Journal of Geotechnical and Geoenvironmental Engineering

UPPER BOUND LIMIT ANALYSIS OF SHALLOW STRIP FOOTING RESTING ON

TWO 1LAYERED SOIL
--Manuscript Draft--

Manuscript Number:

Full Title: UPPER BOUND LIMIT ANALYSIS OF SHALLOW STRIP FOOTING RESTING ON
TWO 1LAYERED SOIL

Manuscript Region of Origin: INDIA

Article Type: Technical Paper

Manuscript Classifications: 13.04.05: Soil mechanics; 13.04.06: Subsurface investigation; 13.05.01: Foundations

Funding Information:

Abstract: The ultimate bearing capacity of shallow substrata is mostly predicated on the
simplified conception that bearing layer is illimitable and homogeneous.. However, in
practice multi layered soils are mainly being used . Resoluteness of reliable ultimate
bearing capacity of substratum in multi layer soil stratum is required in the safety
assessment and design analysis of a substratum structure. In present study kinematic
approach of Limit analysis is used to evaluate the seismic bearing capacity of strip
footing. The bearing capacity of strip footings over a two-layer substructure soil is
considered. The failure mechanism considered was similar to Debnath and Ghosh
(2018) with different wedge angles. The critical wedge angles were found in each case.
Bearing capacity is calculated according to soil resistance beneath the foundation and
a virtual retaining wall method. To determine the bearing capacity of the footing in the
said method, the active and passive forces on the retaining wall are considered equal
along the edges. A MATLAB program was used for calculation and deduction. In order
to study the effects of sundry parameters on two layers of soil, results were compared
with the bearing capacity of the footing on one soil layer in sundry situations. Bearing
capacity charts for footings are presented by varying cohesion in layers,with the same
friction angle and unit weight. The charts are yarely utilizable for any cumulation of
Common.EditSubmissionSteps.Transform.EquationText . The bearing capacity of the
footing was then compared with anterior experimental methods, and the results
obtained were reliable.

Corresponding Author: Litan Debnath

National Institute of Technology Agartala
AGARTALA, Tripura INDIA

Corresponding Author E-Mail: litandbnth4@gmail.com

Order of Authors: Litan Debnath

Suggested Reviewers:

Opposed Reviewers:

Additional Information:

Question Response

The flat fee for including color figures in No

print is $800, regardless of the number of
color figures. There is no fee for online
only color figures. If you decide to not
print figures in color, please ensure that
the color figures will also make sense
when printed in black-and-white, and
remove any reference to color in the text.
Only one file is accepted for each figure.
Do you intend to pay to include color

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
figures in print? If yes, please indicate
which figures in the comments box.

Authors are required to attain permission No

to re-use content, figures, tables, charts,
maps, and photographs for which the
authors do not hold copyright. Figures
created by the authors but previously
published under copyright elsewhere may
require permission. For more information
see
http://ascelibrary.org/doi/abs/10.1061/978
0784479018.ch03. All permissions must
be uploaded as a permission file in PDF
format. Are there any required
permissions that have not yet been
secured? If yes, please explain in the
comment box.

ASCE does not review manuscripts that No

are being considered elsewhere to include
other ASCE Journals and all conference
proceedings. Is the article or parts of it
being considered for any other
publication? If your answer is yes, please
explain in the comments box below.

Is this article or parts of it already No

published in print or online in any
language? ASCE does not review content
already published (see next questions for
conference papers and posted
theses/dissertations). If your answer is
yes, please explain in the comments box
below.

Has this paper or parts of it been No

published as a conference proceeding? A
conference proceeding may be reviewed
for publication only if it has been
significantly revised and contains 50%
new content. Any content overlap should
be reworded and/or properly referenced. If
your answer is yes, please explain in the
comments box below and be prepared to
provide the conference paper.

ASCE allows submissions of papers that No

are based on theses and dissertations so
long as the paper has been modified to fit
the journal page limits, format, and
tailored for the audience. ASCE will
consider such papers even if the thesis or
dissertation has been posted online
provided that the degree-granting

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
institution requires that the thesis or
dissertation be posted.

Is this paper a derivative of a thesis or

dissertation posted or about to be posted
on the Internet? If yes, please provide the
URL or DOI permalink in the comment
box below.

Each submission to ASCE must stand on No

its own and represent significant new
information, which may include disproving
the work of others. While it is acceptable
to build upon one’s own work or replicate
other’s work, it is not appropriate to
fragment the research to maximize the
number of manuscripts or to submit
papers that represent very small
incremental changes. ASCE may use
tools such as CrossCheck, Duplicate
Submission Checks, and Google Scholar
to verify that submissions are novel. Does
the manuscript constitute incremental
work (i.e. restating raw data, models, or
conclusions from a previously published
study)?

Authors are expected to present their No

papers within the page limitations
described in <u><i><a
href="http://dx.doi.org/10.1061/978078447
9018" target="_blank">Publishing in
ASCE Journals: A Guide for
Authors</a></u></i>. Technical papers
and Case Studies must not exceed 30
double-spaced manuscript pages,
including all figures and tables. Technical
notes must not exceed 7 double-spaced
manuscript pages. Papers that exceed the
limits must be justified. Grossly over-
length papers may be returned without
review. Does this paper exceed the ASCE
length limitations? If yes, please provide
justification in the comments box below.

All authors listed on the manuscript must No

have contributed to the study and must
approve the current version of the
manuscript. Are there any authors on the
paper that do not meet these criteria? If
the answer is yes, please explain in the
comments.

Was this paper previously declined or No

withdrawn from this or another ASCE
journal? If so, please provide the previous
manuscript number and explain what you

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
have changed in this current version in
the comments box below. You may
upload a separate response to reviewers
if your comments are extensive.

Companion manuscripts are discouraged

as all papers published must be able to
stand on their own. Justification must be
provided to the editor if an author feels as
though the work must be presented in two
parts and published simultaneously.
There is no guarantee that companions
will be reviewed by the same reviewers,
which complicates the review process,
increases the risk for rejection and
potentially lengthens the review time. If
this is a companion paper, please indicate
the part number and provide the title,
authors and manuscript number (if
available) for the companion papers along
with your detailed justification for the
editor in the comments box below. If there
is no justification provided, or if there is
insufficient justification, the papers will be
returned without review.

If this manuscript is intended as part of a

Special Issue or Collection, please
provide the Special Collection title and
name of the guest editor in the comments
box below.

Recognizing that science and engineering

are best served when data aremade
available during the review and discussion
of manuscripts andjournal articles, and to
allow others to replicate and build on
workpublished in ASCE journals, all
reasonable requests by reviewers
formaterials, data, and associated
protocols must be fulfilled. If you are
restricted from sharing your data and
materials, please explain below.

Papers published in ASCE Journals must Upper bound limit analysis in layered soil is an modern technique to calculate the
make a contribution to the core body of bearing capacity
knowledge and to the advancement of the
field. Authors must consider how their
new knowledge and/or innovations add
value to the state of the art and/or state of
the practice. Please outline the specific
contributions of this research in the
comments box.

When submitting a new and revised d. Some or all data, models, or code generated or used during the study are

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
manuscript, authors are asked to include proprietary or confidential in nature and may only be provided with restrictions.
a Data Availability Statement containing
one or more of the following statements,
with specific items listed as appropriate.
Please select any of the statements below
that apply to your manuscript. Also,
please include the selected statements in
a separate "Data Availability Statement"
section in your manuscript, directly before
the acknowledgements or references. The
statement(s) listed in your manuscript
should match those you select in your
response to this question.

