Documente Academic
Documente Profesional
Documente Cultură
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering1
Madhira R Madhav2
Key words Abstract: The paper revisits the basic tenets of Geotechnical Engineering as propounded by
Consolidation, non-linear theory, its founders in the early years such as (i) rigid plasticity, (ii) leaning instability, (iii) thin
consolidation by PVDs, bearing capacity, versus thick layer response, (iv) inelastic response, (v) stress or strain induced non-
leaning instability, uncoupled moduli, homogeneity, etc. Conventional approaches for the estimation of the ultimate capacities of
strain induced non-homogeneity both shallow and deep foundations use limit equilibrium methods and are based on only
the strength parameters apart from the geometry of the foundation element. Ground/soil
being a much more complex material than metals from which the original theories have
been developed, requires the consideration of stiffness or compressibility as well as the
strength parameters for the estimation of ultimate loads. The results for the bearing
capacity of shallow foundations, the ultimate axial capacity of piles and for leaning
instability of tall structures show dependence on the stiffness of the ground. The
conventional one-dimensional consolidation theory developed by Terzaghi is based on linear
void ratio-effective stress relationship and for thin layers only. A theory of non-linear one-
dimensional consolidation of a thick clay deposit considering linear void ratio-log effective
stress relationship that incorporates the variation of initial in-situ effective stress with depth
indicates that the conventional thin layer theory underestimates the degree of consolidation
and overestimates the degree of dissipation of excess pore pressures. The degree of
settlement and degree of dissipation of pore pressures are sensitive to the magnitude of
applied load unlike in the conventional thin layer theory. The effects of the non-linear theory
for radial flow in to vertical drains has similar effects on degree of settlement, degree of
dissipation of excess pore pressure and variation of excess pore pressures with respect to
time, radial distance and depth. An interesting outcome is the induced vertical flow in a
radial flow problem. Response of ground is inelastic and moduli of deformation in
compression and on unloading are significantly different from each other. The moment
influence coefficient considering inelastic response of ground (as part of the foundation
gets unloaded due to moment loading), for different rectangular, circular and annular
foundations is a function of the relative modulus, Rk, the ratio of subgrade moduli in
unloading to that in compression. Estimation of distribution of displacements and stresses
in the soil mass resulting from the loads applied at or below the surface is one of the basic
There is Life in the Ground: challenges in foundation engineering/soil-foundation interaction studies. Estimating the
it goes into the seed and it also, parameters E and G from by the direct shear test, a new simple two parameter model based
on uncoupled deformation, E, and shear, G, moduli, is proposed for analysing stresses and
when stirred up, goes into the displacements under a rigid or uniformly loaded circular footing on a finite layer. The elastic
continuum and Vlasov models are obtained as particular cases of the proposed solution. An
man who stirs it. analysis that incorporates strain dependency of deformation modulus of soils leads to strain
C. D. Warner induced non-homogeneity for the ground. Results obtained highlight the importance of
considering strain induced non-homogeneity on settlement of foundations.
1 IGS Ferroco Terzaghi Oration 2010 delivered on 4 October 2010 at Bangalore, India
2 Emeritus Professor, JNT University Hyderabad, India, Email: madhavmr@gmail.com
208
Indian Geotechnical Journal 40(4), 2010
Stress, MPa
The stress – strain relationships of various 300 Fe250
engineering materials are shown in Figures 1(a) and (b)
in linear and log-log scales. Soils are so weak and highly 200
deformable that their stress – strain curves for are not Dense sand
visible when plotted on a linear scale in comparison with Loose sand
100 Stiff Clay
those of steel and concrete but become discernable if M50
M15 Soft Clay
plotted on log-log scale. The ultimate strength and 0
stiffness of steel are respectively 415 MPa and 210 GPa
0.00 0.05 0.10 0.15 0.20 0.25 0.30
while those of soils possibly range between 10 (soft Strain
clay) to 400 kPa (dense sand) and 0.5 to 10 MPa
a) Natural
a) Scale
Natural Scale
respectively. Thus these plots contrast the significantly
large differences in the strength and stiffness of soils in
comparison with those of steel and concrete.
