishihara 1993), where sand was mixed with kaolinite and bentonite and the finding that the cyclic resistance of the mixes were not influenced by the clay mineral type up to a plasticity index of 10 was presented. The resistance increased for higher IP values, albeit with considerable dispersion of the data points. It was accordingly decided to investigate this issue comprehensively by mixing the infamous Adapazari silt by an inactive clay mineral like kaolinite and highly active montmorillonite.

3 Experimental study

An attempt was made to investigate some of the predominant factors that influence the rise of porewater pressures to critical levels using properties such as consistency and clay content and its type, while keeping testing conditions unchanged, in order that an evaluation of the existing criteria may be achieved. The practicing engineer or a designer merely wishes to know whether a soil layer can be branded as “liquefiable” or cleared as “safe” under the existing seismic conditions at a site under study.

This research program consists of two parts. In the first stage, reconstituted mixes of Adapazari silt with kaolin and bentonite were prepared and dynamically sheared. The second part is a similar study on undisturbed samples procured from the field, to be reported in a separate paper.

Adapazari silt is the parent material encountered at sites where widespread ground failure around the City was observed during the 1999 Kocaeli earthquake (Bray et al. 2004). The properties of the components used in the reconstituted mixtures are summarized in Table 1. By starting with reconstituted samples prepared by the “pouring” method in accordance with Article 8.3.1, ASTM D5311/D5311M, it was foreseen to keep a close control over the texture and the physical properties. It was also intended that the structure of the mixes would not be significantly different, by utilizing the same preparation procedure for the cyclic triaxial (CTX) and dynamic simple shear (DSS) samples.
Table 1

Properties of the components used in preparing the blends

Component

Soil class

wL (cone)

wL (percussion)

wP

FC %

Clay %

Adapazari silt (washed)

ML

36

NP

NP

100

4.5

Bentonite

CH

135

133

34

100

82

Kaolin

MH

57

55

34

100

58

About 250 kg of airdried silt was stirred with ample water in a large bath, out of which its clay was floated, decanted, air dried and pulverized to achieve uniformity. Increasing percentages of bentonite or kaolinite were then added to the silt and mixed with distilled water to form a slurry with water contents at around 1.5 times the liquid limit, deaired and poured into consolidation cylinders with water on the top. Percentage of clays added to the samples were limited to 50% because the silt started floating in the clay matrix and, clay rather than the silt would have been tested otherwise. Cylindrical samples of 100 mm diameter and 200 mm high were consolidated in Lucite cylinders with top/bottom and side drainage before being placed in the dynamic testing systems. A standard consolidation pressure of 100 kPa was selected because liquefaction is predominant in the top layers of silt in situ and also the use this stress is recommended (Ishihara 1993). It may be pertinent to point out here that all samples tested in the program were finer than 0.074 mm (minus No. 200). Consolidation phase lasted 3–15 days, depending on the type and percentage of the added clay. The completion of consolidation was decided as the stage when outflow from the top and bottom drainage valves ceased and the dial gauge readings measuring compression did not increase. The samples became anisotropically consolidated by this technique, which closely represents the DSS testing conditions. In addition, all samples were normally consolidated because they were subsequently dynamically tested at the same cell/normal pressure. The sampling tools were pushed into the consolidated slurry while the water level was still above the top of the sample in the cylinder and they were extruded from the tube to be cut into the ring (DSS) or the tube (CTX) at the base. The saturation of the CTX sample was secured by applying a back pressure. The ring of the DSS was placed on the Teflon rings that were already immersed, to prevent air intruding into the sample and pushed into place. Each sample was carefully removed from the cells at the end of a test and the degree of saturation was rechecked by weighing. The condition of consolidation is slightly different in the CTX because the sample is under isotropic pressure in the cell. It is nevertheless obvious that this difference has not constituted a major obstacle, because the CTX sample was not subjected to further consolidation as the cell pressure was kept unchanged σ3 = 100 kPa.