If there is anything else you wish to

communicate to the editor of the journal,
please do so in this box.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
 Click here to access/download;Manuscript;Manuscript.docx

1 UPPER BOUND LIMIT ANALYSIS OF SHALLOW STRIP FOOTING

2 RESTING ON TWO 1LAYERED SOIL
3 Litan Debnath1 and Sima Ghosh2

5 Abstract
6 The ultimate bearing capacity of shallow substrata is mostly predicated on the simplified
7 conception that bearing layer is illimitable and homogeneous.. However, in practice multi
8 layered soils are mainly being used . Resoluteness of reliable ultimate bearing capacity of
9 substratum in multi layer soil stratum is required in the safety assessment and design analysis of
10 a substratum structure. In present study kinematic approach of Limit analysis is used to evaluate
11 the seismic bearing capacity of strip footing. The bearing capacity of strip footings over a two-
12 layer substructure soil is considered. The failure mechanism considered was similar to Debnath
13 and Ghosh (2018) with different wedge angles. The critical wedge angles were found in each
14 case. Bearing capacity is calculated according to soil resistance beneath the foundation and a
15 virtual retaining wall method. To determine the bearing capacity of the footing in the said
16 method, the active and passive forces on the retaining wall are considered equal along the edges.
17 A MATLAB program was used for calculation and deduction. In order to study the effects of
18 sundry parameters on two layers of soil, results were compared with the bearing capacity of the
19 footing on one soil layer in sundry situations. Bearing capacity charts for footings are presented
20 by varying cohesion in layers,with the same friction angle and unit weight. The charts are yarely
21 utilizable for any cumulation of c1 / c2 , D f / B0 ,  1 /  2 , 1 / 2 . The bearing capacity of the footing was
22 then compared with anterior experimental methods, and the results obtained were reliable.

23 KEYWORDS

24 Limit analysis, Pseudo-static, Layered soil, c   soil, Velocity hodograph

25 Introduction

1
Research Scholar, Civil Engineering Department, National Institute of Technology Agartala, PIN-799046,INDIA,
(Corresponding author) Email- litandbnth4@gmail.com,Mobile- (+91) 8732873529
2
Assistant professor, Civil Engineering Department, National Institute of Technology Agartala, PIN- 799046,
INDIA, Email- sima.civil@nita.ac.in, Mobile- (+91) 9862472558
26 One of the important problems in soil mechanics is the ultimate bearing capacity of soils. Even
27 through the problem of punch indentation (or bearing capacity) into a perfectly plastic
28 homogeneous half-space has been described for many decades(Prandtl 1920; Reissner (1924)
29 ;Hill 1950; shield 1954; Chen 1970), effective solutions for layered soils do not exist. Attempts
30 at solving the problem for Mohr-coulomb or Tresca material using the method of characterics
31 were reported by Rychlewski (1966), Michalowski (1983), Mandel and Salencon (1972) . The
32 ultimate bearing capacity of a horizontally layered medium is usually estimated in engineering
33 practice using semiempirical formulas (eg, Brown and Meyerhof 1969; Madej 1969; Chen 1970;
34 Meyerhof and hanna 1978; Hanna and Meyerhof 1979; Hanna 1981). The methods for
35 calculating the bearing capacity of multilayer soils range from averaging the strength parameter
36 Bowles (1988). Using limit equilibrium consideration(Reddy and Srinivasan 1967; Meyerhof
37 1974; Ghazavi and Eghbali 2008; Ghosh and Debnath 2017). Considerations of stability of
38 geotechnical structures include a renewed interest in limit analysis. This is due to new
39 applications, such as reinforced soil (Michalowski,1997), it is also due to development of
40 numerical techniques in limit analysis, such as presented by Tamura et al.(1984), Sloan (1988)
41 and Sloan and Kleeman(1994). The first rational approach to solve bearing capacity problem for
42 a two-layer clay foundation soil was shown by Button (1953), while not indicated at the
43 time,Button’s calculations were equivalent to limit analysis based on the kinematic approach.
44 Close upper-bound and lower bound solutions can be found for ultimate loads over
45 nonhomogeneous soil layers (Mandel and Salencon 1972; Chen and Davidson 1973;Florkiewicz
46 1989; Michalowski and Shi 1993; Michalowski and Shi 1995; Michalowski and Shi 2002).
47 Ahmadi and Kouchaki (2016) investigate the ultimate bearing capacity of two-layered clay.
48 Lotfizadeh and Kamalian (2016) has determined the bearing capacity of strip foundations over
49 two-layered sandy soils. Haghbin (2016) proposed an analytical method to determine the bearing
50 capacity of strip footing resting on granular soil overlaying soft clay. Mosallanezhad and
51 Moayedi (2017) is investigated ultimate bearing capacity of layered soil according to different
52 conventional techniques. While many researches have been done on ultimate bearing capacity of
53 shallow strip footing resting on two layered soil. Most of the works are for the calculation of
54 bearing capacity under static loading condition. Only Debnath and Ghosh (2018), Debnath and
55 Ghosh (2019) has given the solution for seismic loading condition using limit equilibrium
56 method with pseudo-static approach. But Limit Equilibrium method does not give stress strain
57 criteria of the system. So, here an attempt is made to introduce limit analysis method for the
58 evaluation of ultimate bearing capacity of shallow strip footing resting on two-layered c   soil.
59 Attempt is also made to consider the coincident effect of unit weight, surcharge and cohesion.

60 Model Definition

61 Consider a strip footing of width B0 resting on the top of a two layered c   soil (Fig.1). The

62 bearing capacity of a strip footing, qult , determined using the following formulation

1
63 qult   B0 N  , whereas N  is single bearing capacity coefficient for equivalant resistance of
2
64 unit weight, surcharge and cohesion.  is the weighted averaged  1 and  2

65 Analysis Procedure

66 An earthquake can effects on a soil-foundation system by two possible ways. Due to

67 increase the driving forces and reduction of the shearing resistance of soil are the two ways
68 the foundation can fail due to earthquake forces. When the magnitude of the earthquake
69 exceeds a certain limit and the ground conditions are not so much favourable the reduction
70 in the shearing resistance of a soil can only occurs . In present study due to the increase
71 in driving forces is investigated under seismic loading conditions. It has been assumed that
72 the shear strength of the soil constant by the seismic loading. The assumptions made in
73 the analysis are as follows.
74 (a) The soil is homogeneous and isotropic, it is postulated to be an associated flow
75 rule Coulomb material complying with Hill’s maximal work principle. The effect
76 of pore water pressure is not included.
77 (b) A one-sided failure mechanism is surmised to occur.
78 (c) Only the reduction of the bearing capacity due to the increase in driving forces
79 is investigated under seismic loading conditions. The shear vigor of the soil is
80 postulated to remain unaffected by the seismic loading.
81 (d) As was mentioned before, all inertias of the soil-structure system are considered.
82 (e) The earthquake acceleration for both the soil and the structure is assumed to be
83 the same.
 84 (f) The earthquake load on the structure is represented by the base shear load
85 acting at the substratum level . The moment due to the seismic load on the structure
86 is not considered. Only the base shear load will be taken into account.
87 (g) The water table is assumed to be well below the footing.
88 (h) The failure mechanism consists of two active wedge angle and passive wedge angle,
89 which are admitted as the variable of this present analysis.
90 To determine the bearing capacity coefficient for a strip foundation located on a two-layered
91 soil by the limit analysis method, collapse mechanism of the problem is depicted in Fig.2 (a), (b).
92 Parameters involved in the present study are as follow:
93 c1 = cohesion of soil in top layer, c2 = cohesion of soil in bottom layer, 1 = angle of friction of

94 soil in top layer, 2 = angle of friction of soil in bottom layer,  1 = unit weight of soil in top layer,

95  2 = unit weight of soil in bottom layer,  A1 = Angle of slip surface at top layer in active zone,
96  A 2 = Angle of slip surface at bottom layer in active zone,  p1 = Angle of slip surface at top

97 layer in passive zone,  p 2 = Angle of slip surface at bottom layer in passive zone,  i = Friction

98 angle along surface between active and passive zone at ith layer. The thickness of the second
99 layer contributing the failure wedge h2 is expressed as:
tan  A 2
100 h2  B0 tan  A 2  h1
tan  A1 (1)
101 h1 is the thickness of top layer
102 Upper bound theorem of analysis:

103 This theorem says, if a compatible mechanism of plastic deformation ij p* , Vij p* is assumed,

104 which satisfies the condition Vij p*  0 on the displacement boundary S v , then the load Ti , Fi

105 determined by equating the rate at which the external forces do work at the rate of internal
106 dissipation of energy will be either higher or equal to the actual limit load, and this can be
107 written as

 ij p* dv   TV
i i ds   FV
p* p* p*
108 ij i i dv (2)
v s v
109 The collapse mechanism was chosen to comprise 2 trapezoidal block ABED, BDKJ and 2
110 triangular block EDF, DKF as shown in Fig.2(a), (b). At collapse, the footing and the underlying
111 rigid block ABED, EDF were assumed to move in phase with absolute velocities V1 ,V2 , whereas

112 V3 ,V4 are the absolute velocities of triangular and trapezoidal block EDF, BDKJ. v21 is the

113 relative velocity of the block EDF with respect to the block ABED. V32 is the relative velocity of

114 the block EDF with respect to the block EDF. V34 is the relative velocity of the block BDKJ with

115 respect to the block FDK. V41 is the relative velocity of the block BDKJ with respect to the block
116 AEDB. The interface BD, DF, ED, DK, as well as the lines AE, EF, FK, KJ were treated as
117 velocity discontinuity lines. The direction V1 ,V4 ,V2 ,V3 makes an angle 1 , 2 with the

118 corresponding velocity discontinuity lines. Whereas, V21 , V34 makes an angle 2  1 with the

119 discontinuity lines ED, DK. V41 , V32 makes an angle 1 ,  2 with discontinuity lines GF, FC.