1000 Fe415
M50 M15
Prandtl’s theory originally developed for metals 10
with compressive and tensile strengths of nearly the
Stress, MPa
Dense sand
same magnitude, i.e. with friction angle equal to zero, is 1
the starting point for the estimation of bearing capacity Loose sand
of shallow foundations. Terzaghi (1943) modified the 0.1 Stiff Clay
same and proposed his famous theory for strip footing
resting on cohesive-frictional soils (Figure 2) an active 0.01
Soft Clay
wedge defined by the angle of shearing resistance, ,
and the passive wedges defined by the angle, (/4-/2) 0.001
both with respect to the horizontal. Prandtl’s solution
0.0001 0.001 0.01 0.1 1
modified for c- soils (c being the cohesion) with the Strain
active wedge defined by (/4+/2) instead is adopted b) Log-Log Scale
as the basis for the estimation of ultimate bearing b) Log-Log Scale
capacity, qult, of foundations, as
Fig. 1 Stress-Strain Variations for
qult = cNcsc + qNq + 0.5BNs (1) Various Engineering Materials
pl = h + cu Nc* (2)
For Ir<Ir(cr);
1 c/D=0.5
B
1 1.5 4.4 0.6 L tan '
Rq 2 F c Fqc exp (6)
(3.07sin ')(log2Ir )
0.5 1 sin '
1
45 40 30
1 =20° 25
qu/qur
30
= 20° 35
Rq 40
0.5 45
0.75
0.5 0
0 1.5 3 1 10 100 1,000
Irr Ir
Fig. 7 Variation of Rq with Irr and for Fig. 8 Variation of Ultimate Bearing Capacity Ratio qu /qur
dVp = 0.3 and c/D = 0.5 with Ir and for B/L=0, c'/B=0, D/B = 1
211
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
qu/q
General shear failure is indicated for Ir values
greater than the limiting values given above. Figures 9, 0.3 c’/γB = 0, 0.25,
10 and 11 show the variations in the ratio, qu/qur, with Ir 0.5, 0.75 and 1
and c'/B for B/L = 0, D/B = 1 and = 20°, 30° and
45° respectively. The ratio, qu/qur, is very small, of the
order of 0.1 or less for very compressible soils and
increases gradually with Ir till the ratio attains the value
of 1.0 corresponding to general shear failure at the 0
limiting value of Irr given above. The limiting value of Ir is
1 10 Ir 100 1000
independent of c/B for a given value of . The ratio,
qu/qur, is nearly independent of c'/B for = 45°. Fig. 11 Variation of qu/qur, with Ir and c'/B
Similarly the ratio, qu/qur, is not very sensitive for B/L = 0, = 45° and D/B = 1
(Figure 12) to the ratio D/B varying from 0.5 to 3
especially at higher Ir value. 1
qu/qur
0.5
1
0.25
c/B = 0 3
0.5
qu/qur
D/B = 0.5
1 0
0.5
1 10 Ir 100 1000
0.5 the ground is more for soils with higher friction angles.
qpu/qpur values for all values of converge at higher
values of Ir. The variation of qpu/qpur with Ir is
independent of D/B (Figure 17) in the range 10 to 100
for = 30° and c'/γB = 0.5. c'/γB has no effect
0 (Figure 18) on the variation of qpu/qpur with Ir for =
30° and D/B =10. Figure 19 shows the variation in
1 10 100 1000 qpu/qpur with relative stiffness or rigidity factor, Ir, and
Ir
volumetric strainfor c'/γB = 0.5, =30° and
Fig. 14 Variation of qu/qur with Ir and B/L D/B =10. qpu/qpur increases with increase in Ir for any
for D/B = 1, , c'/B = 0
1 1
1
0.5
qu/qur
qp/qpr
0.2 = 20°
0.5 30°
B/L = 0 0.5 40°
50°
0 0
1 10 100 1000 1 10 100 1000
Ir Irr (=Ir)
Fig. 15 Variation of qu/qur with Ir and B/L Fig. 16 Variation of Ultimate Point Bearing Capacity Ratio
for D/B = 1, and c'/B = 0.5 qpu/qpur with Ir and for D/B = 10, c'/B = 0.5
213
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
∆
1
3%
qpu/qpu
2%
0.5 = 1% W
h
0 GL
1 Ir 10 100
ks
The restoring moment, dMt, generated due to the undrained strength. According to conventional theories,
variation in foundation pressure can be written as the bearing capacity of the foundation of the tower was
the same for the three cases considered. If instability is
dMt = ksIe. d (10) governed by bearing capacity failure of the foundation,
all the three analyses indicate the weight of the
where Ie is the second moment of the area of the structure to be the same for the three cases.