Figure 1 shows the changes in the liquid limits as determined in the fall cone device (BS1377 1990) and the plasticity indices of the tested samples as a function of percentage of added clay. The marked increase in the liquid limit of Adapazari silt mixed with bentonite was not observed for the kaolin mixes, supposedly due to the low activity of kaolinite which classifies as MH.
Fig. 1

The influence of clay content on consistency a liquid limit, b plasticity index

4 Test results: porewater pressure ratios

The samples were tested in a cyclic triaxial system (CTX) manufactured by Wykeham Farrance in accordance with ASTM D5311 (2011) and in the dynamic simple shear device (DSS) manufactured by Geocomp complying with ASTM D6528 conditions (2007). A total of ninety samples were prepared and tested. A constant and reasonably high cyclic stress ratio (CSR = 0.35) was applied at a frequency of f = 0.5 Hz (Fig. 2a) possibly representing a medium rate cycle where the rise of porewater pressures could be recorded unimpeded during CTX and DSS testing. The cyclic shear ratios are defined as (Seed and Idriss 1971; Boulanger and Idriss 2004):
Fig. 2

Behavior of the 18% bentonite mix in dynamic simple shear (DSS). a Applied cyclic stress, b rise of pore pressure, c increase of shear strain γ, d deformation loops during shear, e descent of normal stress by increasing number of cycles

$$ {\text{CTX}} \to {\text{CSR}} = \frac{{\sigma_{\text{dev}} }}{{2\upsigma_{3}^{\prime } }} $$
(1)
$$ {\text{DSS}} \to {\text{CSR}} = \frac{{\uptau_{\text{av}} }}{{\upsigma_{{{\text{V}}0}}^{\prime } }} $$
(2)
Porewater pressures are measured directly at the pedestal in the CTX test. The value of uw on the other hand, is inferred from the changes in the vertical stress in the DSS test, because this test is performed at constant volume. The details of all the tests performed are presented in Tables 2, 3, 4, 5 for the use of researchers of the subject.
Table 2

CTX test results: bentonite blends

Bentonite %

Residual-ru10

Residual-ru15

Peak

Number of cycles

Specific gravity (Gs)

Void ratio (e)

Natural unit weight (ρn)

Plasticity index (Ip, %)

9.5

10.0

14.5

15.0

ru10

ru15

N15

N10

N5

  

(kN/m3)

Fall cone

Percussion

0.0

0.750

0.683

0.800

0.741

0.89

0.98

16.386

14.450

10.250

   

NP

NP

1.0

0.835

0.688

0.844

0.732

0.89

0.92

26.290

23.503

13.742

2.69

  

7

6

1.0

0.883

0.797

0.898

0.859

0.99

1.00

22.994

25.894

17.063

     

1.5

0.869

0.947

0.946

0.928

0.98

0.99

19.255

16.397

8.976

2.74

1.11

19.01

NP

NP

1.5

0.762

0.718

0.799

0.743

0.97

0.99

13.868

13.868

18.922

     

3

0.893

0.863

0.911

0.917

1.00

1.00

30.000

28.111

26.445

2.70

0.94

18.91

10

5

3

0.889

0.864

0.898

0.951

1.00

1.00

26.395

27.847

21.424

     

4

0.825

0.835

0.884

0.884

1.00

1.00

22.647

23.977

18.892

2.70

0.91

19.18

NP

NP

6

0.874

0.903

0.917

0.864

0.91

0.92

12.558

18.692

13.949

2.70

0.80

19.02

8

5

7.5

0.831

1.000

0.908

1.000

1.00

1.00

21.350

24.230

24.361

2.71

0.96

18.86

9

6

9

0.810

0.771

0.791

0.815

0.97

0.99

26.835

28.884

19.785

2.65

0.90

19.11

12

10

9

0.866

0.762

0.881

0.813

0.99

0.98

26.802

27.692

24.107

     

10.5

0.740

0.773

0.823

0.863

0.85

0.88

17.243

21.855

23.288

2.66

0.92

19.08

11

10

12

0.681

0.667

0.771

0.746

0.72

0.81

16.884

15.217

20.012

2.73

0.95

19.06

13

9

12

0.857

0.824

0.863

0.920

0.99

1.00

21.377

16.577

3.291

     

12

0.755

0.717

0.883

0.836

0.93

0.93

28.601

28.530

25.430

2.73

    

15

0.827

0.874

0.913

0.961

0.94

0.97

18.233

21.461

14.877

 

0.95

19.02

15

13

16

0.787

0.769

0.836

0.838

0.80

0.86

23.915

26.844

29.711

2.66

0.87

18.87

19

16

18

0.825

0.773

0.864

0.841

0.92

0.96

28.561

28.608

28.490

2.65

0.95

 