120 Deformation pattern in present analysis is taken 1  2 . It is considered velocities at weak soil
121 layer is higher . The velocity hodograph is shown in Fig.3.

122 Representation of all velocities with respect to V1 (Fig 3)

123 From the hodograph all the relative velocities can be represented as initial velocity v 1 by usine
124 lemis sine formula. All the velocities with respect to initial velocity v1 is shown as follows:

sin  1 + 2  21 
V2  V1
125 sin   2  1  (3)

sin  1   2 + 2  1 
126 V21  V1 (4)
sin   2  1 

cos   2 + 2   2 
127 V3  V2 (5)
cos   3   2  2 

sin   2   3 
128 V32  V2 (6)
cos  3   2  2 

sin  3 + 1 
129 V4  V3 (7)
sin   4  1  2 
 sin   4   3 + 1  2 
130 V43  V3 (8)
sin  2  1   4 

V41  (V43  V21 ) cos(2  1 )  V32 sin  2   (V43  V21 )sin(2  1 )  V32 cos  2  
2 2
131 (9)
 

132 Incremental External work

133 The different elements of the incremental external work can be calculated as follows:

134 (a) Incremental external work due to the foundation load and corresponding inertia forces are

135 WP1  PL B0 sin   A1  1 1  kv   cos   A1  1  kh  V 11 (10)

WP2  PL1 ED sin   A2  2 1  kv   cos   A2  2  kh  V 2

136 (11)
 PL1  B0  h1 cot  A1  sin   A2  2 1  kv   cos   A2  2  kh  V 2

137
(b) Incremental external work due to the surcharge loading and the corresponding inertia forces

WQ1  Q1   sin   P1  1 1  kv   cos   P1  1  kh  V 4

138 (12)
  1D f  h1 cot  P1  h2 cot  P 2   sin   P1  1 1  kv   cos   P1  1  kh  V 4

WQ2  Q2   sin   P 2  2 1  kv   cos   P 2  2  kh  V3

q  h cot  P1  h2 cot  P 2 
2

139  1 1   sin   P 2  2 1  kv   cos   P 2  2  kh  V3 (13)

 h1  h1 cot  P1  h2 cot  P 2  
 1 D f  h1 cot  P1  h2 cot  P 2 
2

   sin   P 2  2 1  kv   cos   P 2  2  kh  V3

 h1  h1 cot  P1  h2 cot  P 2 

140 (c) Incremental external work due to self weight and inertia force of trapezoidal ABDE

WA  WA sin   A1  1 1  kv   cos   A1  1  kh  V1

141 2 B0  h1 cot  A1 (14)
 h1 1 sin   A1  1 1  kv   cos   A1  1  kh  V1
2

142 d) Incremental external work due to self weight and inertia force of triangle BDE
 WB  WB sin   A 2  2 1  kv   cos   A 2  2  kh  V2
143 B0  h1 cot  A1 (15)
 h2 2 sin   A2  2 1  kv   cos   A2  2  kh  V2
2

144 (e) Incremental external work due to self weight and inertia force of trapezoidal ABDE

WC  WC   sin   P1  1 1  kv   cos   P1  1  kh  V4

145 h1 cot  P1  2h2 cot  P 2 (16)
 h1 1   sin   P1  1 1  kv   cos   P1  1  kh  V4
2

146 (f) Incremental external work due to self weight and inertia force of triangle CDE

WD  WD   sin   P 2  2 1  kv   cos   P 2  2  kh  V3

147 h22 cot  P 2 (17)
  2   sin   P 2  2 1  kv   cos   P 2  2  kh  V3
2

148 The total incremental external work is the summation of these six contributions; i.e; equations
149 (10),(11), (12),(13), (14),(15), (16),(17)

150  W  ext
 W p1  W p2  WQ1  WQ2  WA  WB  Wc  WD (18)

151 Incremental internal Energy dissipation

152 The different elements of the incremental internal energy dissipation can be calculated as
153 follows:

154 (i) Along AE

DAE  c1 AE cos 1V1

155 (19)
 c1 h1 cos ec A1 cos 1V1

156 (ii) Along EF

DEF  c2 EF cos 2 v2
157 (20)
 c2 h2 cos ec A2 cos 2 v2

158 (iii) Along FK

 DFK  c2 FK cos 2V3
159 (21)
 c2 h2 cos ec p 2 cos 2V3

160 (iv) Along KJ

DKJ  c1 KJ cos 1V4

161 (22)
 c1 h1 cos ec p1 cos 1V4

162 (v) Along DF

DDF  c2 DF cos  2 V32

163 (23)
 c2 h2 cos  2 V32

164 (vi) Along BD

165 DBD  c1 h1 cos 1V41 (24)

166 The total incremental energy dissipation is the summation of these six parts: that is, Equations
167 (19),(20), (21),(22), (23),(24)

168  D  D AE  DEF  DFK  DKJ  DDF  DBD (25)

169
Since the soil mass was considered to be cohesive-friction (c-ϕ), the rate of total external work
170
dobe by all the body and the total incremental internal energy dissipation should become equal.
171 The collapse load is pL per unit length of the footing due to the component of soil unit weight,
172 surcharge and cohesion can be expressed by using the following expression

1
173 pL   N  (26)
2

174 N  is a single coefficient for coincident effect of cohesion, unit weight and surcharge. In seismic

175 condition N  is expressed as N E and static condition as N  S

176 The most critical N  values is obtained by minimization pL with respect to the variables  A1 ,

177  A 2 ,  p1 and  p 2 angles describing the collapse mechanism.

178  is delimited as the equivalent unit weight demoing the proportion of each layer in the rupture
179 zone and evinced by

A1 1  A2 2
180  
A1  A2 (27)

181 Where A1 and A2 are the effective area of each layer in breach zone and dependent on h1 and h2

182 Unity coefficient is a function of several parameters including cohesion, surcharge and unit
183 weight. It can be expressed

a b 2c d1 
184 N    1  1   (28)
 e1 e1  B0 e1 

185 Where, c is defined as the weight averaged cohesion showing the proportion of each layer in the
186 rupture slip line and shown by

c1h1  c2 h2
187 c (29)
h1  h2

188

  c1 h1 h2   h h  h  h  
    2  1 cot  A1  1  1  1 cot  A1  2  
2c  2c    c2 B0 B0   B0  B0  B0  B0  
189   2 
 B0   2 B0    h1 h2   h1  h1  1  h1  h2  
    2  cot  A1    1  cot  A1  
  B0 B0   B0  B0  2  B0  B0  

190 (30)

191 The detail equations for a1 , b1 , d1 and e1 are given in “ Appendix”

192 Results and Discussion

193 Since the heuristic algorithms give us low ramification and high execution and these methods are
194 relatively new can be applied in the geotechnical problem. Out of these methods a brief
195 discussion on particle swarm optimization is given here as it is used in the analysis.
196 Particle swarm optimization

197 Particle swarm optimization (PSO) is a robust stochastic optimization technique based on
198 movement and intellegence of swarms. PSO applies the concept of social interaction to problem
199 solving. It was first introduced by Kennedy and Eberhart (1995). It uses a number of agents
200 (particles) that constitute a swarm moving around in a search space looking for the best
201 solutions. Each particle is treated as a point in N dimentional space which adjusts its “flying”
202 according to its own flying experience of other particles. Let X and V denote a particle’s
203 position and velocity in a search space. Then the ith particle can be interpreted as
204 X i ( xi1 , xi 2 , xi 3, ................, xid ) and the velocity of the ith particle is delimited by

205 Vi   vi1 , vi 2 , vi 3 ,................, vid  , where d comprises the dimension of the problem. The best

206 previous particle of the ith particle is recorded and represented as

207 Pi   pi1 , pi 2 , pi 3 ,....................., pid  , the index of the best particle among all the particles, is

208 comprised by Pg   pg1 pg 2 pg 3 ........... pgd  . The velocity and position of each particle can be

209 wangled according to the following equation:

210 Vid   Vid  c1  rand   Pid  X id   c1  rand   Pgd  X id  (31)

211 X id  X id  Vid (32)

212 Whereas c1 and c2 are position constants known as acceleration coefficient,  is the inertia

213 weight coefficient; rand is a random number within the range  0, 1 . In present analysis  is

max  min
214 chartered by  ( gn)  max   gn , where gn is the generation. The flowchart
NI
215 describing the various steps involved in execution of the PSO algorithm is shown in Fig.4.

