Influence of Excavation Rate and Tension Crack on the

Stability of an Unsupported Vertical Cut in Unsaturated Soil

by

Geeshpati Yanamandra

B.Tech. (Civil Engineering), Amity University Rajasthan, 2017

A Thesis Submitted in Partial Fulfillment

of the requirements for the Degree of

Master of Science in Engineering

In the Graduate Academic Unit of Civil Engineering

Supervisor: Won Taek Oh, Ph.D., P.Eng., Department of Civil Engineering

Examining Board: Kripa Singh, Ph.D., P.Eng., Department of Civil Engineering

Othman Nasir, Ph.D., P.Eng., Department of Civil Engineering
Karl Butler, Ph.D., P.Geo., P.Eng., Department of Earth
Sciences

This Thesis is accepted by the Dean of Graduate Studies

THE UNIVERSITY OF NEW BRUNSWICK

April 2020

© Geeshpati S.S. Yanamandra, 2020

 ABSTRACT

Unsupported excavation is an important activity in geotechnical engineering practice since

many projects ranging from mining to infrastructure developments are initiated from

unsupported excavation. Unsupported excavation should be carried out with the utmost

caution since the failure of unsupported cuts can result in not only property losses but also

fatalities. This research focuses on the stability of unsupported vertical cuts in unsaturated

soils. For this, a series of numerical analyses is conducted in an unsaturated glacial till

considering three factors: i) excavation rate, ii) depth and location of tension crack, and iii)

rainfall infiltration into a tension crack. The results showed that the influence of excavation

rate is not significant if an unsupported vertical cut is made to a safe height (i.e. critical

height divided by factor of safety (1.2 in this study)). Tension crack is one of the major

factors that can lead to the failure in unsupported vertical cuts; however, the factor of safety

did not drop below unity if the location and depth of tension crack is limited within 20%

and 30% of the safe height from the cut wall and the ground surface, respectively. Rainfall

infiltration into a tension crack decreased the factor of safety with time and then eventually

led to the failure in unsupported vertical cuts for most cases. However, extremely long

duration of rainfall was required if the initial factor of safety with a tension crack is close

to 1.2. The proposed approaches are then applied to a deep unsupported vertical cut (9.75

m) made into a clay for its validation, which successfully estimated the critical location

and depth of a tension crack.

ii
 ACKNOWLEDGEMENTS

Foremost, I would like to express my deep and sincere gratitude to my research supervisor

Dr. Won Taek Oh for giving me the opportunity by accepting me as his research student

and providing invaluable guidance throughout the study. He taught me methods to widen

my horizons on carrying out this research and present the works as clearly as possible. I

am extremely grateful for the freedom he offered me to pursue all possible avenues

diverging from academics. I would also like to thank him for his empathy, patience and all

the great tennis sessions shared with him.

I am extremely grateful to my parents for their love, prayers, and sacrifices for educating

and preparing me for my future. I am thankful to all my friends for making this journey a

lot more pleasant and becoming an extended part of my family in this home away from

home. I extend my special thanks to not just my colleagues but also great friends, Gregory

Brennan, and Mehdi Poormousavian. Greg, I couldn’t thank you more for the patience and

assistance you offered throughout my initial stage of study. It is safe to say that you’ve

successfully taught me the unconventionally effective principal of “fitting a square in a

circle”. Mehdi, I thoroughly enjoyed all the lectures we shared together and the

brainstorming sessions on soil mechanics which frequently ended up making us more

confused than we were.

Finally, I would like to thank the staff of UNB Civil Engineering for always being so

flexible to my needs. I could not have pursued the opportunities I did throughout my degree,

if it wasn’t always for the assistance provided by you.

iii
 TABLE OF CONTENTS

ABSTRACT ....................................................................................................................... ii
ACKNOWLEDGEMENTS ............................................................................................ iii
TABLE OF CONTENTS ................................................................................................ iv
LIST OF TABLES ........................................................................................................... vi
LIST OF FIGURES ........................................................................................................ vii
LIST OF SYMBOLS & ABBREVIATIONS ................................................................ xi
1. GENERAL INTRODUCTION .............................................................................. 1
1.1 Problem Statement ............................................................................................... 1
1.2 Objectives of the Thesis ....................................................................................... 5
1.3 Scope of the Thesis .............................................................................................. 5
1.4 Organization of the Thesis ................................................................................... 6
2. TENSION CRACKS ............................................................................................... 8
2.1 Introduction .......................................................................................................... 8
2.2 Formation and Propagation of Tension Crack ................................................... 11
2.3 Estimating Depth of Tension Crack ................................................................... 15
2.4 Estimating Permeability Function of Tension Crack ......................................... 20
2.4.1 Zhou et al. (1998) ........................................................................................ 21
2.4.2 Hu et al. (2000) ........................................................................................... 22
2.4.3 Ping et al. (2005) ......................................................................................... 23
2.4.4 Zhang and Li (2012) ................................................................................... 24
2.4.5 Zhang et al. (2020) ...................................................................................... 25
2.4.6 Studies Questioning Modelling of Infiltration into Tension Crack ............ 26
3. EFFECTIVE AND TOTAL STRESS APPRAOCHES IN UNSATURATED
SOILS ..................................................................................................................... 27
4. METHODOLOGY OF NUMERICAL MODELING ....................................... 31
4.1 Soil Properties .................................................................................................... 31
4.2 Initial In-Situ Stresses ........................................................................................ 34
4.3 Simulation of Excavation ................................................................................... 35

iv
 4.4 Slope Stability Analysis ..................................................................................... 36
4.5 Simulation of Tension Crack and Rainfall Event............................................... 39
5. ANALYSIS RESULTS AND DISCUSSION ...................................................... 44
5.1 Influence of Excavation Rate on the Stability of UVC ...................................... 44
5.1.1 Estimation of Safe Height (FOS = 1.2)....................................................... 44
5.1.2 Estimation of Critical Height (FOS = 1) ..................................................... 49
5.1.3 Summary and Conclusions ......................................................................... 55
5.2 Determining the Influence of Tension Crack on Stability of Unsupported Vertical
Cut…. ................................................................................................................. 56
5.2.1 Determining the Critical Height of UVC using Numerical Method ........... 56
5.2.2 Estimating a Depth of Tension Crack using Pufahl et al. (1983) approach 62
5.2.3 Determination of Critical Tension Crack .................................................... 66
5.2.4 Summary and Conclusions ......................................................................... 70
5.3 Influence of Rainfall Infiltration into Tension Crack on the Stability of UVC . 71
5.3.1 Methodology ............................................................................................... 71
5.3.2 Stand-Up time of UVC with Tension Crack under Different Rainfall
Intensities .................................................................................................... 76
5.4 Summary and Conclusions ................................................................................. 78
6. CASE STUDY ........................................................................................................ 79
6.1 Description of Project......................................................................................... 79
6.2 Soil Properties .................................................................................................... 82
6.3 Methodology ...................................................................................................... 85
6.4 Summary and Conclusions ................................................................................. 93
7. GENERAL CONCLUSIONS ............................................................................... 94
7.1 Recommendations for Future Research ............................................................. 96
REFERENCES ................................................................................................................ 97
CURRICULUM VITAE

v
 LIST OF TABLES

Table 4.1. Basic soil properties of Indian Head Till (Vanapalli 1996) ............................. 32
Table 5.1. Variation of factor of safety for different combinations of DPtc and DStc with
GWT at 1, 3 and 5 m......................................................................................................... 61
Table 5.2. Stand-up time of UVC with a tension crack under different rainfall events
(Initial FOS of UVC is close to 1.2). ................................................................................ 77
Table 6.1. Material properties used in the numerical analysis (adopted from Kwan 1971
and Banerjee et al. 1988). ................................................................................................. 83

vi
 LIST OF FIGURES

Figure 1.1. Variation of deformation, pore-water pressure, and FOS with time for (a) 10,
(b) 250 (c) 500, and (d) 750 seconds after 1.3m excavation stage in sand with initial
ground water table at 0.7 m (Richard 2018). ...................................................................... 3
Figure 2.1. Semi-infinite cohesive mass with horizontal surface: (a) stresses at boundaries
of prismatic element; (b) graphic representation of state of stress at failure; (c) shear
pattern for active state; (d) shear pattern for passive state; (e) stresses on vertical section
through the mass (Terzaghi 1943). ..................................................................................... 8
Figure 2.2. Monthly rainfall and FOS during the period concerned (Gofar et al. 2006). . 10
Figure 2.3. Schematic of test specimen and loading used to determine critical state energy
release rate (Lee et al. 1988). ............................................................................................ 12
Figure 2.4. Crack patters for UVC with different depth of cut (H) and Poisson’s ratio (ν):
(a) H = 6 m, ν = 0.48, (a) H = 12 m, ν = 0.41, (a) H = 18 m, ν = 0.41 (Lee et al. 1988) .... .
........................................................................................................................................... 14
Figure 2.5. Earth pressure distribution diagram used to calculate a factor of safety of the
temporary vertical cut against general failure using the field measurement data in
Whenham et al. (2007) (after Richard et al. 2020). .......................................................... 20
Figure 2.6. Simulating cracked upper layer as an equivalent weaker layer (Hu et al.
2000). ................................................................................................................................ 22
Figure 2.7. Permeability function of a tension crack used by Hu et al. (2000). ............... 23
Figure 2.8 Permeability functions for a fine-grained soil and a crack used by Zhang and
Li (2012). .......................................................................................................................... 25
Figure 3.1. Variation of load, stress, pore pressure, strength, and factor of safety at point
A due to excavation in saturated clay (modified after Bishop and Bjerrum 1960). ......... 28
Figure 3.2. Drops in pore-water pressure due to excavation measured from piezometer
sensors located near excavation and those estimated using FLAC (Galera et al. 2009). . 29
Figure 4.1. Grain size distribution curve of Indian Head Till (Oh and Vanapalli 2010) .. 32
Figure 4.2. SWCC of Indian Head Till (Oh and Vanapalli 2018). ................................... 33
Figure 4.3. Permeability function of Indian Head Till...................................................... 33

vii
Figure 4.4. Meshes and boundary conditions established using ‘Insitu’ analysis type in
SIGMA/W. ........................................................................................................................ 35
Figure 4.5. Simulating staged excavation and assigning water total head hydraulic
boundary condition in SIGMA/W. ................................................................................... 36
Figure 4.6. Vertical stress contours computed with SIGMA/W (SIGMA/W manual) ..... 37
Figure 4.7. Slope stability analysis in SLOPE/W using ‘Entry and Exit’ surface option. 38
Figure 4.8. Example of analysis tree. ................................................................................ 39
Figure 4.9. Simulating tension crack using tension crack line feature in SLOPE/W. ...... 40
Figure 4.10. Example of stability analysis result using tension crack line feature in
SLOPE/W. ........................................................................................................................ 40
Figure 4.11. Slope stability analysis considering a tension crack. Tension crack was
simulated as a void. ........................................................................................................... 41
Figure 4.12. Simulating rainfall infiltration in SIGMA/W. .............................................. 42
Figure 4.13. Total head versus time relationships used to simulate the infiltration of
rainfall into a tension crack under different rainfall intensities. ....................................... 43
Figure 5.1. Slope stability analyses results with GWT at 1 m for different excavation rate:
(a) 1 second; (b) 4 hours; (c) 12 hours; and (d) 24 hours. ................................................ 45
Figure 5.2. Stability analyses results with GWT at 3 m for different excavation rate: (a) 1
second; (b) 4 hours; (c) 12 hours; and (d) 24 hours. ......................................................... 46
Figure 5.3. Sability analyses results with GWT at 5 m for different excavation rate: (a) 1
second; (b) 4 hours; (c) 12 hours; and (d) 24 hours. ......................................................... 47
Figure 5.4. Variation of safe height with respect to excavation rate for different levels of
ground water table (i.e. 1, 2, 3, 4, and 5 m). ..................................................................... 48
Figure 5.5. FOS of UVC (a) prior and (b) post equilibrium condition with respect to pore-
water pressure (excavation rate = 1 second, ground water table at 1 m). ......................... 50
Figure 5.6. Stability analyses with 1 second excavation rate for different levels of GWT
(1, 3, and 5 m): (a), (b), (c) coupled - SIGMA/W stress method; (d), (e), (f) Bishop’s
simplified method. ............................................................................................................ 52
Figure 5.7 Comparison of critical height estimated using coupled - SIGMA/W stress and
Bishop’s simplified methods at 1 second excavation rate for different levels of GWT. .. 53

viii
Figure 5.8. Variation of the critical height of UVC in sand (Unimin 7030) with respect to
the level of GWT from coupled – SIGMA/W stress (excavation rates = 10 and 10000 s)
and Morgenstern-Price method (limit equilibrium method) (modified after Richard 2018).
........................................................................................................................................... 54
Figure 5.9. Simulating tension crack for the safe height of 2.55 m with distance ratio
(DStc) fixed at 0.1 and six different depth ratios (DPtc): (a) 0; (b) 0.1; (c) 0.2; (d) 0.3; (e)
0.4; and (f) 0.5. .................................................................................................................. 57
Figure 5.10. Simulating tension crack for the safe height of 2.55 m with depth ratio (DPtc)
fixed at 0.5 and five different distance ratios (DStc): (a) 0.1; (b) 0.2; (c) 0.3; (d) 0.4; and
(e) 0.5. ............................................................................................................................... 58
Figure 5.11. Analysis tree for various depth ratio (DPtc) in SIGMA/W to determine
critical depth of tension crack. .......................................................................................... 59
Figure 5.12. Stability analysis considering tension crack with different combinations of
DStc and DPtc: (a) DStc = 0.4, DPtc = 0.4 (GWT at 1 m); (b) DStc = 0.5, DPtc = 0.5 (GWT
at 3 m). .............................................................................................................................. 60
Figure 5.13. Determination of depth of tension crack based on net active earth pressure
distribution in vadose zone (Eq.(2.15)) extending the approach by Pufahl et al. (1983). 62
Figure 5.14. Positive, negative, and net active earth pressure distribution (GWT = 1 m).63
Figure 5.15 Positive, negative, and net active earth pressure distribution (GWT = 3 m).64
Figure 5.16 Positive, negative, and net active earth pressure distribution (GWT = 5 m). 65
Figure 5.17. Contours of FOS for different combinations of DStc and DPtc with GWT at 1
m ....................................................................................................................................... 67
Figure 5.18. Contours of FOS for different combinations of DStc and DPtc with GWT at 3
m. ...................................................................................................................................... 68
Figure 5.19. Contours of FOS for different combinations of DStc and DPtc with GWT at
5 m. ................................................................................................................................... 69
Figure 5.20. Water pressure and water flux vector distribution in a tension crack
associated with rainfall infiltration. .................................................................................. 72
Figure 5.21. Variation of pore-water distribution with time around a tension crack under
25 mm/hr rainfall intensity (DStc = 0.2, DPtc = 0.3, GWT at 5 m). .................................. 73