18

16

25

0.485

0.480

0.556

0.541

0.49

0.56

26.728

27.741

25.629

2.65

0.88

18.45

24

23

25

0.868

0.840

0.890

0.855

0.98

1.00

19.714

14.798

9.858

     

30

0.408

0.386

0.451

0.454

0.56

0.60

26.033

27.712

19.004

2.64

0.96

18.59

27

25

35

0.480

0.448

0.547

0.529

0.58

0.64

32.439

32.439

31.351

2.73

1.09

17.97

27

24

40

0.241

0.223

0.296

0.265

0.39

0.46

1.288

1.035

0.795

2.63

1.01

18.34

27

27

50

0.243

0.188

0.324

0.255

0.34

0.43

3.086

1.577

0.965

2.66

1.15

17.81

39

35

ru10: residual pore pressure ratio after 10 cycles of loading, %εN15: percent strain at 15 cycles

Table 3

DSS test results: Bentonite blends

Bentonite %

Residual-ru10

Residual-ru15

Peak

N15

N10

N5

Gs

e

9.5

10.0

14.5

15.0

ru10

ru15

N15

N10

N5

0.0

0.875

0.927

0.962

0.983

0.981

1.000

9.432

6.394

3.106

2.69

 

0.0

0.860

0.851

0.928

0.916

0.860

0.929

13.746

9.114

4.531

  

1.5

0.943

0.915

0.943

0.932

0.950

0.946

19.611

15.733

4.171

2.74

0.81

3

0.895

0.913

0.917

0.915

0.913

0.919

19.039

14.443

8.910

2.70

0.75

3

0.901

0.913

0.904

0.906

0.913

0.927

20.012

15.730

8.556

  

4

0.858

0.854

0.860

0.860

0.858

0.860

30.000

30.652

14.869

2.70

0.79

6

0.918

0.902

0.910

0.913

0.918

0.919

30.969

23.343

8.251

2.70

0.76

7.5

0.739

0.802

0.739

0.802

0.814

0.814

30.000

30.000

18.954

2.71

0.84

9

0.824

0.827

0.842

0.850

0.839

0.855

25.527

26.224

8.291

2.65

0.74

10.5

0.806

0.815

0.848

0.848

0.838

0.862

25.909

19.205

7.263

2.65

0.73

12

0.736

0.746

0.882

0.850

0.747

0.882

26.872

15.249

4.584

2.73

0.80

15

0.870

0.885

0.938

0.938

0.885

0.941

30.000

31.072

6.376

2.66

0.77

16

0.804

0.741

0.863

0.802

0.804

0.863

31.057

21.869

8.094

2.66

0.77

18

0.670

0.717

0.861

0.871

0.717

0.871

41.712

11.660

3.770

2.65

0.77

20

0.419

0.425

0.499

0.510

0.428

0.510

9.521

5.126

2.989

2.72

0.83

23

0.383

0.386

0.482

0.477

0.394

0.494

4.976

3.475

2.550

2.71

0.88

25

0.383

0.428

0.480

0.537

0.428

0.543

5.692

4.234

2.431

2.65

0.81

30

0.314

0.320

0.369

0.374

0.320

0.374

2.733

2.432

1.988

2.64

0.84

35

0.252

0.245

0.296

0.281

0.263

0.309

3.263

2.975

2.106

2.73

1.04

40

0.211

0.212

0.246

0.248

0.217

0.253

2.478

2.371

1.561

2.63

0.88

50

0.250

0.261

0.293

0.296

0.261

0.303

3.535

3.256

2.376

2.66

1.03

Table 4

CTX tests results: Kaolin blends

Kaolin %

Residual-ru10

Residual-ru15

Peak

Number of cycles

Specific gravity (Gs)

Void ratio (e)

Natural unit weight (ρn)

Plasticity index (Ip, %)

9.5

10.0

14.5

15.0

ru10

ru15

N15

N10

N5

  

(kN/m3)

Fall cone

Percussion

0.0

0.750

0.683

0.800

0.741

0.893

0.976

16.386

14.450

10.250

   

NP

NP

1.5

0.802

0.750

0.820

0.766

0.951

0.961

20.006

17.674

12.524

  

19.21

NP

NP

3

0.820

0.830

0.888

0.868

0.995

1.000

19.989

18.236

11.212

2.73

0.93

19.33

NP

NP

3

0.868

0.829

0.902

0.917

0.976

0.971

21.471

23.429

16.197

     