216 Results of the calculations are presented in a dimensionless form in Fig. 5-8. All diagrams
217 represents least upper bounds obtained from analyses based on mechanisms in Fig.2 (a)
218 whichever yields the minimum. The independent angles describing the geometry of the
219 mechanism were varied in an optimization scheme.

220 PSO Applications in Geotechnical Engineering

221 Involution of analysis of geotechnical deportment is due to multivariable dependencies of soil
222 and rock replications. Most of the materials that geotechnical engineers deals with show dubious
223 deportment in consequence of the involute formation of these materials. Consequently, in some
224 geotechnical engineering quandaries, the objective function is non-convex and discontinuous.
225 Consequently, simple optimization techniques may have difficulties in finding the ecumenical
226 optimum solution due to get trapped in local solutions. To surmount this inhibition, utilizing a
227 potent optimization method to obtain the ecumenical optimum solution is of interest. In recent
228 years, soft computing techniques have been widely used to prognosticate geotechnical
229 parameters (Sharma et al. 2017a, b, c; Singh et al. 2017; Sharma and Singh 2017). Accordingly,
230 as a potent optimization technique, PSO has entered in the field of geotechnical engineering to
231 solve its quandaries. In the following sections, a review of the PSO applications in geotechnical
232 engineering is presented.

233 Pile and Foundation Design

234 In General, the application of PSO in pile and foundation design is mainly in predicting the
235 behavior of axially loaded piles and ultimate bearing capacity of foundations. In addition, PSO
236 has been used to estimate the secant pile construction time. A hybrid PSO-BP model was
237 developed by Ismail and Jeng (2012) to predict the relationship of load/ settlement in a single
238 pile. They considered pile settlement as a non-linear function of related parameters as follow:
239 S  f ( P , k s , E p , D, L )

240 in which S, P, Ks, Ep, D and L are pile settlement, pile load, soil stiffness, pile modulus, pile
241 diameter and pile length, respectively. A database of 92 static pile loading tests on concrete piles
242 was used to conduct the modeling. The results of their study show high values of correlation
243 coefficient between predicted settle-ments obtained by PSO-ANN model and measured
244 settlements obtained by static load tests. In the field of geotechnical engineering, construction
245 time is a critical factor for control the cost and planning in construction projects. Regarding this,
246 Chen et al. (2009) compared the construction time of a secant pile wall utilizing two optimization
247 methods including self-organizing map predicated optimization (SOMO) and PSO. A database
248 consists of 207 primary and secondary bored piles for a secant pile wall was utilized in this
249 study. The results of the study revealed that the total preserved time for the SOMO and PSO was
250 27.21 and 23.79 h respectively. They concluded that the SOMO method yields better
251 construction compared to the PSO method.
252 Rock Mechanics
253 The application of PSO in rock mechanics included but not constrained to determining the
254 roughness profiles of rocks, back analysis of geomechanical parameters, analysis of uniaxial
255 compressive vigor (UCS), identifying the structure of rocks, and apperceiving the structure of
256 altered rocks. An incipient method predicated on hybrid PSO algorithm and multi-layer
257 perceptron (MLP) neural network was developed by Babanouri et al. (2013) to estimate fractal
258 dimension of roughness profiles (D) and standard deviation (r) of rock. It is worth noting that
259 resoluteness of D is still a quandary in geotechnical engineering due to attributed different values
260 of the fractal dimension. There are two kinds of errors for prognostication of fractal dimension;
261 stochastic and systematic. Modelling of these errors is difficult because of the intricacy
262 relationship between the fractal dimension and the quantifiable variables. An immensely colossal
263 number of fractional Brownian including 39900 profiles was engendered and their statistical
264 features were extracted. They analyzed this model for 10 standard profiles of roughness and
265 concluded that the proposed estimator gives an error 15 times more than the roughness-length
266 method. Zhao and Yin (2009) proposed an incipient keenly intellective displacement back
267 analysis method predicated on the amalgamation of fortification vector machine, PSO, and
268 numerical analysis.
269 Slope Stability Analysis
270 Slope instability is one of the main concerns in geotechnical engineering. It has been identified as
271 one of the most frequent natural disaster in mountainous regions that can lead to solemn
272 economic loss, property damage, and communication passage interruption (Singh et al. 2016;
273 Sharma et al. 2017a). More importantly, it is an incessant cause of suffering because it puts
274 human life in hazard (Sharma et al.2017b). A single factor or more likely a cumulation of
275 different factor including slope geometry, properties of slope composing material, groundwater
276 condition, structural discontinuity, development of impuissant zones, disruption in geological
277 formation, and cumbersomely hefty rainfall can initiate slope failure (Umrao et al. 2017). The
278 application of PSO in slope stability analysis is mainly within the framework of limit equilibrium
279 method for soil slopes (Kalatehjari and Ali 2013). PSO is mainly used to the second step, since
280 the shape and location of the critical slip surface is generally unknown in soil slopes (Bolton et
281 al. 2003).
282 Soil Mechanics
283 PSO have successfully encountered with some problems in soil mechanics such as determination
284 of soil erosion characteristics, behavior of unsaturated soil, soil-structure interaction, and soil
285 parameters. Yunkai et al. (2010) predicted soil erosion characteristics in small watersheds using
286 a combination of SVM and PSO. In fact, they introduced the application of PSO for automatic
287 selection of SVM parameters and presented a model by linking PSO and SVM for small sample
288 data analysis. According to the results, the proposed model can simulate successfully the erosion
289 characteristics in small watersheds with low degree of average error (3.85%). Zhang et al. (2009,
290 2013a, b) utilized a hybrid moving boundary PSO (HMPSO) method to minimize the difference
291 between measured (field data) and computed values on the cavity pressure-cavity strain curve in
292 unsaturated soil. They used the HMPSO algorithm to select parameter values in the Barcelona
293 Basic Model (BBM) which is one of the best known constitutive models for unsaturated soils.
294 Parametric Study

295 The parametric study is done for the variation of pseudo-static seismic bearing capacity
296 coefficients with the different soil parameter as shown in Fig.9-14

Df
297 Variations of Seismic Bearing capacity Coefficient for Different Valuesof using Particle Swarm
B0
298 Optimization Algorithm:

2
299 Figure 9 depicts the variation of Seismic Bearing capacity Coefficient  N  E  at 2  300 ,  2 
2
1  k  h 2c c
300 ,  0.8 , 1  0.8 , kv  h , 1  0.8 , 1  0.1 , 2  0 , 1  0 with kh . From the plot it is seen
2 2 2 2 B0  2 B0 c2

Df
301 that, N E increases with the increase in the value of . It is additionally conspicuous as increase in
B0
302 depth increases the confinement between the soil grains of the substructure.

h1
303 Variations of Seismic Bearing capacity Coefficient for Different Valuesof using Particle Swarm
B0
304 Optimization Algorithm:
305 Figure 10 shows the variation of Seismic Bearing capacity Coefficient N E at 2  300 ,

2 1  k  Df 2c c
306 2  ,  0.8 , 1  0.8 , kv  h , 1  0.8 ,  0.5 , 2  0 , 1  0 with kh . From the plot
2 2 2 2 2 B0  2 B0 c2

h1
307 it is seen that, N E decreases with the increase in the value of . Here, h1 is the depth of the top layer
B0
308 and it is considered in the analysis that it is weaker than the bottom layer. So, more impotent
309 layer will provide less resistance and hence increase of the thickness of this layer decreases the
310 value of bearing capacity coefficient.