ix
Figure 5.22. Pore-water pressure distribution and factor of safety for different time step
under 25 mm/hr rainfall intensity with a tension crack (DStc = 0.4, DPtc = 0.1, GWT at 5
m). ..................................................................................................................................... 75
Figure 5.23. Variation of factor of safety with time based on the results in Figure 5.22
and the definition of stand-up time. .................................................................................. 76
Figure 6.1. Soil profile and variation of water content, shear strength parameters and
coefficient of permeability with depth at test site (after Kwan 1971). ............................. 80
Figure 6.2. Timeline of excavation (after Kwan 1970). ................................................... 81
Figure 6.3 Excavation profile and location of piezometers (after Kwan 1970)................ 81
Figure 6.4. Soil-Water Characteristic Curve for materials used in numerical analysis. ... 84
Figure 6.5. Permeability function for materials used in numerical analysis. .................... 84
Figure 6.6. Mesh and boundary conditions used in the numerical analysis (11853 nodes,
4523 elements). ................................................................................................................. 85
Figure 6.7. Comparisons of measured and estimated (coupled analysis) pore-water
pressure contours (0 (i.e. phreatic line), 60, and 120 kPa) 12 days after removal of top
sediment layer. .................................................................................................................. 86
Figure 6.8. Comparison of measured and estimated pore-water pressure contours (0
(phreatic line) and 60 kPa) prior to failure........................................................................ 87
Figure 6.9. Pore-water pressure distribution and FOS prior to failure without tension
crack (coupled - SIGMA/W stress method)...................................................................... 88
Figure 6.10. Pore-water pressure distribution and FOS prior to failure without tension
crack (Bishop’s simplified method). ................................................................................. 89
Figure 6.11. Identified potential slip surface with the first tension crack (after Kwan
1971). ................................................................................................................................ 90
Figure 6.12. Actual slip surfaces with the second tension crack that led to the failure of
UVC (after Kwan 1971).................................................................................................... 90
Figure 6.13. Stability analysis with different locations of tension crack from the vertical
cut face: (a) 4.5 m; (b) 3.45 m; and (c) 2.75 m and (d) 2.43 m (coupled – SIGMA/W
stress method). .................................................................................................................. 92

x
 LIST OF SYMBOLS & ABBREVIATIONS

Abbreviations

FOS = factor of safety

GWT = groundwater table

IHT = Indian Head till

PWP = pore-water pressure

SWCC = soil-water characteristic curve

UVC = unsupported vertical cut

Symbols

A = Skempton pore-pressure parameter

bs = critical degree of joint opening

C = total cohesion (kPa)

cL = crack length

c’ = effective cohesion

cu = undrained shear strength

cu(unsat) = undrained shear strength for unsaturated soil

cu(sat) = undrained shear strength for saturated soil

DPtc = depth ratio for tension crack

DStc = distance ratio for tension crack

E = elastic modulus (kPa)

e = void ratio

Fc = applied force

Fr = total resisting force

Fm = total mobilized shear force

xi
g = gravitational acceleration

Gs = specific gravity of soil particles

Gc = critical state energy release rate

H = height of unsupported vertical cut

Hsafe = safe height of unsupported vertical cut (FOS = 1.2)

Hw = depth of water outside slope, measured above toe

hw = distance from ground surface to water table

Ip = plasticity index

k = coefficient of permeability

K0 = Rankine’s earth pressure coefficient for at-rest conditions

K1c = fracture toughness

Ka = coefficient of active earth pressure

kc = coefficient of permeability of tension crack

ksat = coefficient of permeability for saturated condition

m, n, a = fitting parameters for Fredlund & Xing’s SWCC model (1994)

mvG, nnG, αvG = fitting parameters for van Genuchten’ s SWCC model (1980)

Ncf = stability number for φ > 0

No = stability number for φ = 0

Pa = atmospheric pressure (i.e. 101.3 kPa)

Pc = capillary pressure at rock joint

Pd = driving force term

q = rainfall intensity

r = radial length of the crack

ru = pore-pressure coefficient

S = degree of saturation

xii
Se = effective degree of saturation

Sr = residual degree of saturation

t = thickness of specimen

uc = load point displacement

ua - uw = matric suction

ua = pore-air pressure

uw = pore-water pressure

wc = width of tension crack

v = Poisson’s ratio

z = depth from ground surface

zt = depth of tension crack

γ = unit weight of soil

γsat = saturated unit weight of soil

γw = unit weight of water

αt = ratio of effective cohesion to tensile strength

θ = volumetric water content

θr = residual volumetric water content

θs = volumetric water content for saturated condition

θc = contact angle

λ = specimen compliance

σ n - ua = net normal stress

σn = total normal stress

σa = active earth pressure

σx = stresses in X direction

σy = stresses in Y direction

xiii
(σθθ)max = maximum circumferential tensile stress

σT = interfacial tension

τunsat = shear strength of unsaturated soil for drained condition

τxy = shear strength in XY direction

φ’ = effective internal friction angle

φ = internal friction angle

ϕb = angle describing rate of change between suction and shear strength

ψ = suction

κ = fitting parameter for shear strength of unsaturated soil

μ = kinematic viscosity for water (10-6 m2/s in normal temperature)

μq = correction factor for surcharge

μt = correction factor for tension crack

μw = correction factor for water pressure

ξ, ζ = fitting parameters for undrained strength of unsaturated soil

xiv
1. GENERAL INTRODUCTION

1.1 Problem Statement

Unsupported excavation is an important activity in geotechnical engineering practice since

most projects ranging from mining to infrastructure developments are initiated from

unsupported excavation. Unsupported excavation should be carried out with the utmost

caution because the failure of unsupported cuts can lead directly to work-related injuries

and deaths (Thompson and Tanenbaum 1977, Suruda et al. 1988, White 2008, Bureau of

Labor Statistics 2010). Between 2000 and 2009, an average of 39 fatalities were reported

annually in the U.S. in association with the failure of unsupported cuts (BLS 2010). Due

to this reason, authorities enforce their own regulations for excavation to safeguard workers

from injuries or deaths. For example, Canadian provinces specify the safe height,

maximum slope angles, benching angles, and minimum distance from an unsupported

vertical cut to stockpiling of excavated or backfill materials. Occupational Safety and

Health Administration (OHSA 2019) also revised excavation manual to make excavation

regulations and standards easier to understand.

Safe height of an unsupported vertical cut (hereafter referred to as UVC) is considered a

governing factor for ensuring safe excavation practices. Canadian provinces limit the

maximum allowable height of UVC to 1.2 m – 1.5 m. On the other hand, OHSA (2019)

allows UVC only in stable rock. However, previous studies showed that UVC with more

than several meters can remain stable in case where excavations are made into unsaturated

soils. (Tsidzi 1997, Whenham et al. 2007, De Vita et al. 2008, Stanier and Tarantino 2013,

Richard 2018).

1
Safe height of UVC in an unsaturated soil can be estimated using a critical height

(maximum excavation depth without failure, Factor of Safety = 1) divided by a certain

safety margin (i.e. factor of safety greater than unity). Critical height can be estimated

based on the net active earth pressure distribution with depth considering the influence of

matric suction on either the cohesion (Richard 2018, Ileme 2019) or the coefficient of

active earth pressure (Vahedifard et al. 2015). Various approaches are available in the

literature to estimate the active earth pressure in unsaturated soils (Pufahl et al. 1983,

Leshchinsky and Zhu 2010, Stanier and Tarantino 2013, Zhang et al. 2010, Vahedifard et

al. 2015).

Excavation causes a temporary drop in the phreatic line, which eventually rebounds with

time after the completion of excavation. In other words, stability of an unsupported cut

continuously varies throughout the excavation process. Richard (2018) conducted

numerical analysis to investigate the critical height of UVC in cohesionless soil. The results

showed that the factor of safety (hereafter referred to as FOS) decreases with time due to

the rebound of phreatic line until the pore-water pressure reaches an equilibrium condition.

The magnitude of drop in phreatic line and its rebound time are governed by excavation

rate and permeability function of a soil, respectively (Figure 1.1).

2
 1.30
Factor of Safety

-6
0.0 ≤ 1.00 - 1.10 0.0
-0.2 1.10 - 1.20 -0.3
-0.4 1.20 - 1.30
-2 (a) 10s -0.6
1.30 - 1.40
-0.6
-0.8 -0.9

Depth (m)
1.40 - 1.50

Depth (m)
-1.0
2
-1.2 1.50 - 1.60 -1.2
-1.4 1.60 - 1.70 -1.5
2
6
-1.6 4 1.70 - 1.80
-1.8 -1.8
1.80 - 1.90
-2.0 -2.1
10 -2.2
8 ≥ 1.90
-2.4 -2.4
12
14 -2.6 -2.7
-2.8
16 -3.0
-3.0

1.32
Factor of Safety
0.0 ≤ 1.00 - 1.10 0.0
-4
-0.2 1.10 - 1.20 -0.3
(b) 250s -0.4
-0.6
1.20 - 1.30 -0.6
-0.8
1.30 - 1.40
-0.9

Depth (m)
1.40 - 1.50

Depth (m)
-1.0
-1.2 1.50 - 1.60 -1.2
4 -1.4 4 1.60 - 1.70 -1.5
-1.6 1.70 - 1.80
-1.8 -1.8
1.80 - 1.90

8
8
-2.0 -2.1
-2.2 ≥ 1.90
12
12
-2.4 -2.4
-2.6 16 -2.7
16 -2.8
-3.0 -3.0

1.12
Factor of Safety
0.0 ≤ 1.00 - 1.10 0.0
-6
-0.2
1.10 - 1.20 -0.3
-0.4
-2 (c) 500s -0.6 1.20 - 1.30 -0.6
-0.8 1.30 - 1.40
-0.9

Depth (m)
Depth (m)

2 -1.0 1.40 - 1.50

-1.2 4
1.50 - 1.60 -1.2
6 -1.4 1.60 - 1.70 -1.5
-1.6 8 1.70 - 1.80
-1.8 -1.8
10 1.80 - 1.90
-2.0 12 ≥ 1.90 -2.1
-2.2 -2.4
14
-2.4 16
-2.6 -2.7
18 -2.8 20 -3.0
-3.0

750s

Figure 1.1. Variation of deformation, pore-water pressure, and FOS with time for (a)
10, (b) 250 (c) 500, and (d) 750 seconds after 1.3m excavation stage in sand with initial
ground water table at 0.7 m (Richard 2018).

3
Tension cracks are often found at the crest of slopes and cuts in case the tensile stress

exceeds the tensile strength of a soil (Baker 1981, Bagge 1985). It is well known that

tension cracks have adverse impact on the safe height of UVC due to following reasons: i)

Tension crack usually forms a part of slip surface. This shortens the length of slip surface

and subsequently reduces resistance to slope failure, ii) Additional driving force can be

generated if a tension crack is filled with water, and iii) Tension crack can act as a pathway

for rainfall to seep through the soil and further reduces its shear strength. To reliably

estimate the influence of tension crack on the stability of UVC, three factors should be

considered; i) depth of tension crack, ii) location of tension crack, and iii) penetration of

water into tension crack associated with rainfall events. Several closed-form equations or

approaches are available to estimate the depth and/or location of tension crack (Taylor

1948, Spencer 1968, Bagge 1985, Lee et al. 1988, Kutschke and Vallejo 2011, Baker and

Leshchinsky 2003, Michalowski 2013, Li et al. 2018). However, limited studies have been

undertaken to estimate the safe height and location and depth of tension cracks in UVC in

unsaturated soil. Research on the stability of slopes under rainfall events clearly showed

that rainfall infiltration into tension cracks significantly decreases the factor of safety (Ping

et al. 2005, Wang et al. 2012, Gofar et al. 2006, Sasekaran 2011, Sun et al. 2019, Zhang et

al. 2020). Although these studies successfully addressed the importance of considering

tension cracks under a rainfall event, it is still challenging to simulate the seepage into a

tension crack in numerical analysis.

4
1.2 Objectives of the Thesis

The main objective of this study is to investigate the stability of UVC in unsaturated soil

considering various practical scenarios. More details are as follows:

- Estimate the safe height of unsupported vertical cuts in an unsaturated cohesive soil

(Indian Head till) considering excavation rate.

- Estimate the influence of tension cracks on the stability of unsupported vertical cuts

considering various depths and locations of tension cracks

- Estimate the influence of rainfall infiltration into tension cracks on the stability of

unsupported vertical cuts.

- Validate the adopted methodologies/approaches through a case study (i.e. deep

unsupported cut in a clay, Kwan 1971).

1.3 Scope of the Thesis

This research is carried out for UVC excavated into an unsaturated glacial till (i.e. Indian

Head till). Multiple excavation scenarios were simulated in numerical analysis considering

four factors: i) level of the groundwater table, ii) excavation rate, iii) depth and distance

(i.e. distance from cut wall) of tension crack, and iv) infiltration of rainfall into tension

crack under several different rainfall intensities. It was assumed that UVC was initially

excavated to a safe height with factor of safety = 1.2. Slope stability analyses were

conducted based on the stress from finite element analysis. Geotechnical modelling

software, SIGMA/W and SLOPE/W (GEO-SLOPE International Ltd.) were used to

simulate excavation/rainfall events and to perform slope stability analysis, respectively.

5
1.4 Organization of the Thesis

This thesis consists of seven chapters including ‘General Introduction’ (Chapter 1) and

‘General Conclusions’ (Chapter 7).

Chapter 2 presents literature reviews on the formation and propagation of tension crack,

and existing methodologies/approaches to estimate the depth of tension cracks. Various

techniques to estimate the permeability function of tension cracks in unsaturated soils are

introduced.

Chapter 3 describes the mechanism of drop in phreatic line and the variation of factor of

safety of unsupported vertical cut (UVC) with time due to excavation. Two different

stability analysis methods; effective and total stress approaches are explained taking

account of the change in drainage condition of pore-water pressure during excavation

process.

Chapter 4 presents the numerical modelling technique adopted in this research along with

the soil properties. Details include the methodologies to i) simulate excavation and

subsequent change in phreatic line, tension crack and rainfall infiltration into tension crack,

and to ii) analyze slope stability. Descriptions of the numerical modeling software,

SIGMA/W and SLOPE/W used in this study are presented.

In Chapter 5, the influence of excavation rate on the stability of UVC is investigated. In

addition, the critical combinations of tension crack depth and location are determined for

different levels of ground water table. Lastly, the variation of factor of safety of UVC under

different rainfall intensities are studied.

6
Chapter 6 revisits an instrumented large-scale UVC failure case that took place in Welland,

Ontario (Kwan 1971). This case study is used to validate the methodologies and approaches

adopted in this research.

7
2. TENSION CRACKS

2.1 Introduction

Tension cracks are found near the crest of slopes and unsupported cuts when tensile stress

exceeds tensile strength (Bagge 1985). Figure 2.1 illustrates the state of stresses in a

cohesive soil. The diameter of circle Ct in Figure 2.1(b) represents the tensile strength of

soil at the surface. Desiccation, differential settlement and temperature changes can also

be the main reasons for the formation of tension cracks (Li and Zhang, 2018).

Figure 2.1. Semi-infinite cohesive mass with horizontal surface: (a) stresses at
boundaries of prismatic element; (b) graphic representation of state of stress at
failure; (c) shear pattern for active state; (d) shear pattern for passive state; (e)
stresses on vertical section through the mass (Terzaghi 1943).

8
Gofar et al. (2006) performed numerical analysis on an open coal mine landslide caused

by rainfall infiltration through surface fissures in Airlaya, Indonesia using GeoStudio (i.e.