3

0.815

0.840

0.928

0.922

0.990

1.000

23.445

22.292

19.783

     

4

0.813

0.855

0.884

0.957

0.855

1.000

22.767

23.956

20.782

2.70

0.90

19.41

NP

NP

6

0.906

0.887

0.892

0.917

1.000

1.000

22.304

21.253

14.978

2.67

0.87

19.65

NP

NP

6

0.868

0.966

1.000

0.888

1.000

1.000

22.552

23.323

20.563

     

7.5

0.803

0.842

0.837

0.947

1.000

1.000

19.950

20.091

14.833

2.70

0.83

19.79

NP

NP

9

0.863

0.888

0.932

0.990

1.000

1.000

21.412

24.277

18.761

2.70

0.84

19.32

9

6

10.5

0.839

0.845

0.864

0.903

1.000

1.000

21.874

23.960

20.000

2.65

0.75

19.50

10

7

12

0.908

0.985

0.941

0.927

0.995

0.971

16.754

21.909

22.710

2.69

0.79

19.66

9

6

12

0.912

0.927

0.922

0.946

1.000

1.000

23.795

23.930

18.168

     

15

0.797

0.845

0.849

0.841

1.000

0.990

25.409

26.881

20.668

2.70

0.77

20.01

10

8

15

0.789

0.888

0.834

0.966

1.000

1.000

21.475

21.453

21.431

     

16

0.743

0.912

0.821

0.965

1.000

1.000

26.394

24.839

18.011

2.70

 

20.13

9

6

16

0.704

0.616

0.714

0.647

0.839

0.879

27.115

25.517

18.503

     

18

0.789

0.777

0.829

0.840

0.986

0.986

20.373

22.158

16.816

2.70

0.74

20.45

9

7

20

0.743

0.912

0.784

0.951

1.000

1.000

21.625

21.172

19.396

2.70

0.88

20.30

10

8

25

0.926

0.891

0.975

0.950

1.000

1.000

20.830

25.201

19.548

2.68

0.66

20.53

9

7

25

0.821

0.783

0.883

0.870

0.927

0.927

21.951

24.755

22.327

     

30

0.854

0.917

0.893

0.917

0.917

0.940

15.373

20.721

23.829

2.69

0.70

20.28

11

9

40

0.760

0.589

0.791

0.668

0.904

0.937

23.930

23.955

16.411

2.71

0.78

20.24

13

10

50

0.680

0.660

0.768

0.730

0.765

0.837

25.424

26.578

25.976

2.68

0.75

19.72

12

10

ru10: residual pore pressure ratio after 10 cycles of loading, %εN15: percent strain at 15 cycles

Table 5

DSS tests results: Kaolin blends

Kaolin %

Residual-ru10

Residual-ru15

Peak

N15

N10

N5

Gs

e

9.5

10.0

14.5

15.0

ru10

ru15

N15

N10

N5

0.0

0.875

0.927

0.962

0.983

0.981

1.000

9.432

6.394

3.106

2.69

 

1.5

0.895

0.891

0.917

0.914

0.906

0.920

12.476

9.177

5.173

  

3

0.865

0.859

0.887

0.891

0.880

0.904

13.724

9.653

5.267

2.73

0.77

4

0.855

0.885

0.896

0.898

0.889

0.904

20.645

10.981

6.395

2.70

0.73

6

0.873

0.845

0.882

0.879

0.892

0.904

17.141

11.099

6.654

2.67

0.66

7.5

0.805

0.804

0.885

0.831

0.839

0.851

29.145

19.496

9.373

2.70

0.68

9

0.865

0.859

0.880

0.880

0.877

0.895

32.948

32.948

15.843

2.70

0.75

10.5

0.932

0.886

0.960

0.919

0.932

0.963

27.063

19.773

13.852

2.65

0.71

12

0.799

0.827

0.849

0.858

0.838

0.865

33.756

23.412

16.012

2.69

0.68

15

0.817

0.820

0.830

0.845

0.836

0.849

26.693

22.726

15.357

2.70

0.64

16

0.803

0.822

0.831

0.831

0.822

0.840

28.635

19.465

11.337

2.70

 