foundation about horizontal axis through the centroid. At
limiting equilibrium, W = 0, and the following However, if the rotation of the tower is plotted
relationship is derived (Figure 22) against its weight, the effect of the relative
stiffness of the ground with respect to the undrained
he.we/re2 = 1 (11) strength becomes significant. The weights of the tower
for failure are of the order of 65, 120 and 135 MN for
where he is the limiting height, we-average G/cu values of 10, 100 and 1,000. Failure is very abrupt
settlement and re=Ie/A, A-area of the foundation. Thus for very stiff soils compared to that of relatively softer
the structure starts to lean (leaning instability) when the soils.
product of the height to the centre of gravity and the
average settlement, we (=W/A.ks), equals the radius of
gyration, re (=Ie/A). If the structure is loaded non- 4
centrally, an upper bound limit load, Wy, can be G/c u=10 G/c u=100
estimated in a similar manner, as
3
Wy = ks.Iy/hy (12)
Rotation, deg.
2
where Iy is the second moment of the area of the
foundation about horizontal axis at edge of foundation
and hy – is the limiting height of the tall structure. Eqs. 1 G/c u=1000
(11) and (12) demonstrate the effect of the stiffness of
the ground on the instability of tall structures.
0
Lancellotta (1993) uses the concept of non-linear
0 50 100 150
moment restraint to explain the leaning instability.
Weight of Tower, MN
Potts (2003) models the Leaning Tower of Pisa (Figure
21) a typical tall structure, as resting on a uniform Fig. 22 Rotation with Increase in Weight of Tower
deposit of clay defined by its undrained strength, cu, and (after Potts 2003)
shear stiffness, G, with an initial imperfection of 0.50 tilt.
The undrained strength of the clay was assigned a value
of 80 kPa while the value of shear modulus, G, was Effect of Thickness of Layer on
aried as equal to 10, 100 and 1,000 times the Consolidation
One Dimensional Problem
Formulation
Cv, m2/s
Considering linear e-log σ΄ relationship,
assuming (1+e) constant during consolidation and
constant coefficient of consolidation, cv, (Figure 24) the
following equation of consolidation can be derived (Ayub
Khan et al. 2010) following Davis and Raymond (1965)
in non-dimensional form as
'v, kPa
w w 1 w
2
Fig. 24 Variation of Coefficient of Consolidation (cv)
(13)
T Z 2 10w. Z with Effective Vertical Stress 'v)
(Madhav and Miura, 2004)
where w log10 , ’= ('z+q-u) and ’ is the
.H H
u ( z, t )
submerged unit weight of the soil, Z and T are the Up 1 dz (15)
normalized parameters; Z=z/H and T=cv t/H2, σ΄ is the 0 uo
effective vertical stress; t - time; u - the excess pore
water pressure and z- depth from surface. where uo = q the initial excess pore pressure.
Two boundary conditions, viz., (i) pervious top Eq. (13) is solved numerically using the finite difference
impervious base (PTIB) and (ii) pervious top pervious approach. The layer was divided in to hundred sub-
base (PTPB) are considered. Average degree of layers and convergence ensured. The average degree of
Settlement, Us, is settlement for the entire thickness, Us, by the proposed
non-linear theory, is presented in Figures 25 and 26 for
H
PTPB and PTIB cases respectively and compared with
log( ) dz the conventional one-dimensional consolidation theory
Us 0 o
(14) by Terzaghi. The degree of consolidation for a thick layer
H
decreases with increase of q* at all times. The results
0 log( of ) dz from the proposed theory agree with the conventional
thin layer theory for q* (=q/(γ’H)) ≥ 10,000 (for H
Average degree of dissipation of excess pore tending to zero, thin layer). The degree of consolidation
pressure, Up, is is relatively more in the case of thick layer of clay
compared to thin layer for a given load intensity.