VADOSE/W and SLOPE/W). The slope consisted of three layers: Layer 1 was the dumped

mine material which exhibited high swelling and shrinkage characteristics under wet and

dry conditions, respectively; Layer 2 was a thin high organic material between fill material

and natural soil, which was attributed to lack of clearance before dumping; Layer 3 was

the natural clay stone material. A crack of 40 m deep existed, passing thorough Layer 2.

The analyses results showed that if the tension crack was neglected, only limited infiltration

took place through the layers with no remarkable rise in the ground water table (hereafter

referred to as GWT). Figure 2.2 shows the monthly rainfall data and the variation of factor

of safety (hereafter referred to as FOS) from the slope stability analyses. It was noticed that

FOS became lowest on two occasions, June 21, 2001 and November 27, 2002. However,

the failure only took place on the latter, which was attributed to the development of tension

crack shortly after the end of dry season. This tension crack has allowed the water to seep

into Layer 2, and it started flowing horizontally within the layer while generating a weak

plane, which led to failure eventually. The numerical analyses performed with a tension

crack successfully captured this failure condition.

9
 lowest FOS

Figure 2.2. Monthly rainfall and FOS during the period concerned (Gofar et al. 2006).

Sasekaran et al. (2011) conducted slope stability analysis on a homogenous silty clay slope,

taking account of tension crack location and depth and rainfall intensity. The results

showed that, with increase in the depth of tension crack, pore-water pressure around its

vicinity increases as water infiltrates into a deeper section of the soil, further raising the

GWT. This led to a decrease in matric suction and drop in FOS. However, the tension crack

did not have any significant impact on stability of slope when it was located at more than

30 m from the crest of the slope. When rainfall was simulated it was noted that, for the

same duration, the higher intensity rainfall has the greater influence towards slope failure.

Similar research was also carried out by Zhang and Li (2012) considering the same factors

as Sasekaran et al. (2011). The lowest FOS was observed with a tension crack located on

the crest rather than on the slope since the slip surface initiated from the tension crack when

it is on the crest. The FOS decreased with increasing the depth of tension crack. It was also

observed that light rainfall for long duration was more detrimental than a storm like rainfall

10
for short duration. This is because rainfall with longer duration will have extended seepage

period, which allows water to further infiltrate into the soil.

These examples clearly show that tension crack is one of the governing factors that affect

the stability of slopes. Tension cracks become detrimental during rainfall as they act as

pathway for rainfall to seep through the soil. As rainfall infiltrated through the tension

cracks, the pore-water pressure around the cracks increase, which leads to decrease in

matric suction and shear strength of the soil (Hu 2000). Typically, pre-existing cracks are

highly influential towards the deep failures (Hu 2000, Ping et al. 2005, Wang et al. 2010).

In practice, however, this aspect is often overlooked in the design of unsupported cuts.

This chapter discusses the formation mechanism and methodologies to estimate the depth

of a tension crack. In addition, techniques to estimate the permeability function of tension

cracks in numerical seepage analysis are revisited.

2.2 Formation and Propagation of Tension Crack

Lee et al. (1988) investigated the initiation and propagation of tension cracks in an over-

consolidated marine clay specimen. For this, the concept of critical state energy release

rate (Gc) was adopted, which was determined using an experimental setup as shown in

Figure 2.3.

11
 Fc Fc

uc
cL

Figure 2.3. Schematic of test specimen and loading used to determine critical state
energy release rate (Lee et al. 1988).

A slot of 0.5 mm was cut in an over-consolidated clay specimen, and a force Fc was applied

by drawing the pins apart. Movement of the pins were then read using a dial gauge and

exerted tension on the specimen was measured. Crack propagation from tip of the crack

was then determined at various stages of applied force using a microscope positioned

directly above the specimen.

Gc is defined as the difference between rate of work done by applied loading and rate of

increase in strain energy in a specimen (Eq.(2.1))

1 2 dλ
Gc = ( Fc ) (2.1)
2t dcL

where, Fc is applied force, t is thickness of specimen, dλ is change in specimen compliance,

and dcL is change in crack length

12
The specimen compliance, λ can be calculated using Eq. (2.2)

uc
λ= (2.2)
Fc

where, uc is load point displacement

The plot of Gc against crack length, cL remained fairly constant, which justifies that Gc can

be adopted as a material constant. The crack was then assumed to propagate from the tip

of an existing crack in direction of normal to maximum circumferential tensile stress,

(σθθ)max (Eq. (2.3)).

(σθθ )max 2π r = K1c (2.3)

where, r is radial length of crack, and K1c is fracture toughness (Erdogan and Sih 1963)

Fracture toughness, K1c can be corelated to Gc by following Irwin (1958)’s expression for

plane stress condition (Eq. (2.4)).

K12c
Gc = (1 − v 2 ) (2.4)
E

where, E is elastic modulus, and v is Poisson’s ratio

K1c in Eq. (2.4) can be estimated based on Gc, E and ν determined through laboratory test.

Eq. (2.3) then can be used to estimate the direction of crack propagation.

Lee et al. (1988) extended this approach to numerical modelling to study the influence of

Poisson’s ratio and depth of cut on crack patterns in UVC. It turned out that lateral extent

13
of the crack zone is a function of ν (i.e. the higher ν the greater extent of crack zone),

while an increase in the depth of cut increased both depth and lateral extent of cracking

(Figure 2.4). However, no attempt was made to determine depth and location of crack that

are critical to the failure of UVC.

Figure 2.4. Crack patters for UVC with different depth of cut (H) and Poisson’s ratio
(ν): (a) H = 6 m, ν = 0.48, (a) H = 12 m, ν = 0.41, (a) H = 18 m, ν = 0.41 (Lee et al.
1988).

14
Kutschke and Vallejo (2011) performed finite element analysis to study stability of UVC

in over-consolidated clays considering tension cracks. Analyses were performed on a

homogenous domain with elastic-plastic model. Excavation was simulated by removing

elements from the in-situ condition and the slope was allowed to achieve the equilibrium

condition in terms of stress. During excavation, horizontal movement was observed along

the slope face due to stress relief, which led to the formation of tension cracks. It was

inferred that tension cracks were formed when lateral stress induced due to stress relief

exceeds tensile strength of the soil. This indicates that the formation and propagation of

tension cracks are highly dependent on in-situ earth pressure coefficient at-rest (i.e. K0).

To study the influence of tension crack on failure plane, a crack with zero tensile strength

(or stiffness) was simulated on the crest of UVC. The analysis results with various in-situ

K0 values showed that a failure in an UVC takes place when the depth of tension crack

ranges between 0.45 and 0.57 times the height of UVC.

2.3 Estimating Depth of Tension Crack

Maximum expected depth of tension crack in UVC can be estimated extending either

effective stress approach or total stress approach. Most traditional equations to estimate the

depth of a tension crack, zt were proposed by Taylor (1948) for drained (Eq.(2.5)) and

undrained (Eq.(2.6)) conditions.

2c ′  φ′ 
zt = tan  45° +  (2.5)
γ  2

2cu
zt = (2.6)
γ

15
where, zt is depth of tension crack, c’ is effective cohesion, φ’ is effective internal friction

angle, cu is undrained shear strength, and γ is unit weight of soil

Taylor (1948)’s equations are applicable only to UVC and do not consider the influence of

pore-water pressure. Janbu (1968) developed charts that can be used to estimate the critical

height of UVC. The FOS can be calculated using Eq. (2.7) and Eq. (2.8) extending total

and effective stress approaches, respectively taking account of surcharge, submergence,

seepage, and depth of tension crack. According to Janbu (1968), critical slip surface passes

through the toe for both drained and undrained conditions.

cu
FOS = N 0 ( for φ = 0) (2.7)
Pd

c′
FOS = N cf ( for φ > 0) (2.8)
Pd

where No and Ncf are stability numbers , and Pd is driving force term (Eq. (2.9))

γ H + q − γ wH w
Pd = (2.9)
μ q μ w μt

where, γw is unit weight of water, H is height of UVC, μq, μw, μt are correction factors for

surcharge, water pressure, and tension crack, respectively, and Hw is depth of water outside

slope, measured above toe.

Spencer (1973) used Eq. (2.10) to estimate the depth of tension crack in embankment

extending the effective stress approach. The stability analysis was carried out based on zero

16
lateral effective stress considering geometry of slope and influence of pore-water pressure

(i.e. pore-pressure coefficient, ru).

2 c′  φ′ 
zt = tan  45 +  (2.10)
γ (1 − ru )  2

The approach proposed by Baker (1981) can be used to determine location and depth of

tension crack in UVC for both drained and undrained conditions. According to Baker

(1981), maximum depth of tension crack takes place in UVC and it never exceeds one

quarter of the slope height. Bagge (1985) suggested that maximum depth of tension crack

in UVC can be estimated considering the change in pore-water pressure due to stress relief

during vertical cutting. Eqs (2.11) and (2.12) can be used to estimate the depth of tension

crack for undrained and drained conditions, respectively. If a water table is at the ground

surface and the tensile strength of soil is neglected, Eq. (2.11) reduces to Eq. (2.6).

2 Acu − γ whw  K0 + A (1 − K0 )  − αt c′
zt = (2.11)
 K0 + A (1 − K0 )  (γ − γ w )

 cos φ ′ − 12 αt (1 + sin φ ′) 
2 c′ 
1 − sin φ ′  − γ whw
zt =   (2.12)
γ −γw

where A is Skempton pore-pressure parameter, αt is ratio of effective cohesion to tensile

strength, and hw is distance from ground surface to water table

These conventional analytical solutions can be more effectively used for saturated

homogeneous soils. However, in reality, shear strength of soil is governed by the location

of water table and varies with respect to matric suction. Pufahl et al. (1983) investigated
17
lateral earth pressure in UVC extending the unsaturated soil mechanics considering the

influence of matric suction on the shear strength of soil. Based on the conventional Rankin

earth pressure theory, the active earth pressure, σa, for unsaturated condition can be written

as Eq. (2.13) assuming air-pressure is atmospheric pressure.

σ a = γ zK a − 2  c′ + ( ua − uw ) tan φ b  K a (2.13)

where z is depth from the ground surface, (ua – uw) is matric suction, ua is pore-air pressure,

uw is pore-water pressure, and φb is angle indicating the rate of increase in shear strength

with respect to a change in matric suction

Eq. (2.13) can be rewritten as Eq. (2.14) considering the nonlinear variation of unit weight

and shear strength of unsaturated soil with depth (Vanapalli et al. 1996).

 Gs + θ (1 + e) 
σa =   γ w zKa − 2 c′ + ( ua − uw ) S κ tan φ ′ Ka (2.14)
 (1 + e) 

where S is degree of saturation, κ is fitting parameter (function of plasticity index), Gs is

specific gravity, θ is volumetric water content, and e is void ratio

Hence, the depth of tension crack is obtained for σa = 0 as shown in Eq. (2.15)

2  c′ + ( ua − uw ) S κ tan φ ′
zt = (2.15)
 Gs + θ (1 + e ) 
 (1 + e ) γ w  K a
 

More recently, Baker and Leshchinsky (2003) studied the spatial distribution of safety

factors in a cohesive UVC by utilizing the safety map notion originally proposed by Baker

18
and Leshchinsky (2001). Michalowski (2013) proposed a method that can be used to

estimate the maximum depth of tension crack based on limit analysis. The location of

tension crack was determined as the one with the most adverse influence on the stability.

A closed-form solution proposed by Li et al. (2018) can also be used to determine depth of

tension crack, which was based on limit equilibrium method while taking account of

linearly increasing undrained strength. Numerical analyses were also carried out to study

the influence of tension crack on the stability of UVC (Lee et al. 1998; Kutschke and

Vallejo 2011). The main advantage of numerical analysis is that progressive failure

mechanism can be taken into account during analysis.

In the present study, Eq. (2.15) was used to estimate the maximum depth of tension crack

analytically. Richard et al. (2020) used Eq. (2.15) to estimate critical height of an

instrumented temporary large scale UVC (3 m deep, 6 m wide and 20 m long) in an

unsaturated soil (Whenham et al. 2007) taking account of rainfall infiltration (Figure 2.5).

The factor of safety at the moment of general failure matched the one obtained extending

the concept in Eq. (2.15). However, it should be noted that Eq. (2.15) can only be used to

determine theoretical maximum depth of a tension crack, not the location.

19
 Positive earth pressure Negative earth pressure
(σvKa) {-2[c'+(ua - uw)Sκtanφ'] Ka}

1.5m γsat =
20.2 kN/m3

P3
3m
γunsat = P1
1m 18.8 kN/m3 1m
+
(ua - uw ) = 8.5 kPa P2 P4
p4
at 2.5 m
0.5m 0.5m
(ua - uw ) = 18 kPa
at 3.0 m p3
6m p1 p2 p5

p3 = 2c' Ka
p1 = γsat x 1.5 m x Ka
p4 = {[(ua - uw)Sκtanφ'] Ka} at 2.5 m
p2 = γunsat x 1.5 m x Ka
P5 = {[(ua - uw)Sκtanφ'] Ka} at 3.0 m

P1, P2, P3, P4: active thrust

Figure 2.5. Earth pressure distribution diagram used to calculate a factor of safety of
the temporary vertical cut against general failure using the field measurement data
in Whenham et al. (2007) (after Richard et al. 2020).

2.4 Estimating Permeability Function of Tension Crack

As mentioned earlier, tension crack acts as a pathway for rainfall to seep through the soil,

which results in a significant decrease in shear strength of unsaturated soil. Hence, it is

important to estimate permeability of tension crack in case the stability of UVC is carried

out through numerical analysis considering rainfall events. Seepage of fluids such as water,

gasoline or oil through unsaturated soil is far more complicated than single-phase saturated

percolation. This is because the coefficient of permeability of an unsaturated soil is

governed not only by the properties of fluids and pore-size distribution of a soil, but also

by the degree of saturation (or matric suction). This indicates that permeability of a tension

crack in an unsaturated soil is also governed by both the width of tension crack and matric

suction. Following subsections summarize the methodologies or techniques used to

20
estimate the permeability function (i.e. variation of coefficient of permeability with respect

to matric suction) of tension cracks in unsaturated soils.

2.4.1 Zhou et al. (1998)

Zhou et al. (1998) introduced an analytical model to determine the permeability function

of openings in rock joints. For this, van Genuchten’s (1980) Soil-Water Characteristic

Curve (SWCC) model was adopted to establish the relationship between capillary pressure

and effective degree of saturation of a rock joint, shown in Eq. (2.16)

1
Se = mvG (2.16)
1 + (α vG Pc ) nvG 
 

where, Se is effective degree of saturation (Eq.(2.17)), Pc (Eq.(2.18)) is capillary pressure

at rock joint, and αvG, mvG, nvG are fitting parameters

S − Sr
Se = (2.17)
1 − Sr

where, S is degree of saturation and Sr is residual degree of saturation

2σ T cos θ c
Pc = (2.18)
bs

Where, σT is interfacial tension, θc is contact angle, and bs is critical opening at which joints

begin to excrete under pressure, Pc

Based on the fitting parameters obtained from Eq. (2.16), the permeability function of a

rock joint, k(Se) can then be estimated as shown in Eq. (2.19).