1
311 Variations of Seismic Bearing capacity Coefficient for Different Values of using Particle Swarm
2
312 Optimization Algorithm:

313 Figure 11 shows the variations of Seismic Bearing capacity Coefficient N 

E at 2  300 ,

2 h1  k  Df 2c c
314 2  ,  0.25 , 1  0.8 , kv  h , 1  0.8 ,  0.5 , 2  0 , 1  0 with kh . From the
2 B0 2 2 2 B0  2 B0 c2

1
315 plotit is seen that,coefficient N E increases with the increase in the value of as one would expect.
2
1
316 Here, ratio is increased keeping  2 as constant.
2
1
317 Variations of Seismic Bearing capacity Coefficient for Different Values of using Particle Swarm
2
318 Optimization Algorithm:

319 Figure 12 shows the variation of Seismic Bearing capacity Coefficient N E at 2  300 ,

2 h1  k  Df 2c c
320 2  ,  0.25 , 1  0.8 , kv  h , 1  0.8 ,  0.5 , 2  0 , 1  0 with kh .From the
2 B0 2 2 2 B0  2 B0 c2

1
321 plotit is seen that, coefficient N E increases with the increase in the value of . Here increase in 1
2
1
322 value is made keeping the  2 value as constant. So, obviously due to the increase in ratio, the
2
323 value N E will increase.
324 Variations of Seismic Bearing capacity Coefficient for Different Values of kv using Particle Swarm

325 Optimization Algorithm:

2 h1   
326 Figure 13 shows the variation of N E at 2  300 ,  2  ,  0.25 , 1  0.8 , 1  0.8 , 1  0.8
2 B0 2 2 2
Df 2c2 c
327 ,  0.5 ,  0 , 1  0 with kh . From the plot it is seen that, N E decreases with the increase in
B0  2 B0 c2

328 kv . It is obvious because increase in the value of kv will increase the disturbance of base soil and this
329 result decrease the value of N E

c1
330 Variations of Seismic Bearing capacity Coefficient for Different values of using Particle Swarm
c2
331 Optimization Algorithm:

2 h1  k 
332 Figure 14 shows the variation of N E at 2  300 ,  2  ,  0.25 , 1  0.8 , kv  h , 1  0.8 ,
2 B0 2 2 2

Df 2c2
333  0.5 , 0 with kh . From the plot, it is seen that the coefficient N E increases with the
B0  2 B0
c1 c
334 increase in the value of . Here, 1 ratio is increased while keeping c2 as constant. So, obviously due
c2 c2
c1
335 to the increase in ratio, the value N E will increase.
c2

336 Comparison with experimental results and published solutions

337 The aim of the comparison is to validate the computational methodology with experimental
338 results, subsisting semi-empirical solutions and later determine the reasons abaft the observed
339 discrepancies.With known formulation for bearing capacity coefficient N  , a computer

340 programing software ‘MATLAB’ code has been developed utilizing PSO algorithm which is
341 able to calculate the ultimate bearing capacity, qult , for various combinations of soil properties in
342 each layer. Table 1 presents the comparison of the avarage limit pressure calculated utilizing the
343 proposed method to the experimentals tests performed by Purushothamaraj et al. (1973), Carlson
344 and Fricano (1961), Koizumi (1965), Desai and Reese (1970). From this table it has been
345 observed the present analysis getting virtually more proximate value with other researchers.
346 Table 2-4 shown the Tabular representation of the seismic bearing capacity coefficient for
347 different thickness to footing width ratio (h1 / B0 ) . From these value it has been observed that
348 values obtained by present method is lower than Debnath and Ghosh (2018). It withal observe
349 that limit analysis method are signaficantly different (higher value) from Debnath and Ghosh
350 (2018). This optical discernment is due to the reason that Debnath and Ghosh (2018) analyzed
351 the bearing capacity using limit equilibrium method. The advantage of the upper-bound
352 solution to the limit equilibrium solution is that the kinematics are completely verified
353 and the solution is obtained for an associated flow-rule material. Table 5 has shown a
354 comparison between present analysis with Eskavari et al. (2019). From Table 5 it is visually
355 perceived that present analysis getting higher upper bound values than Eskavari et al. (2019).
356 The bearing-capacity factors Nc*(qult/c1) for strip, circular, and square footings on two-layered
357 clays are presented in Table 6. The bearing-capacity factor Nc* was found to be virtually
358 identical for square and circular footings for both the soft-over-vigorous strongover-soft and
359 layer cases. In all cases, the bearing-capacity factor for circular footings was around 1-2% above
360 that of a square footing. As discovered by Merifield et al. (1999), Merifield and Nguyen (2007)
361 all of the analyses herein denote that an intricate relationship subsists between the optically
362 canvassed modes of shear failure and the ratios c1/c2 and h1/B0. The modes of failures can best be
363 described as being either ‘general shear’, ‘partial punching shear’, or ‘full punching shear’ akin
364 to that described by Merifield et al. (1999) , Merifield and Nguyen (2007) . When comparing the
365 results for square and circular footings to those for strip footings (present analysis), in general it
366 was found that the bearing-capacity factors Nc* for square and circular footings were much more
367 proximate for strip footings.To demonstrate the potentiality of the present analysis, Finite
368 element (FE) analysis has been compared with present values. FE analyses have been performed
369 by Ghazavi and Eghbali (2008) utilizing PLAXIS. The physical and mechanical properties of
Df
370 soil for each layer are depicted in Table 8, while ratios for different geometries are evinced
B0
371 in Table 7. Table 9 compares present analysis of values with analytical and FE values obtained
372 by Ghazavi and Eghbali(2008). Values obtained by present method is higher than FE analysis for
373 (2  330 , 1  300 ) , (2  370 , 1  340 ) . As visually perceived, the results obtained from the
374 present analysis relatively well commensurable when B0= 2m. As optically discerned, the results
375 obtained from the present analysis relatively well commensurable.

376 Conclusions
377 A rigorous upper bound solution was presented for the bearing capacity of a two-layer
378 substratum soil. The simplicity of the solution to the relatively perplexed quandary of the bearing
379 capacity of footings over two different soils was achieved by introducing a failure mechanism
380 where the velocity discontinuities were bent at specific angles at the interface between layers.
381 This allowede one to construct a simple hodograph. The solution is predicated on the classical
382 approach in which the geometry of hodograph is modified in search of the least upper bound.
383 Optimization of the geometry of the mechanism led to the least upper bound. This approach
384 seems to be very efficacious, and it yields results that, in some regions of parameters, are lower
385 than those engendered by the numerical technique predicated on linear programming (Debnath
386 and Ghosh; 2018). An attempt has also made to evaluate seismic bearing capacity for
387 simultaneous resistance of unit weight, surcharge and cohesion. On the basis of analysis it is seen
388 seismic bearing capacity value decreases if the values of horizontal and vertical seismic
389 acceleration coefficients are increased. Comparison of present analysis with finite element
390 analysis and other researchers has done. It is obvious from the comparative study the present
391 analysis gives closer value with other reseachers. The bearing capacity factor N  for the case of
392 two layer soil with constant friction angle with cohesion varying has been presented in the form
393 of design charts which are readily usable in Practical field.

394 Data availability statement

395 The authors confirm that some data and code generated during this study are proprietary or
396 confidential in nature and may only be provided with restrictions e. g.,

397 (a) MATLAB Code

398 (b) Particle Swarm optimization code
399 (c) Limit Analysis hodograph
400 (d) Conversion of Relative velocity
401 (e) Graphical representation of all parameters
402 References

403 Ahmadi, M. M., and Kouchaki, B. (2016). “New and simple equations for ultimate bearing
404 capacity of strip footings on two layered clays: Numerical study.” Int. J. Geomech., DOI:
405 10.1061/ (ASCE) GM.1943-5622.0000615.

406 Babanouri N, Nasab SK, Sarafrazi, S. (2013). “A hybrid particle swarm optimization and multi-
407 layer perceptron algorithm for bivariate fractal analysis of rock fractures roughness”. Int J Rock
408 Mech Min Sci 60:66–74. doi:10.1016/j.ijrmms. 2012.12.028.
409 Banks, A., Vincent, J., and Anyakoha, C. 2007. “A review of particle swarm optimization.” Part
410 I: Background and development. Nat Comput. 6, 467–484
411 Bolton H, Heymann G, Groenwold, A. (2003). “Global search for critical failure surface in slope
412 stability analysis”. Eng Optim 35(1):51–65.
413 Bowles (1988). “Foundation analysis and design.”4th Ed. McGraw Hill, New York,N.Y.