18

0.816

0.837

0.852

0.858

0.837

0.862

22.901

16.707

10.517

2.70

0.57

20

0.901

0.882

0.905

0.910

0.916

0.923

31.979

21.820

10.517

2.70

0.60

25

0.796

0.841

0.830

0.830

0.853

0.860

22.230

18.516

11.789

2.68

0.57

30

0.779

0.816

0.835

0.834

0.868

0.903

26.173

19.743

11.219

2.69

0.59

40

0.745

0.755

0.852

0.877

0.783

0.902

15.866

10.434

5.262

2.71

0.59

50

0.740

0.719

0.795

0.770

0.806

0.847

34.383

21.612

11.637

2.68

0.67

100

0.607

0.623

0.687

0.690

0.632

0.698

27.899

14.834

8.155

  

Figure 2 is a typical example of soil response in the DSS. This mix which has a relatively high percentage of bentonite, appears to develop monotonously rising porewater pressure up to N = 15 cycles (Fig. 2b). Although the porewater pressure does not rise sufficiently to lower the effective stress to zero, remaining stationary at around σn = 10 kPa (Fig. 2e), it can be seen from Fig. 2c that the shear strains abruptly rise above ± 15% after 15 cycles of loading, indicating the yield of the sample (Fig. 2d). This confirms the view that intermediate to high plasticity silts can soften and experience significant cyclic straining even though the effective stress does not drop to zero. In other words, failure is manifested by excessive deformation in clay-like soils as opposed to sand-like soils where pore pressure ratio ru rapidly reaches unity (Idriss and Boulanger 2008). These results also imply that clayey soils may reach significant levels of pore pressure which may make it important to evaluate their post-earthquake deformation potential that may be critical for structures founded on such soils.

Figure 3a illustrates the percent peak excess porewater pressures Uwpeak, as a function of the liquid limit measured by the fall cone method. The values indicate that samples with wL < 39% have attained porewater pressure ratios of ru = 0.8 to 1.0 indicating failure by sand-like behavior when they are cyclically sheared. Another observation made is the change to a state of limited excess pore water pressure development after the liquid limit exceeds 40%. Pore pressures did not exceed ~ 60 kPa beyond this liquid limit, corresponding to wL ≈ 35% for the percussion method. This shows that NP or low plasticity soils start developing resilience to cyclic loading above this liquid limit, and may start exhibiting clay-like behavior.
Fig. 3

a Peak, b residual pore water pressures at N = 15 cycles versus liquid limit (fall cone)

It may be pertinent at this point to describe what is meant by sand-like and clay-like soil. These terms have been proposed for the case of non-plastic and low plasticity silts (ASTM D6528 2007; Seed and Idriss 1971; Seed et al. 2003) because damage observations at the surface show that ground failure can be attributed to sand-like soils rather than those with clay-like characteristics. This leads to reserving the term “liquefaction” for describing sand-like response and the term “cyclic failure” for clay-like behavior in fine-grained soils. Clay-like behavior has been conservatively identified for soils with plasticity indices higher than 7% (Boulanger and Idriss 2004) as well as at IP ≥ 12% (Idriss and Boulanger 2008; Bray and Sancio 2006; Donahue et al. 2007).

It has also been argued that peak pore pressures are instantaneous measurements of response and they are also directly influenced by the deviatoric stresses during a test. “Residual” pore pressure uwres was defined as the value of pore pressure where the deviator stress crosses the zero level during a loading cycle, to eliminate stress induced effects (Boulanger et al. 1998). In an example shown in Fig. 4, the residual pore pressures at the beginning and the end of the 10th cycle are evaluated individually and averaged to obtain the residual pore water pressure for this cycle. The data presented in the following sections henceforth utilize residual pore pressures to establish behavioral trends.
Fig. 4

Selecting the residual and peak pore water pressures in DSS (N = 10 cycles)