21
 k ( Se ) = ksat Se1/2 [1 − (1 − Se1/mvG )mvG ]2 (2.19)

where, ksat is coefficient of permeability of rock joint for saturation condition

2.4.2 Hu et al. (2000)

In the study by Hu et al. (2000), the upper cracked soil was modeled as a special type of

soil with relatively high permeability and low strength. The thickness of this equivalent

‘weaker’ layer was approximated as the mean crack depth (Figure 2.6).

Figure 2.6. Simulating cracked upper layer as an equivalent weaker layer (Hu et al.
2000).

Strength of the weaker soil was estimated by adopting strength reduction factor that is a

function of crack spacing. The permeability of crack was assumed to be two magnitudes

higher than uncracked soil (Figure 2.7). The SWCC of the crack was estimated extending

the approach by Zhou et al. (1998).

22
 Figure 2.7. Permeability function of a tension crack used by Hu et al. (2000).

2.4.3 Ping et al. (2005)

Ping et al. (2005) conducted numerical analysis to simulate rainfall infiltration into a slope

considering the depth, width, and location of tension cracks. The coefficient of

permeability of a tension crack was estimated using Eq. (2.20) based on research on the

seepage into fractured rock by Wang and Su (2002).

gwc2
kc = (2.20)
12 μ

where, kc is coefficient of permeability of tension crack, g is gravitational acceleration, wc

is width of tension crack, and μ is kinematic viscosity for water (10-6 m2/s in normal

temperature)

Analyses results suggested that, as the depth of tension crack increases, the profile of

infiltration becomes prominent near the lower boundary of the model. For the analysis with

23
a significantly narrow tension crack (0.29 mm), the influence of infiltration through the

tension crack on the increment of pore-water pressure was negligible. Whereas, the tension

crack widths of 0.63 mm and 1.35mm were wide enough for water to reach the bottom of

tension crack. The infiltration of rainfall into a tension crack increased with increasing

rainfall intensity.

2.4.4 Zhang and Li (2012)

Zhang and Li (2012) conducted numerical analyses using SLOPE/W and SEEP/W (product

of GeoStudio) to study the influence of tension crack on stability of slope taking account

of various crack characteristics and conditions. Wang (2011) proposed a methodology to

estimate SWCC and permeability function of a tension crack by analyzing random aperture

distribution of cracks. Zhang and Li (2012) used the methodology proposed by Wang (2011)

to estimate the SWCC and permeability function, assuming a tension crack as a material

with a distinct SWCC and permeability function rather than as a boundary condition.

Figure 2.8 shows the permeability function used by Zhang and Li (2012) as an example.

24
 10-1

Coefficient of permeability, k (m/s)

10-2 Fine-grained soil
Crack
10-3
10-4
10-5
10-6
10-7
10-8
10-9
10-10
10-11
0.01 0.1 1 10 100 1000

Suction, ψ (kPa)

Figure 2.8. Permeability functions for a fine-grained soil and a crack used by Zhang
and Li (2012).

2.4.5 Zhang et al. (2020)

Zhang et al. (2020) investigated the stability of shallow slope considering tension crack

extension state, which is dependent on soil type and water content. According to Valentin

et al. (2005), the water-tension causes a compaction effect on soil and the void ratio

decreases accordingly. Cracks are then developed as tension exceeds the soil bearing

capacity. Once the initial cracks are opened, depth and width of cracks continuously

increases due to the repetition of dry-wet cycles. The effect of soil cracking state on the

permeability of soil was considered by having different ksat/q ratios (where q is rainfall

intensity).

25
2.4.6 Studies Questioning Modelling of Infiltration into Tension Crack

Sun et al. (2019) assumed that, in unsaturated soil slope stability analysis, water infiltration

quickly fills cracks; however, no seepage takes place into the deeper soils. On the other

hand, Deng and Shen (2006) and Li et al. (2018) suggested that water infiltration into a

tension crack can be neglected since tension cracks formed in clays can close up upon

wetting due to expansion.

26
3. EFFECTIVE AND TOTAL STRESS APPRAOCHES IN
UNSATURATED SOILS

During excavation, an overall mean total stress at a local point decreases, which leads to a

decrease in pore-water pressure. This stress relief also continuously increases the applied

shear stress and becomes maximum at the end of excavation. However, if undrained

condition is maintained during excavation process the shear strength remains constant until

the completion of excavation (i.e. undrained shear strength). Following the completion of

excavation, water flows towards the excavation and pore-water pressure increases until

equilibrium condition is achieved. Due to this reason, the shear strength starts decreasing

with time and the fine-grained soil swells. This phenomenon is well presented by Bishop

and Bjerrum (1960), showing the variation of load, applied shear stress, pore-water

pressure, shear strength, and factor of safety with time due to excavation (Figure 3.1). The

short-term analysis (i.e. φ = 0 analysis) is required to analyze the stability of excavated

slope until the completion of excavation, and the long-term stability analysis thereafter.

27
 .
A
Excavation

Excavation
Load
Stress
At point A

Pressure
Pore
Strength
Factor

Safety
of

Time

Figure 3.1. Variation of load, stress, pore pressure, strength, and factor of safety at
point A due to excavation in saturated clay (modified after Bishop and Bjerrum 1960).

Galera et al. (2009) investigated a drop in pore-water pressure due to an open pit excavation

in Cobre Las Cruces mine. The mine consisted of 150 m overlying marls which exhibited

properties of over-consolidated clay on the copper ore. The water table was 30 m below

the ground level. The coefficient of permeability of the marls was in the range of 10-9 to

10-10 m/s, making them almost impermeable in nature. Multiple piezometers were installed

at mine pit excavation to read the pore-pressure profile throughout the excavation. A

significant pore-water pressure drops up to -864 kPa was observed on the piezometers

28
installed near the excavation (Figure 3.2). Whereas, for the piezometers installed farther

from the excavation, the range of pore-water pressure drop was between -6.2 kPa and -258

kPa. This example clearly shows that a drop in pore-water pressure continuously takes

place throughout the excavation process. Good agreement was observed between the

measured pore-water pressures and those estimated through coupled hydro-mechanical

analysis using FLAC.

Figure 3.2. Drops in pore-water pressure due to excavation measured from

piezometer sensors located near excavation and those estimated using FLAC (Galera
et al. 2009).

As explained using Figure 3.1, total stress approach is required to analyze the stability of

UVC before the redistribution of pore-water pressure initiates. The duration of undrained

condition during excavation process varies depending on soil type (National Bureau of

Standards 1988, Irvine and Smith 1983, Leroueil et al. 1990). However, Banerjee et al.

(1988) concluded that dissipation of the excess negative pore-water pressure starts

29
immediately after excavation, which justifies the use of effective stress approach (Lambe

and Turner 1970, Kwan 1971, DiBagio and Roti 1972, Dysli and Fontana 1982) in

analyzing the stability of UVC.

In unsaturated soil, the variation of shear strength with respect to matric suction can be

estimated using either Eq. (3.1) (Vanapalli et al. 1996) or Eq. (3.2) (Oh and Vanapalli 2018)

extending the effective stress or total stress approach, respectively. GeoStudio (2019 R2)

adopts Eq. (3.1) and allows users to conduct drained stability analysis (i.e. effective stress

approach). However, in case of undrained stability analysis, the undrained shear strength

values need to be manually assigned to the elements based on the redistribution of pore-

water pressure due to excavation (Oh and Vanapalli 2018). In this study, stability analyses

were performed assuming drained condition using Eq. (3.1).

 θ − θr 
τ unsat = c ′ + (σ n − u a ) tan φ ′ + ( u a − u w )   tan φ ′ (3.1)
 θs − θr 

 ( ua − uw ) Pa  ( S )ζ 
cu ( unsat ) = cu ( sat ) 1 +  
 (3.2)

ξ 

where τunsat is shear strength of unsaturated soil under drained condition, (σn – ua) is net

normal stress, σn is normal stress, θ is volumetric water content (subscript s is saturated

condition and r is residual condition), cu(unsat) is shear strength of unsaturated soil under

undrained condition, cu(sat) is shear strength of saturated soil under undrained condition, ζ

and ξ is fitting parameters, and Pa is atmospheric pressure (i.e. 101.3 kPa)

30
4. METHODOLOGY OF NUMERICAL MODELING

SIGMA/W and SLOPE/W (product of GeoStudio 2019 R2) were jointly used to simulate

excavation, tension crack, rainfall and to conduct stability analysis of UVC, respectively.

4.1 Soil Properties

In this thesis, it was assumed that the UVC was excavated into a well compacted glacial

till obtained from Indian Head, Saskatchewan, Canada (i.e. Indian Head till, IHT). Basic

soil properties of the soil are summarized in Table 4.1. The grain size distribution curve

and the SWCC are shown in Figure 4.1 and Figure 4.2, respectively. The Fredlund and

Xing's (1994) model used to achieve the best-fit curve of the SWCC and the fitting

parameters are included in Figure 4.2 as insets. Figure 4.3. shows the permeability function

of IHT estimated using the Fredlund and Xing’s (1994) model using the coefficient of

permeability for saturated condition and the SWCC.

31
 Table 4.1. Basic soil properties of Indian Head Till (Vanapalli 1996).

Properties Value
Plasticity Index, Ip 15.5
Saturated unit weight, γsat (kN/m3) 20.7
Saturated water content, θs (%) 47
Void ratio, e 0.55
Specific gravity, Gs 2.72
Effective cohesion, c’ (kPa) 5

Effective internal friction angle, φ’ (˚) 23.1

Saturated coefficient of permeability, ksat (m/s) 1 × 10-7

Elastic modulus, E (kPa) 10,000
Poisson’s ratio, ν 0.33

100

80
Percent passing

60

40

20

0
0.001 0.01 0.1 1 10
Grain size (mm)

Figure 4.1. Grain size distribution curve of Indian Head Till (Oh and Vanapalli 2010).

32
 1.0

0.8
Degree of saturation

0.6 Fredlund and Xing (1994)

Measured (Pressure plate)

m
0.4  
 1 
Se =  n 
ln
   e + (ψ a ) 
 
a = 11.687
0.2
m = 0.58066
n = 1.4073

0.0
0.1 1 10 100 1000
Suction (kPa)

Figure 4.2. SWCC of Indian Head Till (Oh and Vanapalli 2018).

10-7
Coefficent of permeability, k (m/s)

10-8

10-9

10-10

10-11

10-12

10-13

10-14
0.01 0.1 1 10 100 1000
Suction (kPa)

Figure 4.3. Permeability function of Indian Head Till.

33
4.2 Initial In-Situ Stresses

The initial stress condition was established using ‘Insitu’ analysis type by applying body

load to the elements. The size of domain was 10 m × 10 m. Figure 4.4 shows the defined

mesh and boundary conditions used for the domain. Initial pore-water pressures were

specified by drawing an initial water table, which distributes hydrostatic positive and

negative pore-water pressures below and above the water table, respectively. Stress/strain

boundary conditions were assumed to be restrained in horizontal (X) direction at the

vertical ends (i.e. fixed-X displacement boundaries; hollow red triangles) and restrained in

both horizontal (X) and vertical (Y) directions at the bottom (i.e. fixed-XY boundaries

along the base of the domain). Fine meshes were used in the vicinity of the excavation to

obtain accurate and reliable results, which were generated using quadrilateral and

triangular mesh pattern. Four and three-point integration order was used for quadrilateral

elements and triangular elements, respectively, with a linear interpolation model for

calculating stresses.

34
 10

8
GWT
7

Elevation (m)
6

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

Figure 4.4. Meshes and boundary conditions established using ‘Insitu’ analysis type
in SIGMA/W.

4.3 Simulation of Excavation

Staged excavation was performed by deactivating regions in 0.15 m increments for various

excavation rates in SIGMA/W (Figure 4.5) using ‘Coupled Stress/PWP’ analysis type to

consider the variation of GWT due to excavation. In this study, excavation rate defines the

intervals between excavations; for example, 5 min excavation rate indicates that every 5

min, 0.15 m thickness of soils are removed. During excavation, water total head hydraulic

boundary condition which was equal to the elevation of initial water table, was assigned

along the lateral extents of the soil region on the right side (i.e. solid circles). This boundary

condition was used to maintain a constant hydraulic total head along the right edge of the

domain regardless of the fluctuation of water table due to excavation.

35
 10

8
excavation
7
Elevation (m)

5 Hydraulic boundary

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

Figure 4.5. Simulating staged excavation and assigning water total head hydraulic
boundary condition in SIGMA/W.

4.4 Slope Stability Analysis

After each excavation, slope stability analyses were conducted using SLOPE/W based on

finite element-computed stresses, which are then imported into a conventional limit

equilibrium analysis (i.e. ‘SIGMA/W stress’ analysis type). In other words, the stress

conditions and redistribution of pore-water pressure from finite element analysis in

SIGMA/W were used as the parent analysis in the stability analysis. Figure 4.6 illustrates

the vertical stress contours computed with SIGMA/W. As can be seen, the 50 kPa vertical

stress contour is not a constant distance from the slope surface, but instead, closer to the

slope surface in the vicinity of the toe. This is because the vertical stress is affected by both

36
overburden weight and shear stress. In ‘SIGMA/W stress’ analysis type, σx, σy, and τxy at

the mid-point of base for each slice is first computed. This information is then used to

determine the mobilized shear stress and available shear strength along the base of each

slice. By integrating the mobilized shear stress and shear strength over the length of slip

surface, FOS can be estimated using Eq. (4.1)

FOS =
F r
(4.1)
F m

where, Fr is total resisting shear force and Fm is total mobilized shear force

Figure 4.6. Vertical stress contours computed with SIGMA/W (SIGMA/W manual).

‘Entry and Exit’ surface option was used to define the range of entry and the exit point of

potential slip surfaces in SLOPE/W (Figure 4.7). It was assumed that the slip surfaces pass

through the toe of UVC (i.e. exit point) based on the existing studies (Janbu 1968, Dunlop

and Duncan 1970, Kutschke and Vallejo 2011). This procedure was repeated until the depth

of which FOS = 1.2 is achieved. This depth was denoted as safe height.

37
 10

9 Range of entry

8
Exit point
7
Elevation (m) 6

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

Figure 4.7. Slope stability analysis in SLOPE/W using ‘Entry and Exit’ surface option.

For soil with negative pore-water pressure (i.e. matric suction), total cohesion is computed

with Eq. (4.2) using effective cohesion and the SWCC. The residual volumetric water

content was taken as 5% of the volumetric water content at saturation for calculating total

cohesion (C) in SLOPE/W. The Mohr-Coulomb failure criterion was used as the material

model.

 θ − θr 
C = c′ + ( ua − u w )   tan φ ′ (4.2)
θ
 s − θ r 

Figure 4.8 shows an example of analysis tress, including establishing in-situ condition,

simulating excavation, and slope stability analysis. The slope stability analysis method

38
used in the study is denoted as ‘coupled - SIGMA/W stress’ to distinguish to from the

conventional limit equilibrium method.

Coupled Stress/PWP analysis

SIGMA/W Stress stability analysis

Defined excavation rate

Figure 4.8. Example of analysis tree.