414 Brown and Meyerhof.(1969). “An experimental study of ultimate bearing capacity of layered
415 clay foundations. “Proc., 7th Int. Conf. on Soil Mechanics and Foundations Engineering,
416 Socieded Mexicana de Mecanica de Suelos, Mexico city, 45-51.

417 Button (1953). “The bearing capacity of footings on a two layer cohesive subsoil.”
rd
418 Proceedings, 3 International Conference on soil Mechanics and Foundation Engineering, 111:
419 712-729

420 Carlson, E. D., and Fricano, S. P. (1961). “ Tank foundation in eastern Venezuela.” J. Soil Mech.
421 Found. Div. ASCE, 87, 69-90.

422 Chen (1975). “Limit analysis and soil plasticity.” Elsevier,Amsterdam,The Netherlands.

423 Chen and Davidson, H. l. (1973). “Bearing capacity determination by limit analysis,” J.soil
424 Mech. Found. Div. 99(6), 433-449

425 Chen J, Yang L R, Su M. C. (2009). “Comparison of SOM-based optimization and particle

426 swarm optimization for minimizing the construction time of a secant pile wall”. Autom Constr
427 18(6):844–848. doi:10.1016/j.autcon.2009.03.008.
428 Clerc, M., and Kennedy, J. (2002). “The particle swarm explosion, stability, and convergence in
429 a multidimensional complex space”. IEEE Trans EvolComput, 10.1109/4235.9856926, 58–73.
430 Debnath and Ghosh (2017). “Seismic bearing capacity of shallow strip footing with coloumb
431 failure mechanism using limit equilibrium method”. Geotech. and Geol. Engg., DOI:
432 10.1007/s10706-017-0268-y.

433 Debnath and Ghosh (2018).“ Pseudo-static analysis of shallow strip Footing resting on Two-
434 Layered Soil.” Int. J. Geomech.,DOI: 10.1061/(ASCE)GM.1943-5622.0001049.Div., 107(7),
435 915–927.

436 Debnath and Ghosh(2019). “Pseudo-static bearing capacity analysis of Shallow strip footing over
437 two-layered soil considering punching shear failure.” Geotech. and Geol. Engg., DOI:
438 10.1007/s10706-019-00866-5.
439 Desai, C. S., and Reese, L. (1970). “ Ultimate capacity of circular footing on Layered soils.” J.
440 Indian Nat. Soc. Soil Mech. Found. Eng. 96(1), 41-50.
441 Eberhart and Shi(2001). “Particle swarm optimization : Developments,applications and
442 resources.” Proc.,2001 congress on Evolutionary Computation, 81, IEEE,New York, 81-86.
443
Eshkevari, S. S., Abbo, J. A. and Kouretzis, G. (2019). “ Bearing capacity of strip footings on
444
Layered sands”. Computers and Geotechnics., DOI: 10.1016/j.compgeo.2019.103101.
445 Florkiewicz.A.(1989). “Upper bound to bearing capacity of layered soils.” Can. Geotech. J.
446 26(4). 730-736

447 Ghazavi and Eghbali (2008).”A simle limit equilibrium approach for calculation of ultimate
448 bearing capacity of shallow foundations on two-layered granular soils.” Geotech. Geol.Eng.,
449 26(5),535-542

450 Ghazavi, M., and Eghbali, A. H. (2008).“A simple limit equilibrium approach for calculation of
451 ultimate bearing capacity of shallow foundations on two-layered granular soils.”Geotech. Geol.
452 Eng., 26(5), 535-542.
453 Haghbin, M. (2016). “Bearing capacity of strip footings resting on granular soil overlaying soft
454 clay.” Int. J. Civ. Eng., DOI`: 10.1007/s40999-016-0067-5.

455 Hanna and Meyerhof ((1980). “Design charts for ultimate bearing capacity of foundations on
456 sand

457 Hanna(1980). “Design charts for ultimate bearing capacity of foundations on sand overlaying
458 soft clay. Can Geotech J 17: 300- 303

459 Hill (1950). “ The mathematical theory of plasticity.” Clarendon press,Oxford, United kingdom.
460 Ismail A, Jeng, D. S. (2012). “Empirical method for settlement prediction of single piles using
461 higher order neural network and particle swarm optimization”. In: Geo Congress 2012, State of
462 the Art Practice in Geotechnical Engineering, pp 285–294.
463 Kalatehjari, R. (2013). “An improvised three-dimensional slope stability analysis based on limit
464 equilibrium method by using particle swarm optimization”. Dissertation, Universiti Teknologi
465 Malaysia
466 Kennedy and Eberhart (1995) . “Particle swarm optimization.” Proc., IEEE Int. Conf. on Neural
467 Networks, IEEE, New York, 4 , 1942-1948.

468 Koizumi, Y. (1965). “ Discussion on session 4-Division 3.” Proc. Sixth Int. Conf. Soil Mech.
469 Found. Engg., Montreal, Que. 3. pp. 413-415.

470 Lotfizadeh, R. M., and Kamalian, M. (2016).“Estimating Bearing Capacity of Strip Footings
471 over Two-Layered sandy soils using the characteristic Lines Method”. Int. J. Civ. Engg.,
472 10.1007/s40999-016-0015-4.

473 Mandel and Salencon(1972). “ Force portante d’un sol surune assise rigide 9etude theoretique).
474 Geotechnique, 22: 79-93

475 Merifield, R.S., and Sloan, S.W. (1999).“Rigorous plasticity solutions forthebearing capacity of

476 two-layered clays.”Geotechnique,49(4),471-490.

477 Merifield, R.S., and Nguyen, Q. V. (2006).“ Two- and three-dimensional bearing-capacity
478 solutions for footings on two layered clays.”Geomech. and GeoEngg.,1(2),151-162.
479 Meyerhof and Hanna .(1978). “Ultimate bearing capacity of foundation on layered soils
480 underinclined load.” Can. Geotech.J. 15(4).565-572

481 Michalowski (1983). “On the solution of plane flow of granular media for jump non-
482 homogeneity.” International journal for Numerical and Analytical methods in
483 Geomechanics,7:485-492

484 Michalowski (1995). “Bearing capacity of footings overtwo layer foundation soils.” J.of
485 Geotech.Eng, ASCE,121(5),421-428

486 Michalowski and Shi.(1993). “Bearing capacity of non homogeneous clay layers under
487 embankments.” J.Geotech. Engrg.ASCE.119(10), 1657-1669

488 Michalowski, R .L, Shi, L. (1995). “Bearing Capacity of Footings over two layers foundation
489 soils.” J. Geotech. Engg. Vol. 5 ,No.421, pp.421-428.
490
Michalowski, R. L. (2002). “Collapse loads over two-layer clay foundation soils.” Soils and
491
Foundation, vol. 42, No.1, pp.1-7.

492 Michalowski. R.L (1993). “Limit analysis of weak layers under embankments.” Soils and
493 Fond..33(1). 155-168

494 Mosallanezhad, Mand Moayedi, H (2017).“Comparison Analysis of Bearing Capacity

495 Approaches for the Strip Footing on Layered Soils.”Arab J SciEng.42(9),3711–3722.
496 overlying soft Clay.” Can.Geotech.J..17(2).300-303

497 Prandtl (1920). “Uber die Harte plastischer korper.” Nachr. Ges.wissensch, Gottiingen, math-
498 phys. Klasse. 74-85 (in German)
499
Purushothamaraj, P., Ramiah, K. B. and Venkatakrishna, N. K (1973). “ Bearing Capacity of
500
Strip Footings in Two Layered Cohesive-friction Soils”. Can. Geotech. J. Vol.11, pp. 32-45.