216
Indian Geotechnical Journal 40(4), 2010
Similar observations are made by Gibson et al. clay layer is shown in Figures 27 and 28 for different
(1981). The degree of consolidation increases from 51% values of q* for PTPB and PTIB respectively along with
to 60% (Figure 25 for PTPB) and from 53% to 71% the results from conventional thin layer theory. The
(Figure 26 for PTIB), for q* decreasing from 10,000 to degree of dissipation of pore pressure from non-linear
1, at a time factor, TV of 0.197. The theory of theory is slower than the degree of settlement (Davis
consolidation for thin layers thus underestimates the and Raymond, 1965, Gibson et al., 1981, and Xie and
degree of settlement. Leo, 2004). While the degree of settlement is 51% the
corresponding degree of dissipation of pore pressure is
The excess pore pressures are computed at only 8 % for PTPB (Figure 27) and 53% and 8.30%
different depths and times and the average degree of respectively for PTIB (Figure 28), for q* of 10,000 in
dissipation of pore pressures for the entire thickness of thick layers against a value of 50% from linear theory, at
0 0
q’=1000
100
q’=1
20 10 20
10
100
q’=1000 q’=1
40 40
Up, %
Us
60 60
PTPB
PTPB
80 80 Proposed Theory
Proposed Theory
Terzaghi’s Theory
Terzaghi’s Theory
100 100
0.0001 0.001 0.01 0.1 1 10 0.0001 0.001 0.01 0.1 1 10
Tv Tv
Fig. 25 Degree of Consolidation vs Time Factor (PTPB) Fig. 27 Average Degree of Dissipation of Pore Pressures
vs. Time Factor (PTPB)
q’=1000
20 100
10
Up, %
40 q’=1
Us, %
60
PTIB
80 Proposed Theory
Terzaghi’s Theory
100
Tv
0.0001 0.001 0.01 0.1 1 10
Tv Fig. 28 Average Degree of Dissipation of
Fig. 26 Degree of Consolidation vs Time Factor (PTIB) Pore Pressures vs Time Factor (PTIB)
217
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
Z/H
The excess pore pressure is relatively large or P TP B; q’=10
0.5
dissipation of pore pressure is relatively slow at all times NL Theory 0.80
according to non-linear theory of consolidation 0.6
Terzaghi
compared to the conventional theory. Interestingly, the
0.80
isochrones in the case of PTPB are slightly 0.7
unsymmetrical about the mid depth in contrast to
symmetrical isochrones in the conventional linear theory 0.8
for PTPB boundary conditions. The residual pore
pressures are 82% and 78% of q, at depths of 0.2H 0.9
and 0.8H respectively for q* of 10 at a time factor of
0.20. The corresponding residual pore pressure is 46% 1
of q at the two depths from linear theory.
Fig. 29 Excess Pore Pressure Isochrones (PTPB)
0.80
0.5
stress relationship and constant coefficients of volume
compressibility (mv) and horizontal permeability (kh). 0.6
However, for a relatively large applied stress range the 0.80
void ratio is not proportional to effective stress and the 0.7
0.20
coefficients of compressibility and permeability 0.8
decrease during consolidation. Richart (1957), Basak P TIB; q’=10
NL Theory
and Madhav (1978), Hansbo (1979, 1981), Vaid 0.9
Terzaghi
(1985), Lekha et al. (1998), Teh and Nie (2002),
1
Indraratna et al. (2005a and b), Rujikiatkamjorn and
Indraratna, (2006) and Conte and Troncone (2009),
Fig. 30 Excess Pore Pressure Isochrones (PTIB)
present more realistic theories of consolidation for
radial flow.
Hansbo (1979, 1981) presented a simple (2003) along with the evaluation of their effectiveness
solution for radial consolidation with band shaped in practice. Indraratna et al. (2005a and b) developed a
vertical drains by transforming the drain into an theory for consolidation with radial flow using e - log
equivalent circular one. Full-scale tests conducted by
(Cc and Cr) & e – log kh (Ck) relationships and for
Bergado et al. (2002) on soft Bangkok clay with PVDs
revealed that the degree of consolidation obtained from different loading increment ratios ( / i) . Zhuang et
pore pressure measurements is lower than the al. (2005) presented a semi-analytical solution for
corresponding values obtained from settlement vertical consolidation of clays with variable
measurements. The various modeling aspects of PVDs compressibility and permeability. The variation of
are comprehensively discussed by Indraratna et al. horizontal permeability along the radial direction and the
218
Indian Geotechnical Journal 40(4), 2010
actual cross sectional shape of band drains are triangular patterns respectively, where S is the spacing
considered in 3D FEM analysis of soft soil consolidation of drains. The equivalent diameter of the drain, dw, is dw
improved by PVDs (Rujikiatkamjorn and Indraratna, =2(a+b)/, where ‘a’ and ‘b’ are the width and thickness
2006). Sathananthan and Indraratna (2006) formulated of the PVD respectively.