4.5 Simulation of Tension Crack and Rainfall Event

After a safe height (FOS =1.2) is achieved, the influence of tension crack on the stability

of UVC is studied. In SLOPE/W, tension cracks can be defined using ‘tension crack line’

feature, which specifies constant crack depth along the surface of a soil (Figure 4.9). Each

potential slip surface extends up vertically as the slip surface meets tension crack line. The

location of tension crack that has the most adverse impact on the stability of UVC is then

39
determined based on the minimum FOS. Figure 4.10 shows an example of stability analysis

using tension crack line feature.

Figure 4.9. Simulating tension crack using tension crack line feature in SLOPE/W.

0.82
10

8
Tension crack line
7
Elevation (m)

4 Water Pressure
-30 - -20 kPa
-20 - -10 kPa
3 -10 - 0 kPa
0 - 10 kPa
10 - 20 kPa
2 20 - 30 kPa
30 - 40 kPa
40 - 50 kPa
1 50 - 60 kPa
60 - 70 kPa

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)
Figure 4.10. Example of stability analysis result using tension crack line feature in
SLOPE/W.

40
However, ‘tension crack line’ feature is not useful when a user wants to estimate FOS of

UVC considering a specific location and depth of a tension crack. Moreover, simulating

seepage or infiltration through crack is not possible using this feature. To overcome this

disadvantage, in this research, tension crack was simulated as void (0.1 m opening) in

SIGMA/W as shown in Figure 4.11. In case where a tension crack is included in the

analysis, the bottom of tension crack and the toe of UVC were specified as entry range and

exit point, respectively. In other words, it was assumed that failure is initiated from the

bottom of a tension crack. Various locations and depths of tension cracks were considered

in the analyses to define the most critical combination of tension crack location and depth.

Tension crack

Entry range

Exit point

Figure 4.11. Slope stability analysis considering a tension crack. Tension crack was
simulated as a void.

41
Seepage analyses were carried out to investigate the influence of rainfall infiltration into a

tension crack on the stability of UVC. For this, unit flux boundary (hollow triangles in

Figure 4.12) conditions were applied to the ground surface and the bottom of the excavation.

To simulate seepage through tension crack, water total head (hollow circles) boundary

condition was assigned to the crack’s geometry. It was assumed that tension cracks are

gradually filled with water at the same rate as rainfall intensity and the water level in the

tension crack is maintained at the ground level once it is filled with water to consider worst

case scenario. Example of total head increments with time for different rainfall intensities

are shown in Figure 4.13.

Unit flux

10

8
Unit flux

7
Elevation (m)

5
Water total head
4 (function of time)

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

Figure 4.12. Simulating rainfall infiltration in SIGMA/W.

42
 10.2

10.0

9.8
Total head (m)

9.6

9.4

9.2 10mm/hr
15mm/hr
25mm/hr
9.0

0 20 40 60 80 100 120 140 160

Time (hours)

Figure 4.13. Total head versus time relationships used to simulate the infiltration of
rainfall into a tension crack under different rainfall intensities.

43
5. ANALYSIS RESULTS AND DISCUSSION

5.1 Influence of Excavation Rate on the Stability of UVC

5.1.1 Estimation of Safe Height (FOS = 1.2)

A series of analyses was carried out to investigate the influence of excavation rate (i.e. 1

sec, 1 min, 5 min, 15 min, 30 min, 1 hour, 4 hours, 12 hours, 18 hours and 1 day) on the

FOS for various levels of GWT (1, 2, 3, 4, and 5 m). For this, UVC was excavated at a

constant rate up to a certain depth until FOS = 1.2 is achieved (i.e. safe height). In some

cases, FOSs were slightly higher than 1.2 since excavations were conducted at 0.15 m

increments. Figure 5.1, Figure 5.2, and Figure 5.3 show the stability analysis results with

GWT at 1, 3, and 5 m, respectively for different excavation rates (i.e. 1 sec, 4 hrs, 12 hrs

and 24 hrs). As expected, drop in pore-water pressure in the vicinity of excavated area is

more predominant with the quick excavation rate. This is simply because there was not

enough time to reach equilibrium condition in terms of pore-water pressure when the

duration between the staged excavations is short.

44
 1.24 1.22
10 10

9 9

8 8

7 7

6 6
Elevation (m)

Elevation (m)
5 5

4 4

3 3
Water Pressure
Water Pressure
-20 - 0 kPa
2 2 -10 - 10 kPa
0 - 20 kPa
10 - 30 kPa
20 - 40 kPa
30 - 50 kPa
40 - 60 kPa
1
60 - 80 kPa 1 50 - 70 kPa
80 - 100 kPa 70 - 90 kPa

0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Distance (m) Distance (m)

(a) (b)

1.23 1.23
10 10

9 9

8 8

7 7

6 6
Elevation (m)
Elevation (m)

5 5

4 4

3 3

Water Pressure
Water Pressure
2
2 -10 - 10 kPa
-10 - 10 kPa 10 - 30 kPa
10 - 30 kPa 30 - 50 kPa
30 - 50 kPa 1
1 50 - 70 kPa
50 - 70 kPa
70 - 90 kPa 70 - 90 kPa
0
0 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Distance (m)
Distance (m)

(c) (d)

Figure 5.1. Slope stability analyses results with GWT at 1 m for different excavation
rate: (a) 1 second; (b) 4 hours; (c) 12 hours; and (d) 24 hours.

45
 1.24 1.24
10 10

9 9

8 8

7 7

6 6

Elevation (m)
Elevation (m)

5 5

4 4

3 3
Water Pressure
Water Pressure
2 -40 - -20 kPa 2
-20 - 0 kPa -30 - -10 kPa
0 - 20 kPa -10 - 10 kPa
20 - 40 kPa 10 - 30 kPa
1 1 30 - 50 kPa
40 - 60 kPa
60 - 80 kPa 50 - 70 kPa
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

Distance (m) Distance (m)

(a) (b)

1.24 1.24
10 10

9 9

8 8

7 7

6
Elevation (m)

6
Elevation (m)

5 5

4 4

3
3
Water Pressure
Water Pressure 2
2 -30 - -10 kPa
-30 - -10 kPa -10 - 10 kPa
-10 - 10 kPa 10 - 30 kPa
10 - 30 kPa 1 30 - 50 kPa
1
30 - 50 kPa 50 - 70 kPa
50 - 70 kPa
0
0 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Distance (m)
Distance (m)

(c) (d)

Figure 5.2. Stability analyses results with GWT at 3 m for different excavation rate:
(a) 1 second; (b) 4 hours; (c) 12 hours; and (d) 24 hours.

46
 1.20 1.20

10 10

9 9

8 8

7 7

6 6
Elevation (m)

Elevation (m)
5 5

4 4

3 3
Water Pressure
-60 - -40 kPa Water Pressure
2 2
-40 - -20 kPa -50 - -30 kPa
-20 - 0 kPa -30 - -10 kPa
0 - 20 kPa -10 - 10 kPa
1 1
20 - 40 kPa 10 - 30 kPa
40 - 60 kPa 30 - 50 kPa
0
0
0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Distance (m) Distance (m)

(a) (b)

1.20 1.20

10 10

9 9

8 8

7 7

6 6
Elevation (m)
Elevation (m)

5 5

4 4

3 3

Water Pressure Water Pressure

2 -50 - -30 kPa 2
-50 - -30 kPa
-30 - -10 kPa -30 - -10 kPa
-10 - 10 kPa -10 - 10 kPa
1 10 - 30 kPa 1
10 - 30 kPa
30 - 50 kPa 30 - 50 kPa

0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

Distance (m) Distance (m)

(c) (d)

Figure 5.3. Stability analyses results with GWT at 5 m for different excavation rate:
(a) 1 second; (b) 4 hours; (c) 12 hours; and (d) 24 hours.

47
The result showed that the depth of UVC to achieve FOS = 1.2 (i.e. safe height) was not

affected by excavation rate regardless of the level of GWT, shown in Figure 5.4.. This

indicates that the redistribution of matric suction (or pore-water pressure) between the

ground surface and the toe of UVC does not affect the overall stability of UVC in case

where UVC is excavated with certain safety margin (i.e. 1.2 in this research).
Excavatoin depth with FOS = 1.2 (m)

4.0

3.5

3.0
GWT at 1 m
GWT at 2 m
2.5 GWT at 3 m
GWT at 4 m
GWT at 5 m
2.0

1.5

1.0
1 10 100 1000 10000 100000
Excavation rate (sec)
Figure 5.4. Variation of safe height with respect to excavation rate for different levels
of ground water table (i.e. 1, 2, 3, 4, and 5 m).

48
5.1.2 Estimation of Critical Height (FOS = 1)

Figure 5.5 shows the variation of FOS of UVC prior and post equilibrium state with respect

to pore-water pressure excavated to its critical height (i.e. maximum excavation depth

without failure, FOS = 1) at 1 second excavation rate (GWT at 1 m). Unlike the cases

excavated with FOS = 1.2, the UVC failed as GWT rebounds. Therefore, field works

should pay more attention in case UVC is excavated in a fine-grained soil to its critical

height at fast excavation rate.

49
 1.02
10

Elevation (m) 6

Water Pressure
3
-20 - 0 kPa
0 - 20 kPa
2 20 - 40 kPa
(a) 40 - 60 kPa
60 - 80 kPa
1 80 - 100 kPa

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

0.97
10

7
Elevation (m)

Water Pressure
3
-20 - 0 kPa
0 - 20 kPa
2 20 - 40 kPa
(b) 40 - 60 kPa
60 - 80 kPa
1 80 - 100 kPa

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

Figure 5.5. FOS of UVC (a) prior and (b) post equilibrium condition with respect to
pore-water pressure (excavation rate = 1 second, ground water table at 1 m).

50
For the purpose of comparison, additional stability analyses were carried out using

Bishop’s simplified method (Limit Equilibrium Method) for GWTs at 1, 2, 3, 4, and 5 m.

It was assumed that the GWTs remained unchanged, which represents minimum influence

of matric suction on the stability. Figure 5.6 shows stability analysis results using both

coupled - SIGMA/W stress and Bishop’s simplified methods with GWT at 1 m, 3 m, and

5 m. Figure 5.7 shows the comparison of critical height estimated from both methods with

increasing level of GWT at 1 second excavation rate. The maximum difference in the

critical height was estimated to be 1 m for the GWT at 5 m. This result demonstrates that,

for a deep GWT, the critical height of UVC can be significantly overestimated in case

where UVC is excavated at fast rate with FOS close to unity.

51
 1.01 1.01
10 10

9 9

8 8

7 7

6 6
Elevation (m)

Elevation (m)
5
(a) 5
(d)
4 4

3 3
Water Pressure
-20 - 0 kPa Water Pressure
2 2
0 - 20 kPa -20 - 0 kPa
20 - 40 kPa 0 - 20 kPa
40 - 60 kPa 20 - 40 kPa
1
60 - 80 kPa 1 40 - 60 kPa
80 - 100 kPa 60 - 80 kPa
80 - 100 kPa
0
0
0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Distance (m) Distance (m)

0.99 1.04
10
10

9
9

8
8

7
7

Elevation (m) 6

(b) (e)
6
Elevation (m)

5 5

4 4

3 3
Water Pressure Water Pressure

2 -40 - -20 kPa 2 -40 - -20 kPa

-20 - 0 kPa -20 - 0 kPa
0 - 20 kPa 0 - 20 kPa
1 20 - 40 kPa 1 20 - 40 kPa
40 - 60 kPa 40 - 60 kPa
60 - 80 kPa 60 - 80 kPa
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

Distance (m) Distance (m)

1.01 1.01

10
10

9
9

8
8

7
7

(c)
6
(f)
Elevation (m)

6
Elevation (m)

5
5

4 4

3 3
Water Pressure Water Pressure
-60 - -40 kPa 2 -60 - -40 kPa
2
-40 - -20 kPa -40 - -20 kPa
-20 - 0 kPa -20 - 0 kPa
0 - 20 kPa 1 0 - 20 kPa
1
20 - 40 kPa 20 - 40 kPa
40 - 60 kPa 40 - 60 kPa
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

Distance (m) Distance (m)

Figure 5.6. Stability analyses with 1 second excavation rate for different levels of
GWT (1, 3, and 5 m): (a), (b), (c) coupled - SIGMA/W stress method; (d), (e), (f)
Bishop’s simplified method.

52
 6

Bishop's Simplified Method

Coupled - SIGMA/W stress
5 (current study, excavation rate = 1 sec)
Critical height (m)

1
1 2 3 4 5 6
GWT (m)
Figure 5.7. Comparison of critical height estimated using coupled - SIGMA/W stress
and Bishop’s simplified methods at 1 second excavation rate for different levels of
GWT.

Richard (2018) studied the critical height of UVC in a sand (Unimin 7030) for various

levels of GWT. The critical heights were estimated using both coupled – SIGMA/W stress

at two different excavation rates (i.e. 10 and 1000 seconds) and Morgenstern-Price method

(limit equilibrium method). Comparison of the critical heights estimated using both

methods are shown in Figure 5.8.

53
 1.5
Richard (2018) coupled - SIGMA/W stress
(excavation rate = 10 s)
coupled - SIMGA/W stress
(excavation rate = 10000 s)
Critical height (m)

Morgenstern-Price method
1.0
(Limit equilibrium method)

0.5

0.0
0.0 0.5 1.0 1.5 2.0
GWT (m)
Figure 5.8. Variation of the critical height of UVC in sand (Unimin 7030) with respect
to the level of GWT from coupled – SIGMA/W stress (excavation rates = 10 and 10000
s) and Morgenstern-Price method (limit equilibrium method) (modified after
Richard 2018).

Unlike the results in Figure 5.5 and Figure 5.7, the estimated critical heights using coupled

– SIGMA/W stress method for the two excavation rates (i.e. 10 and 1000 seconds) were

approximately the same regardless of the levels of GWT. This can be attributed to the

relatively high permeability of sand, which brought the equilibrium condition within short

period of time. However, the critical heights estimated using Morgenstern-Price method

were significantly low for the level of ground water table less than 0.9 m. This phenomenon

54
is relevant to concentration of shear stress at the toe of UVC in coupled – SIGMA/W stress

analysis, not the excavation rate.

5.1.3 Summary and Conclusions

The influence of excavation rate on the safe and critical heights of UVC was investigated.

For this, stability analyses were carried out for various excavation rates (from 1 second to

24 hrs) and levels of GWT (1, 2, 3, 4, and 5 m). The coupled - SIGMA/W stress analysis

results showed that the excavation rate does not affect the overall stability of UVC in case

where excavation is performed considering a certain safety margin (i.e. FOS = 1.2 in this

research). However, the UVC eventually failed if UVC is excavated up to a critical height

(i.e. FOS = 1) due to the rebound of the GWT. The critical heights estimated using the

coupled - SIGMA/W stress method were significantly overestimated when compared with

those from the Bishop’s simplified method (Limit Equilibrium Method). However,

research by others has shown that the influence of excavation rate on the critical heights of

UVC in sand is negligible due to its high permeability. This indicates that permeability

function of a soil is a key parameter in estimating the critical height of UVC when

excavation rate is considered.

55
5.2 Determining the Influence of Tension Crack on Stability of
Unsupported Vertical Cut

5.2.1 Determining the Critical Height of UVC using Numerical Method

As mentioned earlier, tension crack was considered and simulated as a void in SIGMA/W.