The residual excess pore pressures result from the progressive collapse of the soil skeleton i.e., plastic deformations, thus alter the effective stresses acting on the soil. Consequently, it can be said that Uwres directly influences the strength and stiffness (Dobry et al. 1982). The residual Uwres values are plotted in Fig. 3b for CTX an DSS tests. Uwres values have accordingly been employed in the subsequent evaluations. The curves in Fig. 3 and subsequent plots were approximated by a sigmoid transfer function of the form shown below. The advantage of the sigmoid function is the way it can handle smooth transitions to sharp drops in different trends of data in relation to x0 and b parameters:
$$ {\text{y}} = {\text{y}}_{0} \frac{\text{a}}{{1 + {\text{e}}^{{ - \left( {\frac{{{\text{x}} - {\text{x}}_{0} }}{\text{b}}} \right)}} }} $$
(3)
Figure 5 depicts a comparison of the liquid limit values measured by the percussion and the fall cone devices versus the residual excess pore pressure recorded. It is observed in Fig. 5a that the excess pore pressures generated drop sharply at around liquid limit 40% and the rures values are 0.6 or lower beyond this limit. This would correspond to a wL ≅ 35% in the percussion method (Koester 1992). The advantage of the fall cone method over percussion is that a liquid limit value is always measured by this method, even though the test may be concluded as “non-plastic” in the percussion device (Fig. 5b). The correlation is improved as the bias of “non-plastic” assessment of the percussion test is replaced by the wL values that could be measured by the fall cone. Figure 5 also indicates that all blends of kaolin reached the state of initial liquefaction.
Fig. 5

Residual pore pressures at N = 15 cycles with respect to liquid limit from a fall cone percussion, b percussion

Several investigators stated that the plasticity index is a better indicator of cyclic failure than the liquid limit or the clay content values (Bray and Sancio 2006; Seed et al. 2003). Its use to improve the judgment has nevertheless not indicated a higher correlation in this study, as shown in Fig. 6. The plasticity indices determined by the liquid limit measured in the fall cone device have shown a slightly lower correlation than when wL was measured in the percussion method. Nevertheless, one may state that a plasticity index of IP = 16% delineates the transition from sand-like to clay-like behavior (Boulanger and Idriss 2006). This transition is defined at a plasticity index of 20%, when the liquid limit was measured in the fall cone test. It is an interesting observation that all the sample points with kaolin are confined to the left of IP ≤ 14% because of the low activity of this mineral. One may accordingly question the efficacy of selecting IP in identifying the dynamic behavior of soils such as silts.
Fig. 6

Residual pore pressures at N = 15 cycles with respect to Plasticity Index, wL from a percussion, b fall cone

Having determined that the porewater pressure ratios will be limited to below 0.6 if a silt blend has a liquid limit wL > 36% or IP > 16%, the next step was to consider the level of cyclic shear strain at N = 15 cycles instead of relying solely on the excess porewater pressure values generated. The limit of N = 15 cycles was selected for evaluation because failure was reached at this level in all the samples tested for the representative CSR value selected.

Another attempt was made to see whether clay content has an appreciable control over the generation of excess pore pressures, as suggested by some investigators (Koester 1992). Figure 7 shows an interesting trend where bentonite blends start developing high pore pressures as clay content drops to below 18%. Kaolin blends however continued generating high porewater pressures even at clay contents as high as 24%, showing the influence of the clay mineral over the process. This finding suggests that using clay content alone to prognose liquefiability cannot be a universal criterion but may be of use at regional level, where similar or identical clay minerals are present in the silt.
Fig. 7

Residual pore pressure at N = 15 cycles with respect to a clay content, b D50 for Bentonite and Kaolinite mixtures

The consistent sigmoidal relationship among the generated pore pressures and physical properties for bentonite blends is also observed when mean size D50 is plotted against Uwres (Fig. 7b). Although 0.03 mm appears as the critical size for bentonite mixes, the correlation is not as high as observed for consistency properties, suggesting it may be the pore size distribution rather than a single grain size limit that controls the process. The observed insensitivity of kaolin mixes to mean particle size may be an indication of this opinion.

5 Study of strains

The initial part of this study was aimed to identify the potential for failure of artificial mixes of different plasticity in dynamic conditions. The porewater pressures generated at reasonable number of cyclic loadings indicated that non-plastic or low plasticity silts may develop initial liquefaction. Therefore, it can be stated that increases in the porewater pressures leading to a reduction of the effective stress towards zero is a clear sign of initial liquefaction in sandy and non-plastic fine grained soil.

The second part of this study aimed to understand what happens to those mixes which fall on the right wing of the sigmoid curve, hence labeled as not to have failed by initial liquefaction. Excess porewater pressures generated in the clay-like samples during cyclic loading indicate only limited increases, suggesting that initial liquefaction does not take place, as clearly shown in Fig. 5, where Δuw does not exceed a threshold level corresponding to a porewater pressure ratio of around 0.6. One can accordingly assume that such clay-like mixes are not vulnerable to initial liquefaction.