equivalent plane-strain consolidation equations for
radial consolidation in soft soils with PVDs and found Formulation
good agreement between equivalent plane-strain and
axi-symmetric solutions. Indraratna et al. (2007) made a
The general equation of non-linear consolidation
critical review of analytical solutions and numerical
with radial flow (Ayub Khan et al. 2009) in terms of a
analysis of soft clay stabilization with PVDs beneath
parameter, w, is
road and railway embankments. Three and two-
dimensional multi-drain finite-element analyses of a
case study of a combined vacuum and surcharge w 2w 1 w
cr 2 (16)
preloading with vertical drains is presented by t r r r
Rujikiatkamjorn et al. (2008). A new technique is
developed by Indraratna et al. (2008) to model ( u)
consolidation by vertical drains beneath a circular where w log10 or w log10 f , where f is
f f
loaded area by transforming the system of vertical
drains into a series of concentric cylindrical drain wall. the final effective stress. ‘w’ varies with depth in thick
Conte and Troncone (2009) developed a simple deposits of clay as the initial effective in-situ stress, σ’o
analytical solution for radial consolidation under time and the final effective stress, σ’f (=σ’o +q0) vary with
dependent loading. Walker et al. (2009) presented the depth due to overburden stresses (Figure 32), where qo
spectral method for analysis of vertical and radial is the applied load intensity. Therefore, the thick clay
consolidation in multilayered soils with PVDs assuming layer of thickness, H is divided equally into m thin layers
constant soil properties within each layer. A theory of of thickness, ΔH=H/m (Figure 33) and the differential
non-linear consolidation for radial flow around a vertical equation governing the consolidation process of each
drain in a thick deposit of clay is developed herein layer can now be written as:
based on the non-linear theory of consolidation for
vertical flow presented by Davis & Raymond (1965). w j 2w 1 w j
cr 2 j (17)
t r r r
Vertical Drains
where the subscript j refers to the layer number and wj =
PVDs which are band or strip shaped about 100 wj (r, t) of the jth layer. Even though the initial and final
mm wide and 4 mm thick, are usually installed in square effective stresses are different in each layer, the flow in
or triangular arrays (Figure 31a). The strip drain and the each layer is assumed to be purely radial and
zone of influence of each drain are replaced by independent of the flows in the adjacent layers.
equivalent circular shapes and the flow pattern around
the drain studied considering the flow to be axi-
symmetric (Figure 31b). The equivalent diameter of the
influence zone, de, is 1.13S & 1.05S for square and
Surcharge, qo
d
dee SS HH H Thick Clay
11 11
Drain
Drain
C c , cr
z z
Fig. 31 (a) Triangular Arrangement of PVDs and Fig. 32 (a) Thick Clay Loading Condition; (b) Initial in-situ
(b) Flow in Unit Cell Stress (’o) Distribution; c) Final Stress (’f) Distribution
219
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
Surcharge, qo
Drain
Pervious
zj
H
σ́o,j & σ́f,j
Layer- j H/m
Layer- m
Impervious
rw
re
Centre Line of Drain
re
UP , j (1 U *avg , j ) (24)
(e
rw
o e) 2 r dr
US , j re (18) Average degree of dissipation of excess pore
(eo ef ) 2 r dr
rw
pressure for the entire thickness is
m
where w j log10
j
, w o, j log10
o , j
and the initial or
U
j 1
P, j .(H) j
f, j f, j UP (21)
H
original effective in-situ stress at the layer j,
o , j .z j , zj the depth of the layer j. Eqn. (13) is identical in form to that of the
conventional Barron’s radial consolidation theory and
Average degree of consolidation for the entire can be solved in the same way with the boundary
thickness, conditions that are the same in terms of u and w.
U s, j .(H) j
Initial and Boundary Conditions
Us j 1 (19)
H
For t = 0 and rw ≤ r ≤ re, uj(r,0) = (f,j - 0,j) or wj
Normalized average excess pore pressure over (r,0) = w0,j ® = log10(’0,j/(’f,j). For t > 0 at r = rw, uj (rw,t)
the radial distance at the layer j is: = 0 or wj (rw,t) = 0 and at r = re uj/r = 0 or wj/r = 0
re
for j = 1, 2, 3, …m. The radius of influence zone, re =
0.5de and radius of equivalent drain, rw =0.5dw. Equation
u
rw
j 2 r dr
(13) in non-dimensional form is
U *avg , j re
(20)
uo 2 r dr w j 2w j 1 w j
(14)
rw
T R2 R R
Degree of dissipation of average excess pore
pressure at the layer j is where R = r/de and T = crt/de2.