UVC was first excavated to a depth where FOS = 1.2 (safe height) under the equilibrium

condition with respect to matric suction. The stability analyses were performed for various

depths (up to 70% of safe height) and distances (up to 50% of safe height from the

excavation wall) of tension cracks with three levels of GWT (i.e. 1, 3, and 5 m). Each

analysis is denoted by its depth ratio and distance ratio as shown in Eq.(5.1) and Eq. (5.2),

respectively.

Depthof atensioncrack
Depthratio ( DPtc ) = (5.1)
Safeheight (FOS = 1.2)

D istance of a tension crack fro m the excavation w all (5.2)

D istan ce ra tio ( D S tc ) =
S afe heigh t ( F O S = 1 .2 )

Analyses were performed keeping one of the two above-mentioned parameters constant

while varying the other until the factor of safety fell below unity; namely, if distance ratio

(DStc) was kept constant then depth ratio (DPtc) is varied till FOS < 1 (Figure 5.9) and vice

versa (Figure 5.10). This analysis pattern efficiently determines the depths and locations of

tension cracks that cause failure in UVC. Figure 5.11 shows an example of analysis tree

with various DPtc of tension crack.

56
 10 10

9 9

8 8

7 7
(a) (b)

10 10

9 9

8 8

7 7
(c) (d)

10 10

9 9

8 8

7 7
(e) (f)

Figure 5.9. Simulating tension crack for the safe height of 2.55 m with distance ratio
(DStc) fixed at 0.1 and six different depth ratios (DPtc): (a) 0; (b) 0.1; (c) 0.2; (d) 0.3;
(e) 0.4; and (f) 0.5.

57
 10 10

9 9

8 8

7 (a) 7 (b)

10 10

9 9

8 8

7 (c) 7 (d)

10

7 (e)

Figure 5.10. Simulating tension crack for the safe height of 2.55 m with depth ratio
(DPtc) fixed at 0.5 and five different distance ratios (DStc): (a) 0.1; (b) 0.2; (c) 0.3; (d)
0.4; and (e) 0.5.

58
 Safe height (FOS = 1.2)

DPtc = 0.1 (0.255 m)

DPtc = 0.2 (0.510 m)

DPtc = 0.3 (0.765 m)

DPtc = 0.4 (1.02 m)

DPtc = 0.5 (1.275 m)

Figure 5.11. Analysis tree for various depth ratio (DPtc) in SIGMA/W to determine
critical depth of tension crack.

Examples of stability analysis results with different combinations of DPtc and DStc are

shown in Figure 5.12. It was assumed that a slip surface begins at bottom of the tension

crack. Analyses were carried out for multiple scenarios to determine critical combinations

of DPtc and DStc. Table 5.1 summarizes FOS for various DPtc and DStc combinations with

GWT at 1, 3, and 5 m.

59
 1.00
10

Elevation (m)
6

4
Water Pressure
-10 - 0 kPa
3 0 - 10 kPa
10 - 20 kPa
20 - 30 kPa
2 30 - 40 kPa
40 - 50 kPa
50 - 60 kPa
1 60 - 70 kPa
70 - 80 kPa
80 - 90 kPa
0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

(a)

0.98
10

7
Elevation (m)

4 Water Pressure
-30 - -20 kPa
-20 - -10 kPa
3 -10 - 0 kPa
0 - 10 kPa
10 - 20 kPa
2 20 - 30 kPa
30 - 40 kPa
40 - 50 kPa
1 50 - 60 kPa
60 - 70 kPa

0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)

(b)

Figure 5.12. Stability analysis considering tension crack with different combinations
of DStc and DPtc: (a) DStc = 0.4, DPtc = 0.4 (GWT at 1 m); (b) DStc = 0.5, DPtc = 0.5
(GWT at 3 m).

60
Table 5.1. Variation of factor of safety for different combinations of DPtc and DStc
with GWT at 1, 3 and 5 m.

GWT = 1 m (Hsafe = 1.5 m)*

DStc
DPtc
0.1 0.2 0.3 0.4 0.5
0.1 1.86 1.4 1.21 1.14 1.12
0.2 1.62 1.24 1.1 1.04 1.03
0.3 1.41 1.11 1.02 1 1
0.4 1.3 1.1 - - -
0.5 1.27 1.08 - - -
0.6 1.27 1.08 - - -
0.7 1.23 1.05 - - -
GWT = 3 m (Hsafe = 2.55 m)*
DStc
DPtc
0.1 0.2 0.3 0.4 0.5
0.1 2.45 1.56 1.29 1.2 1.16
0.2 2.11 1.38 1.19 1.1 1.13
0.3 1.82 1.23 1.07 1.05 1.11
0.4 1.59 1.09 0.98 1.07 1.21
0.5 1.4 0.98 0.92 0.92 0.98
0.6 1.39 - - - -
0.7 1.25 - - - -
GWT = 5 m (Hsafe = 3.45 m)*
DStc
DPtc
0.1 0.2 0.3 0.4 0.5
0.1 2.32 1.55 1.24 1.18 1.13
0.2 2.01 1.38 1.13 1.1 1.11
0.3 1.73 1.22 1.04 1.04 1.1
0.4 1.5 1.1 0.92 0.92 0.95
0.5 1.32 0.99 - - -
0.6 1.05 - - - -
0.7 0.98 - - - -
Hsafe = Safe height estimated using couple – SIGMA/W stress method

61
5.2.2 Estimating a Depth of Tension Crack using Pufahl et al. (1983) approach

For the purpose of estimating the depth of tension cracks, the approach proposed by Pufahl

et al. (1983; Eq. (2.15)) was adopted. In this method, the maximum expected depth of

tension crack in vadose zone can be estimated by locating the depth with zero net active

earth pressure as shown in Figure 5.13.

[σvKa - 2c' Ka
σvKa 2c' Ka (ua - uw )(Sκ)tanφ' - (ua - uw)(Sκ)tanφ']

Depth of tension zone

z
Area I
γunsat

H
+ + =

γsat

Figure 5.13. Determination of depth of tension crack based on net active earth
pressure distribution in vadose zone (Eq. (2.15)) extending the approach by Pufahl et
al. (1983).

The positive, negative, and net active earth pressure diagrams with GWT at 1, 3, and 5 m

are shown in Figure 5.14, Figure 5.15, and Figure 5.16, respectively.

62
 Posive earth pressure (σvKa)
Negative earth pressure (2C Ka)
Net earth pressure (σvKa - 2c Ka)

0.0

zt
0.5
Depth (m)

1.0

1.5

2.0
-12 -8 -4 0 4 8 12 16
Pressure (kPa)
Figure 5.14. Positive, negative, and net active earth pressure distribution (GWT = 1
m).

63
 Positive earth pressure (σvKa)
Negative earth pressure (2C Ka)
Net earth pressure (σvKa - 2C Ka)

0.0

0.5
zt

1.0
Depth (m)

1.5

2.0

2.5

3.0
-20 -10 0 10 20 30
Pressure (kPa)
Figure 5.15. Positive, negative, and net active earth pressure distribution (GWT = 3
m).

64
 Positive earth pressure (σvKa)
Negative earth pressure (2C Ka)
Net earth pressure (σvKa - 2C Ka)

zt
1

2
Depth (m)

5
-20 -10 0 10 20 30 40 50
Pressure (kPa)
Figure 5.16. Positive, negative, and net active earth pressure distribution (GWT = 5
m).

65
5.2.3 Determination of Critical Tension Crack

Based on the obtained FOS in Table 5.1, contours of FOS for different combinations of

DPtc and DStc are plotted in Figure 5.17, Figure 5.18, and Figure 5.19 with GWT at 1, 3,

and 5 m, respectively. The depths of tension crack estimated using Eq. (2.15) were 0.83,

1.34, and 1.56 m with GWT at 1 m, 3 m, and 5 m, respectively, which leads to an average

depth ratio of 0.59 (i.e. 0.55, 0.64, and 0.57 with GWT at 1, 3, and 5 m, respectively). This

depth ratio is similar to the range proposed by Kutschke and Vallejo (2011) based on the

finite element analysis of UVC in stiff clay (i.e. 0.45 – 0.57). As expected, FOS decreases

with increasing the depth ratio. In case where distance ratio is less than 0.2, FOS is greater

than unity up to the depth of tension crack estimated using Eq. (2.15) regardless of the level

of GWT. If a tension crack is developed at a distance ratio greater than about 0.2 failure

can take place before the tension crack can reach the depth estimated by Eq. (2.15) or

proposed by Kutschke and Vallejo (2011). It can also be seen that UVC can remain stable

unless depth ratio exceeds 0.3 for the levels of GWT used in the analysis. According to

OSHA (2019)’s guideline, tension cracks usually form at distance ratio of 0.5 to 0.7

(Information of tension crack depth is not available in OSHA 2019). If this guideline is

implemented in practice, the depth ratio of 0.3 can be used as a maximum ratio to avoid

UVC failure due to tension crack for the range of GWT levels used in the research. Baker

(1981) showed that maximum depth of tension crack is 25% of vertical height of slopes,

which suggests that tension crack does not affect the stability of UVC considered in this

research.

66
 DStc
DS /Htc
safe

0.1 0.2 0.3 0.4 0.5

0.1 1.8
1.4
1.2
1.6

Maximum depth of tension crack

0.2 suggested by Baker (1981)

1.4 1.2
0.3 1.0
1.0
tcsafe

GWT = 1 m
tc/H

0.4
DP

1.0
DP

1.2
Maximum range of tension crack
depth suggested by Kutschke
0.5 1.0 and Vallejo (2011)

0.6
1.2 Maximum depth of tension crack from
current study [Eq. (2.15)]
1.0

0.7

Figure 5.17. Contours of FOS for different combinations of DStc and DPtc with GWT
at 1 m.

67
 DSDS
tc/Htc
safe

0.1 0.2 0.3 0.4 0.5

0.1
1.8 1.2
2.2
1.4
2.0 1.6
Maximum depth of tension
0.2 1.2 crack suggested by Baker
(1981)

1.8
0.3 1.4
1.6 1.2
GWT = 3 m
tcsafe

1.0
tc/H

1.2
0.4
DPDP

1.0
1.4

1.2 Maximum range of tension crack 1.0

0.5 1.0 depth suggested by Kutschke and
Vallejo (2011)

0.8
0.8
0.6
0.6

1.2 0.8 0.6

1.0
0.4
0.7
Maximum depth of tension crack
from current study [Eq. (2.15)]

Figure 5.18. Contours of FOS for different combinations of DStc and DPtc with GWT
at 3 m.

68
 DSDS
tc/Hsafe
tc

0.1 0.2 0.3 0.4 0.5

0.1
2.2 1.2
1.4
2.0 1.6

1.8
0.2 Maximum depth of tension crack
suggested by Baker (1981)
1.2

1.4
0.3 1.6
1.0
tcsafe

1.0 1.0
1.2
tc/H

0.4
DPDP

1.4

Maximum range of tension crack

0.5 1.0 depth suggested by Kutschke and
1.2 Vallejo (2011)

GWT = 5 m

0.6
1.0
1.0
Maximum depth of tension crack
from current study [Eq. (2.15)]

0.7

Figure 5.19. Contours of FOS for different combinations of DStc and DPtc with GWT
at 5 m.

69
5.2.4 Summary and Conclusions

Tension cracks have an adverse impact on the stability of slopes. The location and depth

of a tension crack can be estimated using empirical, semi-empirical, or numerical

approaches available in literature. According to the present research, FOS of UVC does

not drop below unity if tension crack distance and depth ratio can be limited within 0.2 and

0.3, respectively, for the levels of GWT considered in the analyses. The findings from this

study indicate that critical depth and distance of a tension crack should be estimated using

several different approaches since, in some cases, UVC can fail before the tension crack

reaches the depth estimated by approaches and guidelines available in the literature.

70
5.3 Influence of Rainfall Infiltration into Tension Crack on the
Stability of UVC

5.3.1 Methodology

According to Ileme and Oh (2019), the stand-up time of UVC in Indian Head Till becomes

minimum at a rainfall intensity of 10 mm/hr, and then remains constant thereafter up to 25

mm/hr. Hence, three different rainfall intensities; 10, 15, and 25 mm/hr were chosen in the

seepage analyses. As mentioned earlier, in this study, rainfall in the tension crack is allowed

to seep into deeper soil by establishing potential seepage face boundaries along the

perimeter of tension crack. This technique allows users to consider the worst but realistic

field conditions assuming tension cracks with relatively wide openings. Figure 5.20. shows

pore-water pressure and water flux vector distribution in a tension crack caused by rainfall

infiltration. Vectors around the crack are of relatively high magnitude and are spreading

down and out. The variation of pore-water pressure around a crack with time under the

rainfall intensity of 25 mm/hr is shown in Figure 5.21. (DStc = 0.2, DPtc = 0.3, GWT at 5

m).

71
 10

7
Elevation (m)

Water Pressure
4
-60 - -50 kPa
-50 - -40 kPa
3 -40 - -30 kPa
-30 - -20 kPa
-20 - -10 kPa
2 -10 - 0 kPa
0 - 10 kPa
10 - 20 kPa
1 20 - 30 kPa
30 - 40 kPa
40 - 50 kPa
0
0 1 2 3 4 5 6 7 8 9 10

Distance (m)
Figure 5.20. Water pressure and water flux vector distribution in a tension crack
associated with rainfall infiltration.

72
 Water Pressure
-60 - -50 kPa
-50 - -40 kPa
-40 - -30 kPa Water Pressure
-30 - -20 kPa
-20 - -10 kPa
-60 - -50 kPa
-10 - 0 kPa -50 - -40 kPa
0 - 10 kPa -40 - -30 kPa
-30 - -20 kPa
-20 - -10 kPa
-10 - 0 kPa
0 - 10 kPa

Initial 7 hours

Water Pressure
Water Pressure
-60 - -50 kPa
-50 - -40 kPa -60 - -50 kPa
-40 - -30 kPa -50 - -40 kPa
-30 - -20 kPa -40 - -30 kPa
-20 - -10 kPa -30 - -20 kPa
-10 - 0 kPa -20 - -10 kPa
0 - 10 kPa -10 - 0 kPa
0 - 10 kPa

1 day 2 days

Water Pressure Water Pressure

-60 - -50 kPa -60 - -50 kPa
-50 - -40 kPa -50 - -40 kPa
-40 - -30 kPa -40 - -30 kPa
-30 - -20 kPa -30 - -20 kPa
-20 - -10 kPa -20 - -10 kPa
-10 - 0 kPa -10 - 0 kPa
0 - 10 kPa 0 - 10 kPa

6 days 10 days

Figure 5.21. Variation of pore-water distribution with time around a tension crack
under 25 mm/hr rainfall intensity (DStc = 0.2, DPtc = 0.3, GWT at 5 m).

73
It can be clearly seen that, with the presence of tension crack, water seeps into deeper level

of soil at fast rate, which in turn, increases pore-water pressure around vicinity of the crack.

This increment in pore-water pressure has detrimental impact on the stability of UVC as

shear strength of the soil decreases sharply.

Figure 5.22 shows pore-water pressure distribution and FOS for different time step under

25 mm/hr rainfall intensity with a tension crack (DStc = 0.4, DPtc = 0.1, GWT at 5 m). FOS

gradually decreases with time from 1.18 to 1.01 over the period of 14 days. This period

between the start of rainfall and the time of failure is denoted as stand-up time (Figure

5.23).