Assuming that the strains (γ) in the DSS test are related to the axial strain in the CTX(εz) by the relationship based on elastic theory (Vucetic and Dobry 1991);
$$ \upgamma \cong (1 +\upupsilon)\upvarepsilon_{z} $$
(4)
gives a ratio of 1.5 for a Poisson ratio of υ ≈ 0.5, a plot of the test results using this ratio caused excessively high scatter for the DSS results. However, points clustered if measured values of εz and γ are plotted directly, as shown in Fig. 8. It can be seen that a liquid limit value of 42%, corresponding to 38% in the percussion test, drags the points to below 5% Double Strain Amplitude (DSA) line.
Fig. 8

The influence of the number of cycles on the strains a N = 5, b N = 10, c N = 15 cycles

Adopting a strain amplitude of γ = ± 5% as the failure criterion, a liquid limit of wL = 46% measured in the fall cone test defines a clear border beyond which failure under dynamic conditions does not arise, regardless of the number of load cycles applied. It can accordingly be stated that those mixes of intermediate (I) and high plasticity (H) shall under no circumstance reach failure state whether failure is defined by development of excess pore pressure or a limiting strain (Boulanger and Idriss 2006).

Another important observation made is that all mixes of kaolin exceeded γ = 5% DSA in the tests even at N = 5 cycles of loading. The strains of the bentonite mixes nevertheless were measured to be higher than those of kaolin until the threshold value wL = 45% is reached.

It may be debated at this stage that the test conditions imposed to reach these conclusions by applying a constant CSR of 0.35 might have been unduly harsh. A companion study was therefore carried out where the silt was enriched by blending it with increasing percentages of its own clay fraction. Using the sigmoidal function given in Formula 4 for regression, it was found that for N = 15 cycles, none of the samples failed at CSR = 0.15, whereas at least 12% clay was required to prevent failure for CSR = 0.20. All samples failed if CSR was 0.25 or higher. It was a surprising finding of this study that using clay contents to define failure gave a better correlation than the liquid limits.

6 Conclusions

The cyclic failure, or liquefaction, of silt in the laboratory is a complex process where several interactive factors such as the consistency, fines content and type of clay minerals are involved, other factors such as cyclic shear stress, confining pressure and the technique used for specimen preparation being kept unchanged. For the artificial mixes studied, definition of failure in the form of reaching zero effective stress requires liquid limits below 40% as measured by the fall cone apparatus, corresponding to about 36% in the percussion test. This is well above the limit of 33 required by the Adapazari Criteria as well as the other similar criteria. A plasticity index of around 16% is representative of failure conditions at cyclic stress ratios of 0.35. A clay content of 18% is the additional upper limit for initial liquefaction. It may therefore be stated that those samples of intermediate and high plasticity will never reach failure stage in this mode. The reason for the discrepancy between the findings of this study and previous works can be explained by the absence of structure development for the mixes which were only 4–6 weeks old.

Those mixes with wL > 40% which did not develop excess pore pressures to cause failure may nevertheless undergo shear strains that may be unacceptably high. In addition, it was found that those mixes of intermediate and high plasticity shall under no circumstance reach failure state whether failure is defined by ru or by limiting strain.

It must be pointed out at this stage that the deformational behavior of samples with different clay minerals were found to be radically different, mixes with kaolinite reaching failure state much easier than those containing bentonite. The profound effect of clay mineral on the cyclic behavior of silt is confirmed.

It was found that as the natural clay content of Adapazari silt was increased, the mix became more sensitive to increasing cyclic stress ratios. Ongoing testing with undisturbed samples of the natural silt is expected to shed light on this controversial behavior. It is speculated at this stage that the differences in pore structure maybe a major factor in determining the cyclic rigidity of the samples.

The response of the samples in the CTX and the DSS tests were similar. However, the scatter of data was observed to be higher in the CTX, possibly due to the bigger height of the latter samples. On the other hand, the possibility of monitoring the porewater pressures directly in the CTX appears as an advantage over DSS where uw is calculated indirectly.

Notes

Acknowledgements

This work was conducted carried out by the support of the Turkish Foundation for Scientific and Technical Research TÜBITAK under project 106M042. The senior author is thankful for the encouragement given by the Virginia Tech during her stay in 2010–2011. Special thanks are due to Professor James. R. Martin for his support during the preparation of this paper.