501 Reddy and Srinivasan (1967).“Bearing capacity of footings on layered

502 clays.”J.soil.Mech.Found.Div.ASCE. 93(2). 83-99

503 Reissner (1924). “Zum Erddruckproblem.” Proc, First Int. Congr for Appl.Mech. C.B.Biezeno
504 and J.M.Burgers.Eds. Technische Boekhandel en Drukkerii J. Waltman Jr. The
505 Netherlands,Delft .295-311

506 Sharma L K, Singh T. N (2017). “Regression-based models for the prediction of unconfined
507 compressive strength of artificially structured soil”. In: Engineering with Computers, pp 1–12.
508 Sharma LK, Singh R, Umrao RK, Sharma KM, Singh TN (2017a), “Evaluating the modulus of
509 elasticity of soil using soft computing system”. Eng Comput 33(3):497–507.
510 Sharma LK, Umrao RK, Singh R, Ahmad M, Singh T N. (2017b). “Stability investigation of hill
511 cut soil slopes along national highway 222 at Malshej Ghat, Maharashtra, India”. J Geol Soc
512 India 89(2):165–174.
513 Sharma LK, Umrao RK, Singh R, Ahmad M, Singh T N. (2017c). “ Geotechnical
514 characterization of road cut hill slope forming unconsolidated geo-materials: a case study”.
515 Geotech Geol Eng 35(1):503–515.
516 Shi and Eberhart (1998). “A modified particle swarm optimizer.” Proc., IEEE World congress
517 on Computational Intelligence, IEEE, New York.

518 Shield (1954) , “Plastic potential theory and Prandtl bearing capacity solution .” J.Appl.Mech.,
519 21(2), 193-194
520 Singh R, Umrao RK, Ahmad M, Ansari MK, Sharma LK, Singh T. N. (2017). “Prediction of
521 geomechanical parameters using soft computing and multiple regression approach”.
522 Measurement 99:108–119
523 Singh T.N, Singh R, Singh B, Sharma LK, Singh R, Ansari M K. (2016). “Investigations
524 stability analyses of Malin village lslide of Pune district, Maharashtra, India”. Nat Hazards
525 81(3):2019–2030
526 Sloan, S. W. (1988). “Lower bound limit analysis using finite elements and linear
527 programming”. Int. J. Num. Analyt. Mech. Geomech., 12, 61-67.
528
Sloan, S. W. and Kleeman, P. W. (1994). “Upper bound limit analysis using discontinuous
529
velocity fields, Res. Report No. 096.05.1994, University of Newcastle, Australlia.
530 Umrao R. K, Singh R, Sharma L. K, Singh T. N. (2017). “Soil slope instability along a strategic
531 road corridor in Meghalaya, northeastern India”. Arab J Geosci. doi:10.1007/s12517017-3043-8
532 Yunkai, L., Yingjie T, Zhiyun O, Lingyan W, Tingwu X, Peiling Y, Huanxun, Z. (2010).
533 “Analysis of soil erosion characteristics in small with particle swarm optimization, support
534 vector machine and artificial neuronal networks”. Environ Earth Sci 60(7):1559–1568. doi:10.
535 1007/s12665-009-0292-1.
536 Zhang, Y., Gallipoli, D., Augarde, C (2009). “Parallel hybrid particle swarm optimization and
537 applications in geotechnical engineering”. In: Advances in computation and intelligence.
538 Springer, Berlin, pp 466–475.
539 Zhang, Y., Gallipoli, D., Augarde, C. (2013a). “Parameter identification for elasto-plastic
540 modelling of unsaturated soils from pressuremeter tests by parallel modified particle swarm
541 optimization”. Comput Geotech 48:293–303. doi:10. 1016/j.compgeo.2012.08.004.
542 Zhang, Y., Xiong, X., Zhang, Q. (2013b). “An improved self-adaptive PSO algorithm with
543 detection function for multimodal function optimization problems”. Math Probl Eng. doi:10.
544 1155/2013/716952.
545 Zhao, H. B, Yin, S. (2009). “Geomechanical parameters identification by particle swarm
546 optimization support vector machine”. Appl Math Model 33(10):3997–4012. doi:10.
547 1016/j.apm.2009.01.011
548

549 Appendix
  h h 
a1   2  1 cot  A1  1 1 sin   A1  1 1  kv   cos   A1  1  kh  V1
 B0  B0 
 h h 
 1  1 cot  A1  2 2 sin   A 2  2 1  kv   cos   A 2  2  kh  V2
550
 B0  B0 
h h h 
  1 cot  P1  2 2 cot  P 2  1 1   sin   P1  1 1  kv   cos   P1  1  kh  V4
 B0 B0  B0 
h 2 cot  P 2  2
 2   sin   P 2  2 1  kv   cos   P 2  2  kh  V3
B0  

    D f  h1 h2 
b1  2  1    cot  P1  cot  P 2    sin   P1  1 1  kv   cos   P1  1  kh  V4
    B0  B0 B0 
2
  1   D f  h1 h2 
    B  B cot  P1  B cot  P 2 
551
   0  0    sin    1  k  cos    k  V
2 0
  P 2 2  v   P2 2  h  3
 h1 h1 h2 
  cot  P1  cot  P 2 
 B0 B0 B0 

 c1   h1   c2   h2  
    cos ec  A1 cos 1       cos ec  A 2 cos 2 V2 
 c   B0   c   B0  
 
 c2   h2   c1   h1 
552 d1      cos ec  P 2 cos 2 V3      cos ec  P1 cos 1 V4 

  c   B0   c   B0  
 
  c2   h2   c1   h1  
   c   B  cos  2 V32   c   B  cos 1 V14 
   0    0  

e1  sin   A1  1 1  kv   cos   A1  1  kh  v 11

1  h1 
1  cot  A1  sin   A2  2 1  kv   cos   A2  2  kh  v2
553 
h  B0 
1 1
B0

554

555
 Click here to access/download;Table;Table.docx

Table 1. Comparison with Field and Experimental values

Bearing
pressure(tons/ft2)
Source Footing type
d/b c2/c1 ϕ(degrees) Observed Present
method
Carlson and Fricano Circular tank(150ft dia) 0.087 0.18 1.37 5.55
(1961)
Koizumi (1965) Circular(2 in. dia.) 0.8 0.33 0.0 6.00b 4.88
Desai and Reese Circular(3 in. dia.) 0.55 1.80 0.0 2.50b 3.85
(1970b)
Purushothamaraj et Strip footing 0.49 0.50 2.0 0.46 1.22
al. (1973)
bFrom plate load test

Table 2. Comparison with Analytical values

Debnath and Present analysis
Ghosh(2018)
kh=.0.1 Kh=0.1
ϕ2 ϕ1/ϕ2 δ2 δ1/δ2 ɣ1/ɣ2 Kv h1/B0=0.1 h1/B0=0.1
Df/B0 Df/B0
0.25 0.5 1 0.25 0.50 1
0 14.63 19.94 30.55 14.87 20.73 31.23
30 0.8 15 0.8 0.8 kh/2 14.33 19.82 29.48 14.41 20.02 30.30
kh 13.95 19.22 29.18 13.99 19.98 30.28
0 15.39 21.77 34.22 16.18 22.92 35.02
30 1 1 1 kh/2 15.29 21.70 33.87 16.97 22.54 34.73
kh 14.74 21.06 32.93 15.54 22.19 33.98
 Table 3. Comparison with Analytical values
Debnath and Present analysis
Ghosh(2018)
kh=.0.1 Kh=0.1
ϕ2 ϕ1/ϕ2 δ2 δ1/δ2 ɣ1/ɣ2 Kv h1/B0=0.25 h1/B0=0.25
Df/B0 Df/B0
0.25 0.5 1 0.25 0.50 1
0 11.89 16.49 26.79 12.84 17.31 27.88
30 0.8 15 0.8 0.8 kh/2 11.58 15.90 25.97 12.69 16.86 26.54
kh 11.02 15.63 24.95 12.24 16.66 25.55
0 12.29 18.29 28.31 13.68 19.01 29.45
30 1 15 1 1 kh/2 11.98 17.40 28.04 13.47 18.12 28.73
kh 11.62 16.52 27.60 13.42 17.82 28.05