74
 1.18 1.07
10 10
-45 -5

9 -40 9
initial -35
6 day -35
8 -30 8
-25 -25
7 -20
7
-15 -15
6 -10 6
-5 -5
5 5

4 4

1.14 1.04
10 -5 10
-5
9 -40 9 -40

1 days -35
10 days -35
8 8
-25 -25
7 7
-15 -15
6 6
-5 -5
5 5

4 4

1.12 1.01

10 -5 10
-5
9 -40 9
2 days -35 14 days -35
8 8
-25 -25
7 7
-15 -15
6 6
-5 -5
5 5

4 4

Figure 5.22. Pore-water pressure distribution and factor of safety for different time
step under 25 mm/hr rainfall intensity with a tension crack (DStc = 0.4, DPtc = 0.1,
GWT at 5 m).

75
 1.25

initial
1.20
Stand-up time
Factor of safety

1.15

1.10

1.05 failure

1.00
0 100 200 300 400
Time (hour)

Figure 5.23. Variation of factor of safety with time based on the results in Figure 5.22
and the definition of stand-up time.

5.3.2 Stand-Up time of UVC with Tension Crack under Different Rainfall
Intensities

As mentioned earlier, UVC that is initially excavated with FOS = 1.2 can fail due to tension

crack even without any rainfall events. Hence, in this section, seepage and stability

analyses were carried out for the cases with an initial FOS close to 1.2 with a tension crack

from Table 5.1. Slope stability analysis results with three different rainfall intensities (i.e.

10, 15, and 25 mm/hr) are summarized in Table 5.2. As can be seen, UVC remains stable

even under 25 mm/hr rainfall intensity in case where the GWT is shallow (i.e. 1 m). For

the levels of GWT at 3 m and 5 m, FOS drops below unity due to the rainfall events;

76
however, the stand-up time ranged between 7 and 20 days. These durations of rainfall are

unrealistic considering the environmental data of Canada. Hence, it can be concluded that

if FOS of UVC with a tension crack is greater than 1.2, rainfall infiltration into a tension

crack may not affect the overall stability of the UVC except extremely long duration of

rainfall.

Table 5.2. Stand-up time of UVC with a tension crack under different rainfall events
(Initial FOS of UVC is close to 1.2).

GWT DStc DPtc FOSa) (Time in daysb))

(m)
Initial 10 15 25
(with mm/hr mm/hr mm/hr
crack)

1 0.2 0.2 1.25 1.18 (no failure) 1.19 (no failure) 1.18 (no failure)

1 0.3 0.1 1.24 1.08 (no failure) 1.08 (no failure) 1.08 (no failure)

3 0.2 0.3 1.21 0.93 (7) 0.93 (7) 0.93 (7)

3 0.3 0.2 1.22 0.98 (10) 0.98 (10) 0.98 (10)

3 0.4 0.1 1.21 1.04 (10) 1.04 (10) 1.04 (10)

5 0.2 0.3 1.25 0.98 (8.5) 0.98 (8.5) 0.98 (8.5)

5 0.3 0.1 1.24 1.03 (20) 1.03 (20) 1.03 (20)

5 0.4 0.1 1.18 1.01 (14) 1.01 (14) 1.01 (14)

a): minimum FOS
b): time to reach the minimum FOS

77
5.4 Summary and Conclusions

The influence of rainfall infiltration into tension crack on the stability of UVC was

investigated under three different rainfall intensities: 10, 15, and 25 mm/hr. As expected,

rainfall infiltration into a tension crack decreased FOS with time and then eventually lead

to a failure of UVC for most cases. However, extremely long duration of rainfall was

required if the initial FOS with a tension crack is close to 1.2.

78
6. CASE STUDY

To validate the approach used in the present study, instrumented large scale field

excavation results are revisited.

6.1 Description of Project

Kwan (1971) investigated the behavior of a deep UVC cut and an inclined slope in clay.

The excavation was made in Haldimand clay at Welland, Ontario with the measurements

of pore-water pressures and surface deformation. Soil profile and the variation of water

content, shear strength parameters, and coefficient of permeability at test site are provided

in Figure 6.1. The top desiccated sediment of 5.2 m was first removed, followed by the

UVC (9.75 m) and the inclined slope (1:1 slope with 6.1 m berm) excavation. To eliminate

end effects, two vertical trenches were excavated at the ends, which limited the length of

slope to 15.2 m.

79
 WATER CONTENT
(%) PERMEABILITY, k
DESCRIPTION c', φ'
ELEV. m w (cm/s)
wP wL
10 20 30 40 50 60
175.8
c' = 9.58 kPa
LACUSTRINE SILTY CLAY
φ' = 21.5° k < 10-6

STATIFIED LAYER
169.7

CLAYEY SILT TILL

c' = 13.4 kPa k < 10-6
φ' = 24.5°

162.4
STRATIFIED CLAY k < 10-6
159.9
LACUSTRINE SILTY CLAY k < 10-6
156.9 c' = 13.4 kPa
STRATIFIED CLAY k < 10-6
155.1 φ' = 22.6°

SILT AND CLAY k < 10-6

151.7
SANDY SILT TILL k < 10-5
148.7

Figure 6.1. Soil profile and variation of water content, shear strength parameters and
coefficient of permeability with depth at test site (after Kwan 1971).

Staged excavation was carried out for 27 days till the occurrence of failure. Timeline of the

excavation is detailed in Figure 6.2. Multiple piezometers were installed after removal of

top sediment and pressure head readings were taken for 12 days after removal. Excavation

profile and location of piezometers are shown in Figure 6.3

80
EL. 175.87 m

EL. 170.69 m

EL. 166.73 m

EL. 164.00 m

EL. 160.93 m

26 JANUARY 4 FEBRUARY 15 19 22

Figure 6.2. Timeline of excavation (after Kwan 1970).

EL. 175.87 m

3
EL. 170.69 m
1 4 8 12
5 9 13
10 14

6
EL. 160.93 m
7 11 15
2

Bedrock

Figure 6.3. Excavation profile and location of piezometers (after Kwan 1970).

81
The pore-water pressures were significantly dropped in the areas located within a distance

of 24.2 m from the UVC during removal of top sediment layer and vertical excavation.

During the entire excavation period, low overall precipitation was recorded in the range of

0.25 mm to 10 mm. Hence, the precipitation event was not taken into account in the

numerical analysis. The minimum temperature recorded was -25 °C; however, it rarely

dropped below -18 °C. Kwan (1971) conducted a stability analysis of the UVC cut using

both effective and total stress approaches. Later, Banerjee et al. (1988) carried out finite

element analysis to estimate stability of the UVC using the same data. However, the

influence of matric suction on the shear strength of the soils was not considered in either

study.

6.2 Soil Properties

Soil profile was simplified with three different layers; namely, overburden lacustrine silty

clay, clayey silt till and lacustrine silty clay from the top. Material properties of each layer

were adopted from Kwan (1971) and Banerjee et al. (1988) and summarized in Table 6.1.

82
Table 6.1. Material properties used in the numerical analysis (adopted from Kwan
1971 and Banerjee et al. 1988).

No Type of soil Elevation Effective shear strength Saturated

(m) parameters coefficient of
permeability
c’ (kPa) φ’ (deg) ksat (m/s)

1 Lacustrine 169.7 – 9.58 21.5 5 × 10-7

silty clay 175.8

2 Clayey silt till 162.4 – 13.4 24.5 5 × 10-7

169.7

3 Lacustrine 157.7 – 13.4 22.6 5 × 10-8

silty clay 162.4

According to Banerjee et al. (1988), the properties of clayey silt were similar to Weald clay

(Schofield and Wroth 1968). In addition, the plasticity index and percent passing of 425

mm sieve for Weald and London clays are similar as well. Hence, in the present research,

the SWCC of London clay was used for the clayey silt layer. In case of the SWCC of

Lacustrine silty clay, the SWCC of sample material, ‘Silty Clay’ available in GeoStudio

(2019 R2) was used. The SWCC and permeability functions used in the numerical analyses

are shown in Figure 6.4 and Figure 6.5, respectively.

83
 1.0

Volumetric water content, θ

0.8

0.6

0.4

Lacustrine silty clay

0.2 Clayey silt till

0.0
0.01 0.1 1 10 100 1000
Suction (kPa)

Figure 6.4. Soil-Water Characteristic Curve for materials used in numerical analysis.

Bottom Lacustrine silty Clay

1e-6 Top Lacustrine silty Clay
Coefficient of permeability, k (m/sec)

Clayey silt till

1e-7

1e-8

1e-9

1e-10
0.01 0.1 1 10 100 1000 10000

Suction (kPa)

Figure 6.5. Permeability function for materials used in numerical analysis.

84
6.3 Methodology

Figure 6.6 shows mesh and boundary conditions used in the finite element analysis.

Quadrilateral and triangle mesh with secondary nodes were used to define regions with

global size of 0.5 m near the excavated areas based on the sensitivity analysis.

180

175

Stage I
170

Stage II
Elevation (m)

165

160

155

150

145
0 10 20 30 40 50 60

Distance (m)

Figure 6.6. Mesh and boundary conditions used in the numerical analysis (11853
nodes, 4523 elements).

As mentioned in earlier section, a significant drop of pore pressures was observed near the

trench area during excavation. Figure 6.7. shows the comparisons of measured and

estimated (coupled analysis) pore-water pressure contours (0 (i.e. phreatic line), 60, and

120 kPa) 12 days after removal of top sediment layer.

85
 Excavated geometry
0 kPa (measured)
60 kPa (measured)
120 kPa (measured)
0 kPa (coupled analyis)
180 60 kPa (coupled analysis)
120 kPa (coupled analysis)

175
+ location of piezometers
Elevation (m)

170

165

160

155
0 5 10 15 20 25 30
Distance (m)

Figure 6.7. Comparisons of measured and estimated (coupled analysis) pore-water

pressure contours (0 (i.e. phreatic line), 60, and 120 kPa) 12 days after removal of top
sediment layer.

Figure 6.8 shows the comparisons of measured and estimated (coupled analysis) pore-

water pressure contours for 0 kPa (i.e. phreatic line) and 60 kPa after completion of the

stage 2 cut, prior to failure. The pore-water pressures from the coupled analysis were

slightly higher within the cut.

86
 Excavated geometry
0 kPa (measured)
60 kPa (measured)
180 0 kPa (coupled analysis)
60 kPa (coupled analysis)

175
+ location of piezometers
Elevation (m)

170

165

160

155
0 5 10 15 20 25 30
Distance (m)

Figure 6.8. Comparison of measured and estimated pore-water pressure contours (0

(phreatic line) and 60 kPa) prior to failure.

Based on the estimated pore-water pressure distribution, FOS was estimated using both

coupled -SIGMA/W stress and Bishop’s simplified methods, which results in 1.05 and 0.61,

respectively (Figure 6.9 and Figure 6.10). This result indicates that FOS from the coupled

-SIGMA/W stress method better matches the field condition compared to limit equilibrium

method.

87
 Factor of Saf ety
1.05 - 1.15
1.15 - 1.25
1.25 - 1.35
180 1.35 - 1.45 1.05
1.45 - 1.55
1.55 - 1.65
1.65 - 1.75
1.75 - 1.85
1.85 - 1.95
-20
175 -20 ≥ 1.95

-20
0 0

0
170

0
60
60
Elevation (m)

60
165

60
60
60

0
160

Figure 6.9. Pore-water pressure distribution and FOS prior to failure without tension
crack (coupled - SIGMA/W stress method).

88
 Factor of Saf ety
0.61 - 0.71
0.71 - 0.81
0.61
180 0.81 - 0.91
0.91 - 1.01
1.01 - 1.11
1.11 - 1.21
1.21 - 1.31
-20 1.31 - 1.41
175 -20 1.41 - 1.51

-20
≥ 1.51
0 0

0
170

60 0
60
Elevation (m)

60
165

60
60
60

0
160

Figure 6.10. Pore-water pressure distribution and FOS prior to failure without
tension crack (Bishop’s simplified method).

According to Kwan (1971), the reason for failure was due to the formation of a tension

crack at the berm. Two tension cracks appeared after the completion of UVC. The

identified potential slip surface due to the formation of the first tension crack was estimated

to be at 6.4 m from the crest (Figure 6.11). However, the trench unexpectedly failed along

a new slip surface due to formation of the second tension crack at 2.43 m from the crest

(Figure 6.12). Hence, additional stability analysis was carried out as shown in Figure 6.13

with four different locations of tension cracks from the vertical cut face (i.e. 4.5 m, 3.45 m,

2.75 and 2.43 m). The results show that the failure takes place at shallower depth of tension

crack with increasing the distance from the vertical cut face.

89
 3.96 m

9.75 m

Potential
failure
surface

Figure 6.11. Identified potential slip surface with the first tension crack (after Kwan

1971).

2.43 m

2.14 m

6.55 m
9.75 m

54°
Actual
failure
surface

Figure 6.12. Actual slip surfaces with the second tension crack that led to the failure
of UVC (after Kwan 1971).

The ratios of distance and depth of tension crack to the height of UVC for each case are;

(a) = (0.46, 0.14), (b) = (0.35, 0.24), (c) = (0.28, 0.59), and (d) = (0.25, 0.67). Unlike the

results in Figure 5.17, Figure 5.18, and Figure 5.19, these ratios are calculated using the

excavation depth for FOS = 1.05. Hence, it is rational to assume that the ratios can be

higher if the safe height was used in the calculation. As concluded earlier, the tension crack

90
does not have adverse impact on the stability of UVC in case DStc ≤ 0.2 or DPtc ≤ 0.3. Since

the distance ratio for these four cases are greater than 0.25 (or greater than 0.25 if safe

height was used) it can be suggested that these tension crack have potential to lead to a

failure. Since, the depth ratios for cases (a) and (b) are less than 0.24, there is low chance

of failure due to these two tension cracks regardless of distance ratio. Hence, it is

reasonable to conclude that the shape of slip surface can be in the from of case (c) and (d)

if a failure is caused by tension crack. Especially, when crack was located at 2.43 m from

the face of UVC, depth of tension crack was estimated to be 6.55 m with sheared surface

inclined at 52 ° to the horizontal. This failure surface was identical to the geometry of the

second tension crack and failure surface observed in field (Figure 6.12). Through this case

study, it is evident that the methodology proposed in this research can reliably estimate the

location and depth of critical tension crack.

91
 0.98 0.96
3.45 m
4.5 m
(a) (b)

1.345 m

2.34 5 m
0.98 0.98

(c) 2.75 m (d) 2.43 m

5.8 m

6 .55 m

53.62°
52.03°

Figure 6.13. Stability analysis with different locations of tension crack from the
vertical cut face: (a) 4.5 m; (b) 3.45 m; and (c) 2.75 m and (d) 2.43 m (coupled –
SIGMA/W stress method).

92
6.4 Summary and Conclusions

The methodologies adopted in this research was implemented on a case study, to confirm

its applicability. For this, a series of finite element analysis was performed to simulate the

excavation of UVC (depth = 9.75 m) in clay at Welland, Ontario (Kwan 1971). To maintain

the consistency of obtained results with field conditions, the simulated excavation followed

the same timeline as field excavation. Post excavation pore-pressure data obtained from

installed instruments were then compared with the results obtained through performed

numerical modelling. The numerical analysis results showed that the failure in UVC cut

was attributed to a tension crack, which is consistent with field observations.