References

  1. ASTM D6528 (2007) standard test method for consolidated undrained direct simple shear testing of cohesive soilsGoogle Scholar
  2. ASTM D5311 (2011) Standard test method for load controlled cyclic triaxial strength of soilGoogle Scholar
  3. Bol E, Onalp A, Arel E, Sert S, Ozocak A (2010) Liquefaction of silts: the Adapazari criteria. Bull Earthq Eng 8:859–873CrossRefGoogle Scholar
  4. Boulanger RW, Idriss IM (2004) Evaluating the potential for liquefaction or cyclic failure of silts and clays. Report UCD/CGM 04/01, University of CaliforniaGoogle Scholar
  5. Boulanger RW, Idriss IM (2006) Liquefaction susceptibility for silts and clays. J Geotech Geoenviron Eng 132(11):1413–1426CrossRefGoogle Scholar
  6. Boulanger RW, Meyers MW, Mejia LH, Idriss IM (1998) Behavior of a fine-grained soil during the Loma Prieta earthquake. Can Geotech J 35:146–158CrossRefGoogle Scholar
  7. Bray JD, Sancio RB (2006) Assessment of the liquefaction susceptibility of fine grained soils. J Geotech Geoenviron Eng 132(9):1165–1177CrossRefGoogle Scholar
  8. Bray JD, Sancio RB, Durgunoglu T, Onalp A, Youd TL, Stewart JP, Seed RB, Cetin KO, Bol E, Baturay MB, Christensen C, Karadayilar T (2004) Subsurface characterization at ground failure sites in Adapazari, Turkey. J Geotech Geoenviron Eng 130(7):673–685CrossRefGoogle Scholar
  9. BS1377 (1990) Methods of test for soils for civil engineering purposes. British Standards Institution, LondonGoogle Scholar
  10. Dobry R, Ladd R, Yokel F, Chung R, Powell D (1982) Prediction of pore water pressure buildup and liquefaction of sands during earthquakes by the cyclic strain method. NBS Building Science Series 138. National Bureau of Standards, U.S. Department of Commerce, WashingtonGoogle Scholar
  11. Donahue JL, Bray JD, Riemer RM (2007) The liquefaction susceptibility, resistance and response of silty and clayey soils. USGS Research Report, BerkeleyGoogle Scholar
  12. Idriss IM, Boulanger RW (2008) Soil liquefaction during earthquakes. EERInstitute, CaliforniaGoogle Scholar
  13. Ishihara K (1993) Liquefaction and flow failure during earthquakes. Geotechnique 43(3):351–415CrossRefGoogle Scholar
  14. Koester JP (1992) The influence of test procedure on correlation of Atterberg limits with liquefaction in fine-grained soils. Geotech Test J ASTM 15(4):352–360CrossRefGoogle Scholar
  15. Kramer SL, Huang YM, Greenfield MW (2011) Performance based assessment of liquefaction hazards. In: Proceedings of geotechnics for catastrophic flooding events, Taylor & Francis, London, vol 1, pp 17–26Google Scholar
  16. Onalp A, Arel E, Bol E (2001) A general assessment of the effects of 1999 earthquake on the soil-structure interaction in Adapazari. XV ICSMFE-Jubilee Papers in Honour of Prof. Dr. Ergun Togrol, IstanbulGoogle Scholar
  17. Seed HB, Idriss IM (1971) Simplified procedure for evaluating soil liquefaction potential. J Soil Mech Found Div ASCE 97(SM9):1249–1273Google Scholar
  18. Seed RB, Cetin KO, Moss RES, Kammerer AM, Wu J, Pestana JM, Riemer MF, Sancio RB, Bray JD, Kayen RE, Faris A (2003) Recent advances in soil liquefaction engineering. In: 26th annual, ASCE L. A. geotechnical seminarGoogle Scholar
  19. Vucetic M, Dobry R (1991) Effect of soil plasticity on cyclic response. J Geotech Eng Div ASCE 117(1):89–107CrossRefGoogle Scholar
  20. Wang WS (1979) Some findings in soil liquefaction research. Institute of Water Conservancy and Hydroelectric Power, Beijing, PRCGoogle Scholar

Copyright information

© Springer Science+Business Media B.V., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Department of Civil EngineeringIstanbul Kultur UniversityIstanbulTurkey
  2. 2.Department of Civil and Environmental EngineeringVirginia TechBlacksburgUSA

Personalised recommendations