220
Indian Geotechnical Journal 40(4), 2010
20 Up For
q*o =2
40 1
Us and Up, %
0.5 0 100
Us For all q*o
U*av g(Z) For
60
Degree of Settlement, Us
2
80
n =15 40 1 60
Us For all q*o
0.5
100
0.001 0.01 0.1 1 10 60 40
Time Factor, Th
Fig. 36 Variation of Us and Up with Th 80 20
n=15
z/H=0.5
Figure 37 shows variation of degree of
100 0
settlement, Us and normalized average excess pore
pressure, U*avg(z) with time at a depth, z =0.5H for n=15 0.001 0.01 0.1 1 10
and q*o ranging from 0.5 to 4. While the degree of Time Factor, Th
settlement is independent of the q*o and depth, the
normalized average excess pore pressures at the depth Fig. 37 Variation of Us & U*avg(z) with Th –-
are sensitive to q*o values. The average excess pore Effect of Load Intensity
pressure increases or the degree of dissipation of
average excess pore pressure at any depth, Up(z) =
(100-U*avg(z)) decreases with the increase of q*o. For
example, at 90% degree of settlement, i.e. at Th equal to
about 0.60, the degree of dissipation of average pore 0 100
pressure decreases from about 88% for q*o of 0.50 to
about 80% for q*o of 4. Similarly, at 60% degree of U*av g (Z) at
Avg. Pore Pressure, U* ave. Z, %
ground, the possibility of shallow seated rotational The effect of increase of n on the distribution of
bearing failure is to be examined in view of the large the normalized average excess pore pressure, U*avg (z),
residual pore pressures at these depths over longer with depth is shown in Figure 40 for q*o=1, Th=0.20 and
periods of time. n from 5 to 40. The variation of U*avg (z) is relatively
significant at shallow depths for all the n values
The residual average excess pore pressure, uavg compared to the variation at greater depths. As
(z), at different depths of the thick clay are computed for expected, at any given depth, U*avg(z) increases with
various values of q*o, n and Th and the normalized increase of n since the points farther from the drain take
average excess pore pressures, U*avg (z) = relatively longer times to dissipate the excess pore
(uavg(z)/uo).100, are presented in Figure 39 for n=15 at pressures. U*avg (z) increases from 28% to 70% at mid
time factor, Th=0.20 along with the results of linear depth of clay for n increasing from 5 to 40.
theory. The remarkable phenomenon observed is that
average pore pressure values from the non-linear radial
consolidation theory vary with depth in contrast to the
depth-independent U*avg values of linear theory. U*avg(z)
varies from about 77% to 89% near the surface to about Avg. Pore Pressure, U* ave. Z, %
49% to 63% near the bottom for values of q*o ranging
0 20 40 60 80 100
from 0.50 to 4.0 against a constant value of about 44%
through out the depth in linear theory. The difference 0
between the pore pressures of non-linear and linear q* o=1
theory is relatively large at shallow depths due to large
values of σ’f /σ’o compared to those at greater depths. 0.2 Th =0.20
This difference increases with increase of q*o.
Depth, z/H
40
according to non-linear theory are dependent on q*o as
25
10
15
the variation of q*o influences the ratio σ’f /σ’o. The
average pore pressure increases from about 52% to 0.8
69% at mid-depth of clay for q*o increases from 0.5
to 4. The residual average excess pore pressures thus
are underestimated in the conventional linear theory. In 1
view of the above, instead of applying the entire preload
Fig. 40 Variation of U*avg(z) with Depth—
instantaneously on the soft ground, it may be applied in
Effect of n
increments with proper time lag to allow quicker
dissipation of pore pressures and gain of shear strength.