Table 4. Comparison with Analytical values

Debnath and Present analysis
Ghosh(2018)
kh=.0.1 Kh=0.1
ϕ2 ϕ1/ϕ2 δ2 δ1/δ2 ɣ1/ɣ2 Kv h1/B0=0.50 h1/B0=0.50
Df/B0 Df/B0
0.25 0.5 1 0.25 0.50 1
0 9.37 13.95 22.42 9.34 14.15 22.35
30 0.8 15 0.8 0.8 kh/2 9.24 13.60 21.44 9.28 13.88 21.74
kh 9.11 13.47 21.08 9.96 13.66 21.14
0 9.76 14.34 23.70 10.69 16.16 23.81
30 1 15 1 1 kh/2 9.75 14.18 23.26 9.89 15.81 23.99
kh 9.56 14.00 22.72 9.72 15.73 22.90
Table 5. Comparison with numerical values
Ultimate Bearing capacity qu (kPa)
Eskavari et al. (2019) Present Analysis
h1/B0 ϕ2=27.5, ϕ2=30, ϕ2=32.5, ϕ2=35, ϕ2=27.5, ϕ2=30, ϕ2=32.5, ϕ2=35,
ϕ1=40 ϕ1=40 ϕ1=40 ϕ1=40 ϕ1=40 ϕ1=40 ϕ1=40 ϕ1=40
0.6 L.B 360 510 700 970
415.54 541.77 801.377 915.89
U.B 400 530 720 1050
0.9 L.B 400 540 730 1020
468.765 554.15 874.41 1073.90
U.B 440 560 800 1150
1.2 L.B 470 560 780 1100
536.58 672.70 995.94 1437.72
U.B 500 620 870 1200
1.5 L.B 520 600 800 1200
572.814 706.612 1077.117 1623.35
U.B 540 680 910 1260
Table. 6. Comparison of bearing capacity factor (Nc*) with present analysis

UB (
Merifield and
Merifield Present
Nguyen (2006)
et al. Study
Displacement FEM
1999)
h1/B0 c1/c2 Strip Square Circular Strip
Footing Footing Footing Footing
0.2 8.55 7.96 7.95 8.6243
0.25 8.55 7.96 7.95 8.6243
0.4 8.55 7.96 7.95 8.6243
0.5 8.55 7.96 7.89 8.6243
0.8 6.36 6.73 6.85 6.8934
1 5.32 5.95 6.05 6.6538
0.125
1.25 4.52 5.11 5.27 5.5429
1.5 3.93 4.53 4.66 4.9067
2 3.09 3.73 3.85 4.4638
2.5 2.61 3.21 3.32 4.337
4 1.82 2.33 2.41 2.500
5 1.55 2 2.07 2.4432
0.2 6.52 6.35 6.36 6.8208
0.25 6.52 6.35 6.36 6.8208
0.4 6.52 6.35 6.36 6.8208
0.5 6.52 6.35 6.36 6.8208
0.8 6.25 6.27 6.34 6.9867
1 5.32 5.95 6.05 6.570
0.25
1.25 4.6 5.45 5.59 6.1215
1.5 4.08 5.03 5.17 5.3611
2 3.34 4.39 4.51 4.9751
2.5 2.88 3.92 4.02 4.7776
4 2.12 3.04 3.13 3.6041
5 1.85 2.7 2.78 2.8726
0.2 5.49 5.96 6.04 6.3481
0.25 5.49 5.96 6.04 6.3481
0.4 5.49 5.96 6.04 6.3481
0.5 5.49 5.96 6.04 6.3481
0.8 5.49 5.96 6.04 6.3481
1 5.32 5.96 6.05 6.3481
0.5
1.25 4.94 5.94 6.04 6.2816
1.5 4.48 5.82 6.02 5.9871
2 3.89 5.46 5.9 5.8864
2.5 3.47 5.08 5.58 5.2293
4 2.74 4.22 5.23 4.5314
5 2.44 3.89 4.39 4.30
 TABLE 7.Various soil and footing
Geometries for comparative studies
B0 q Overburden h1/B0
Footing Pressure (KN/m )
2

Width
(m)
1 10 For all three
2 17.5 cases h1/B0
varies from 0.1
3 25
up to 1

TABLE 8.Properties of soil of each layer used in comparative studies

Case Friction angle Unit weight Elasticity Poisson's

φ γ (KN/m3) modulus E ratio ν
(KN/m2)
1 30 19 17,500 0.333
2 31 19.3 20,000 0.327
3 33 19.9 25,000 0.313
4 34 20.1 27,500 0.306
5 36 20.5 35,000 0.291
6 37 20.7 40,000 0.285
7 39 20.9 50,000 0.27
8 42 21.1 65,000 0.249
TABLE 9. Comparisons of bearing capacity results obtained from present analysis with other authors and
PLAXIS analysis

No h1/B0 Ghazavi Plaxis Present

and
Eghbali
(2008)
B0=1m 1 0.25 475 572 538.07
2 0.5 450 498 500.11
3 0.75 424 456 480.10
4 1 396 385 469.78
5 1.25 368 382 457.4
B0=2m 1 0.25 904 992 1040.74
2 0.5 856 860 940.05
3 0.75 806 835 835.47
4 1 754 776 819.37
5 1.25 699 766 779.56
B0=3m 1 0.25 1333 1403 1599.0
2 0.5 1263 1210 1108.9
3 0.75 1189 1263 1083.34
4 1 1111 991 1064.03
5 1.25 1029 969 946.34
 Click here to access/download;Figure;Figure.docx

q1   1 D f

c1
Top soil layer 1
1

c2
Bottom soil layer 2
2
FIG.1. Geometry of footing on two-layered soil profile

(a)
(b)
Fig.2. Collapse mechanism of present model

Fig.3. Velocity hodograph of choosen collapse mechanism

 Initialize Particles

Calculate fitness values for

each particles

Is Current fitness
value better than
pBest?
Yes no

Assign current Keep previous pBest

fitness as new pBest

Assign best particle’s pBest value

to gBest

Calculate velocity for each particles

Use each particle’s velocity value to

update it’s data value

No Target or maximum Yes

epochis reached?

End
Fig.4. Flowchart of PSO
 30

28 

26 

24 

22 
20

18
N

16

14

12

10

0.2 0.4 0.6 0.8 1.0

h1/B0

Fig.5. Bearing capacity factor N for  khkv=kh/2

0.2 0.4 0.6 0.8 1.0

38 38
36  36

34 
34

32 32

30  30

28  28
26 26
24 24
22 22
N

20 20
18 18
16 16
14 14
12 12
10 10
8 8
6 6
0.2 0.4 0.6 0.8 1.0
h1/B0
 
Fig.6. Bearing capacity factor N for   kh=0.2, kv=kh/2
 36
34 


32 

30 

28


26
24
N 22
20
18
16
14
12
10
8
0.2 0.4 0.6 0.8 1.0
h1/B0

Fig.7. Bearing capacity factor N for  kh=0.2, kv=kh/2

50
48 
46


44 
42 
40 
38 

36
34
32
30
28
N

26
24
22
20
18
16
14
12
10
0.2 0.4 0.6 0.8 1.0
h1/B0

Fig.8. Bearing capacity factor N for  kh=0.2, kv= kh/2
 32

30
Df/B0=0.25
28 Df/B0=0.50
Df/B0=0.75
26
Df/B0=1
24

22

20

18
N
16

14

12

10

6
0.10 0.15 0.20 0.25 0.30

kh

Fig.9. Variation of N with kh for  /2,
/=0.8,/=0.8,kv=kh/2,h1/B0=0.25,2c2/B0c1/c2=0.8

20

19 h1/B0=0.1
18 h1/B0=0.25
17 h1/B0=0.5
16

15

14

13
N

12

11

10

6
0.10 0.15 0.20 0.25 0.30

kh


Fig.10. Variation of N with kh for  /2,
/=0.8,/=0.8,kv=kh/2,Df/B0=0.50,2c2/B0c1/c2=0.8
 20



18


16

14

N
12

10

8
0.10 0.15 0.20 0.25 0.30

kh

Fig.11. Variation of N with kh for  /2,
c/c=0.8,/=0.8,kv=kh/2,Df/B0=0.50,2c2/B0h1/B0=0.1

20

/=0.6
18 /=0.8
/=1

16
N

14

12

10
0.10 0.15 0.20 0.25 0.30
kh

Fig.12. Variation of N with kh for  /2,
/=0.8, Df/=0.8,kv=kh/2,h1/B0=0.25,2c2/B0c1/c2=0.8
 24

23
kv=0
22
kv=kh/2
21
kv=kh
20

19

18

17

16

N
15

14

13

12

11

10

9
0.10 0.15 0.20 0.25 0.30

kh

Fig.13. Variation of N with kh for  /2,
c/c=0.8,/=0.8,,Df/B0=0.50,2c2/B0h1/B0=0.25

20

c1/c2=0.6
18
c1/c2=0.8
c1/c2=1

16
N

14

12

10
0.10 0.15 0.20 0.25 0.30

kh

Fig.14. Variation of N with kh for  /2,
/=0.8,/=0.8,kv=kh/2,Df/B0=0.50,2c2/B0h1/B0=0.25
h1/h2/=0.3,h1/h2/=0.16,vs1/vs2vp1/vp2=0.8