93
7. GENERAL CONCLUSIONS

In this research, stability of unsupported vertical cut (UVC) excavated into a glacial till

(Indian Head till) is investigated extending the mechanics of unsaturated soils. A series of

numerical analyses were carried out considering practical factors such as excavation rate,

depth and location of a tension crack, and rainfall infiltration into the tension crack.

Conclusions obtained from the research are summarized below.

1. Excavation of UVC causes a drop in the ground water table. The amount of drop

depends on soil type (i.e. permeability function) and excavation rate. This indicates

that the critical height of UVC can be significantly overestimated if it is excavated

at fast rate. However, the numerical analysis results showed that the influence of

excavation rate can be eliminated if UVC is excavated with a certain safety margin

(FOS = 1.2 in this research).

2. Tension crack development is highly detrimental towards the stability of slope;

therefore, UVC can fail upon the development of tension crack behind the crests.

To investigate the influence of tension crack on the stability of UVC, numerical

analyses were conducted for various combinations of depth and distance ratios. The

results showed that FOS of UVC can be maintained greater than unity unless either

depth or distance ratio excess 0.3 or 0.2, respectively. The results also suggest that

UVC can fail before a tension crack reaches the depth calculated using the

approaches available in literature.

94
3. The stand-up time of UVC excavated was investigated by simulating different

rainfall intensities in the numerical analysis. Seepage and stability analyses were

carried out for the cases with an initial FOS close to 1.2 with a tension crack only.

Rainfall infiltration into a tension crack decreases FOS with time and then

eventually led to a failure for most cases. However, extremely long rainfall

durations were required for UVC to fail, which is not realistic considering the

Canadian environmental data. On the contrary, FOS never dropped below unity

with a ground water table at shallow depth (i.e. 1 m in this research).

4. The approaches presented in the research was extended to analyze the stability of

an instrumented large-scale field UVC in clay. There was a good comparison

between the measured and the numerically estimated pore-water pressure

distribution profile prior to failure. The failure of field UVC was attributed to a

tension crack. The distance and depth of the field tension crack was successfully

simulated through the numerical analysis.

95
7.1 Recommendations for Future Research

It is expected that this research can provide field engineers and contractors with guidelines

on determining appropriate depth of UVC considering different field conditions (i.e. soil

type, excavation rate, matric suction distribution, and tension crack) and circumstances (e.g.

rainfall into tension crack).

In this research, worst case scenario was considered to study rainfall induced UVC failure

by assuming tension crack is filled at the same rate as rainfall intensity. Whereas in field,

it is highly unlikely that pressure head is fully developed in tension crack unless the

opening of tension crack is wide enough such that rainfall can infiltrate into tension crack

freely. Hence, further studies must be carried out on this pressure head accumulation and

dissipation characteristic by determining a relationship between rate of infiltration and

opening of a tension crack with respect to time.

The proposed approaches were validated in the research through a case study. To enhance

the reliability and applicability of the proposed approaches, more case studies need to be

done for UVC excavated in various types of soils.

96
REFERENCES

Bagge, G. 1985. Tension cracks in saturated clay cuttings. Proceeding of the 11th

international conference on Soil Mechanics and Foundation Engineering, San

Francisco, (Balkema): 393-395.

Baker, R., and Leshchinsky, D. 2001. Spatial distribution of safety factors. Journal of

Geotechnical and Geoenvironmental Engineering, ASCE, 127(2): 135-145.

Baker, R., and Leshchinsky, D. 2003. Spatial distribution of safety factor: Cohesive vertical

cut. International Journal for Numerical and Analytical Methods in Geomechanics,

27:1057-1078.

Baker. R. 1981. Tensile strength, tension cracks, and stability of slopes. Soils and

Foundations, 21(2): 1-17.

Banerjee, P.K., Kumbhojkar, A., and Yousif, N.B. 1988. Finite element analysis of the

stability of a vertical cut using an anisotropic soil model. Canadian Geotechnical

Journal, 25(1): 119-127.

Bishop A.W. and Bjerrum, L. 1960. The relevance of the triaxial test to the solution of

stability problems. ASCE Conference on Strength of Cohesive soils: 437-501.

BLS (Bureau of Labor Statistics). 2010. Census of fatal occupational injuries (2000−2009).

Washington, D.C.

De Vita, P., Angrisani, A.C., and Di Clemente, E. 2008. Engineering geological properties

of the Phlegraean pozzolan soil (Campania region, Italy) and effect of the suction

97
 on the stability of cut slopes. Italian Journal of Engineering Geology and

Environment, 2: 5-22.

Deng, G., and Shen, Z.J. 2006. Numerical simulation of crack formation process in clays

during drying and wetting. Geomechanics and Geoengineering, 1(1): 27-41.

DiBagio, E., and Roti, J.A. 1972. Earth pressure measurements on a braced slurry-trench

wall in soft clay. In Proceedings of the 5th European Conference on Soil Mechanics

and Foundation Engineering, 1: 473-484.

Dunlop, P., and Duncan, J.M. 1970. Development of failure around excavated slopes.

Journal of Soil Mechanics and Foundation Division, ASCE, 96(SM2): 471-493.

Dysli, M., and Fontana, 1982. Deformations around the excavations in clay soil. In

Proceedings of the International Symposium on Numerical Models in

Geomechanics, Zurich. Edited by R. Dunger and J. A. Studer: 634-642.

Erdogan, F., and Sih, G.C. 1963. On the crack extension in plates under plane loading and

transverse shear. Journal of Basic Engineering, ASME. 85(4):519-527.

Fredlund, D.G., and Xing, A. 1994. Equations for the soil-water characteristic curve.

Canadian Geotechnical Journal. 31: 521-532.

Galera, J.M., Montero, J., Perez C., Varona, P., Vega, L. 2009. Coupled hydromechanical

analysis of Cobre Las Cruces Open Pit. Slope Stability, CSIRO.

GEO-SLOPE International, Ltd. 2013. Stability modeling with SLOPE/W, user’s guide.

GEO-SLOPE International Ltd., Calgary, Canada.

98
GEO-SLOPE International, Ltd. 2013. Stress-Deformation modeling with SIGMA/W,

user’s guide. GEO-SLOPE International Ltd., Calgary, Canada.

Gofar, N., Min, L. L., and Asof, M. 2006. Transient seepage and slope stability analysis

for rainfall-induced landslide: a case study. Malaysian Journal of Civil Engineering,

18(1): 1-13.

Hu, S. 2000. Reliability of slope stability considering infiltration through surface cracks.

Master’s Thesis, Hong Kong University of Science and Technology.

Ileme, V. 2019. Influence of rainfall infiltration on the stand-up time of unsupported

vertical trenches. Master’s thesis, University of New Brunswick, Canada.

Ileme, V., and Oh, W.T. 2019. Estimating the stand-up time of unsupported vertical

trenches in vadose zone. KSCE Journal of Civil Engineering, 23(10): 4259-4273.

Irvine, D.J., and Smith, R.J.H. 1983. Trenching practice, CIRIA Report 97, London.

Irwin, G.R. (1958) Fracture, Elasticity and Plasticity. Springer, Berlin, 551-590.

Janbu, N. 1968. Slope stability computations. Soil Mechanics and Foundation Engineering

Report. Technical University of Norway, Trondheim.

Kutschke, W.G., and Vallejo, L. 2011. Stability and impacts of unsupported vertical cuts

in stiff clay. Geo-Frontiers Congress 2011: 3619-3628.

Kwan, D. 1971. Observation of the failure of a vertical cut in clay at Welland, Ontario.

Canadian Geotechnical Journal, 8(2): 283-298.

99
Lambe, W., and Turner, K. 1970. Braced excavations. Lateral Stresses in the Ground and

Design of Earth-Retaining Structures: 149-218.

Lee, F.-H., Lo, K.-W., and Lee, S-L. 1988. Tension crack development in soils. Journal of

Geotechnical Engineering, 114(8): 915-929.

Leroueil, S., La Rochelle, P., Tavenas, F., and Roy, M. 1990. Remarks on the stability of

temporary cuts. Canadian Geotechnical Journal, 27(5): 687-692.

Leshchinsky, D., and Zhu, F. 2010. Resultant force of lateral earth pressure in unstable

slopes. Journal of Geotechnical and Geoenvironmental Engineering, 136(12): 1655-

1663.

Li, B., Zhang, F., and Want, D. 2018. Impact of crack on stability of slope with linearly

increasing undrained strength. Mathematical Problems in Engineering,

2018(1096513): 1-11.

Michalowski, R.L. 2013. Stability assessment of slopes with cracks using limit analysis.

Canadian Geotechnical Journal, 50(10): 1011-1021.

National Bureau of Standards. 1988. Recommended technical provisions for construction

practice in shoring and sloping of trenches and excavation, NBS Building Science

Series 127.

Oh, W.T. and Vanapalli, S.K. 2010. Influence of rain infiltration on the stability of

compacted soil slopes. Computers and Geotechnics, 37(5): 649-657.

100
Oh, W.T. and Vanapalli, S.K. 2018. Undrained shear strength of unsaturated soils under

zero or low confining pressures in the vadose zone. Vadose Zone Journal.

Oh, W.T. and Vanapalli, S.K. 2018. Modeling the stress versus settlement behavior of

shallow foundations in unsaturated cohesive soils extending the modified total stress

approach. Soils and Foundations, 58(2): 382-397.

OSHA. Excavations: Hazard Recognition in Trenching and Shoring. OSHA Technical

Manual, Section V, Chapter 2. Washington, DC: U.S. Department of Labor,

Occupational Safety and Health Administration. available at

https://www.osha.gov/dts/osta/otm/otm_v/otm_v_2.html (Accessed in July 26, 2019)

Ping, F., Qingquan, L., Jiachun, L., and Jianping, S. 2005. Numerical analysis of rainfall

infiltration in the slope with a fracture. Science in China Series E-Engineering &

Materials Science, 48: 107-120.

Pufahl, D.E., Fredlund, D.G., and Rahardjo, H. 1983. Lateral earth pressures in expansive

clay soils. Canadian Geotechnical Journal, 20(2): 228–241.

Richard, A. 2018. Estimating the critical height of unsupported trenches in unsaturated soil.

Master’s thesis, University of New Brunswick, Canada.

Richard, A., Oh, W.T. and Brennan, G. 2020. Estimating the critical height of unsupported

trenches in the vadose zone. Canadian Geotechnical Journal (accepted for

publication).

101
Sasekaran, M. 2011. Impact of permeability and surface cracks on soil slopes. Master’s

thesis, University of Manchester.

Schofield, A.N., and Wroth, C.P. 1968. Critical state soil mechanics. McGraw-Hill,

London, England.

Spencer, E. 1968. Effect of tension on stability of embankment. ASCE. 94 SM 5: 1159-

1173.

Spencer, E. 1973. Thrust line criterion in embankment stability analysis. Géotechnique,

23(1): 85-100.

Stanier, S.A., and Tarantino, A. 2013. An approach for predicting the stability of vertical

cuts in cohesionless soils above the water table. Engineering Geology, 158: 98-108.

Sun, D., Wang, L., and Li, L. 2019. Stability of unsaturated soil slopes with cracks under

steady-infiltration conditions. International Journal of Geomechanics, 19(6): 1-13.

Suruda, A., Smith, G., and Baker, S. 1988. Deaths from trench cave-in in the construction

industry. Journal of Occupational and Environmental Medicine, 30(7): 552-555.

Taylor, D.W. 1948. Fundamental of soil mechanics. John Wiley & Sons, New York.

Terzaghi, K. 1943. Theoretical Soil Mechanics. John Wiley & Son, New York.

Thompson, L.J., and Tanenbaum, R.J. 1977. Survey of construction related trench cave-

Ins. Journal of the Construction Division, 103(3): 501-512.

102
Tsidzi, K.E.N. 1997. An engineering geological approach to road cutting slope design in

Ghana. Geotechnical and Geological Engineering, 15 (1), 31-45.

Vahedifard, F., Leshchinsky, B., Mortezaei, K., and Lu, N. 2015. Active earth pressures

for unsaturated retaining structures. Journal of Geotechnical and Geoenvironmental

Engineering, 141(11): 1-11.

Valentin, C., Poesen J., Li, Y. 2005. Gully erosion: impacts, factors and control. Catena.

63(2–3):132–53.

Van Genuchten, M.T. 1980. A closed-form equation for predicting the hydraulic

conductivity of unsaturated soils. Soil Science of America Journal, 44: 892-898

Vanapalli, S.K., Fredlund, D.G., Pufahl, D.E. and Clifton, A.W. 1996. Model for the

prediction of shear strength with respect to soil suction. Canadian Geotechnical

Journal, 33(3): 379-392.

Wang, R., Zhang, G., and Zhang, J. M. 2010. Centrifuge modeling of clay slope with

montmorillonite weak layer under rainfall conditions. Applied Clay Science, 50(3):

386-394.

Wang, Y., and Su, B.Y. 2002. Research on the behavior of fluid flow in a single fracture

and its equivalent hydraulic aperture, Advances in Water Science (in Chinese), 13(1):

61-68.

Wang, Z.F. 2011. Unsaturated hydraulic properties of a single crack and its effects on slope

stability. MPhil thesis, Harbin Institute of Technology, PRC.

103
Wang, Z.F., Li, J.H., and Zhang, L.M. 2012. Influence of cracks on the stability of a

cracked soil slope. In Proceeding of the 5th Asia-Pacific Conference on Unsaturated

Soils, Pattaya, Thailand, 2: 594-600.

Whenham, V., Vos, M.D., Legrand, C., Charlier, R., Maertens, J., and Verbrugge, J.-C.

2007. Influence of Soil Suction on Trench Stability. In Experimental Unsaturated

Soil Mechanics, Springer Berlin Heidelberg: 495-501.

White, F. 2008. National Institute for Occupational Safety and Health (NIOSH) Prevention

through Design (PtD) Workshop Closing Remarks. Journal of Safety Research,

39(2): 203-204.

Zhang, C., Zhao, J., Zhang, Q., and Xu, F. 2010. Unified solutions for unsaturated soil

shear strength and active earth pressure. In Experimental and Applied Modeling of

Unsaturated Soils, ASCE, Shanghai, China: 218-224.

Zhang, J., Zhu, D., and Zhang, Shihua. 2020. Shallow slope stability evolution during

rainwater infiltration considering soil cracking state. Computers and Geotechnics,

117: 1-12.

Zhou, C. 1998. A preliminary study on unsaturated permeability of rock joints. Chinese

Journal of Geotechnical Engineering. 20(6): 1-4

104
CURRICULUM VITAE

Candidate’s full name: Geeshpati Sivarama Sarma Yanamandra

Universities attended: B.Tech. (Civil), Amity University Rajasthan, 2017

Journal publications: N/A

Conference presentations: “Influence of tension cracks on stability of unsupported

trenches in vadose zone unsaturated sand”. In Proceedings of the 72nd Canadian

Geotechnical Conference. St. John’s, Canada. (2019).