0.2
0.2
0.50
1
Depth, z/H
0.4
0.05
Linear Theory
Depth, z/H
Non-Linear
0.4
q*o=4
0.4
Th =0.2
0.6
Linear
0.005
0.8
0.6
0.8 n=15
q* o=1
0.8 n=15 1
Th =0.20 100 80 60 40 20 0
Variations of the normalized average excess pore of consolidation. Near the surface at a depth of
pressure, U*avg (z) and degree of dissipation of average z=0.025H, the difference in the degrees of dissipation
excess pore pressure, Up(z) = (100-U*avg(z)) with depth of average excess pore pressures of the two theories is
and time are shown in Figure 41 for n=15 and q*o=1. only 2.70% for Th of 0.005, the difference increasing to
Values from linear theory which remain constant with 37.30% for Th of 0.20 and decreases to 10.40% towards
depth are also shown. At any depth, the degree of the end at Th of 0.80.
dissipation of average excess pore pressure, Up(z), is
relatively smaller in the non-linear theory compared to
that in the linear theory. The expulsion of pore water The distribution of normalized excess pore
depends on permeability of the soil and the permeability pressure at any radial distance and depth, U*(r/rw, z) =
decreases during consolidation due to increase of (u(r/rw,z)/uo.100) with depth at different radial distances
effective stress. The variation of permeability of the soil is presented in Figures 42(a) and (b) (next page) for time
during consolidation is considered in the present non- factors of 0.05 and 0.20 respectively for n=15 and
linear theory. q*o=1. The residual pore pressures are relatively small
being less than 20% close to the drain. The excess pore
The dissipation of pore pressures is relatively pressures vary between 84% to 99% in the outer half of
very slow at shallow depths compared to that at greater the unit cell (i.e. r > rw+0.5(re-rw)) while the variation is
depths in view of the large ratio of final to initial stress relatively sharp ranging from 97% to 0% in the inner half
at shallow depths. For example, at a time factor of 0.20, of the unit cell at time factor of 0.05, The pore pressures
the degree of dissipation of average pore pressure in are nearly constant at all depths farther from the drain
the non-linear theory increases from 18.57% near the compared to the sharp variation in the values in the
surface (z=0.025H) to 47.40% near the bottom middle of the unit cell at all the times. Thus the effect of
(z=0.975H) while it is 55.9% constant with depth in the thickness of the deposit is significantly more at half the
linear theory. Obviously, at the initial stage of distance from the drain to the edge of the unit cell. The
consolidation (Th=0.005) and towards the end variations of pore pressures at different radial distances
(Th=0.80), the difference in the degrees of dissipation of with depth at Th equal to 0.2 is similar to those for Th
average excess pore pressures of the two theories is equal to 0.05 but the magnitudes of the pore pressures
relatively less compared to those at intermediate stages are lesser as can be expected at larger times.
Avg. Pore Pressure, U* (r/rW, Z), % Avg. Pore Pressure, U* (r/rW, Z), %
0 20 40 60 80 100 0 20 40 60 80 100
0 0
T h =0.05
0.2 n=15 0.2
q*o =1
= z/H
Depth, Z=z/H
Depth,ZZ=z/H
P2
0.4 0.4
P3
1
P6
At Point, P
P4
P5
1
At Point, P
Depth,
0.6 0.6
P2
P3
P4
P5
T h =0.2
P6
n=15
0.8 0.8
q*o =1
1 1
(a) (b)
re
P1 P3 P4 P5 P6
P2
P1: At r=rw+ (re‐rw)/32 P2: At r=rw+ (re‐rw)/16 P3: At r=rw+ (re‐rw)/8
rw
P4: At r=rw+ (re‐rw)/4 P5: At r=rw+ (re‐rw)/2 P6: At r=re
Drain Centre Line
Fig. 42 Pore Pressure Distribution with Depth at (a) Th=0.05 and (b) Th=0.2
224
Indian Geotechnical Journal 40(4), 2010
r/rW z/H=0.50
0 5 10 15 0
100
Fig. 44 Variation of Pore Pressures with
z/H=0.025 Radial Distance—Effect of Load Intensity
80
Pore Pressure, U* (r/rW, Z), %
0.10
0.25
0.5
60
r/rW
0.975 0 5 10 15
100
40
n=15 Th =0.05
n=15
q*o=1
80
Pore Pressure, U* (r/rW, Z), %
20 q*o=1 0.1
z/H=0.5
Th =0.20
60 0.2
0
20 0.6
The effect of intensity of loading, q*o, on the
variation of excess pore pressure with radial distance at 0.8
a depth of z=0.5H is shown in Figure 44 for n=15 and 0
Th=0.20. The pore pressure increases from about 50%
to 68% at the centre of the unit cell and from 58% to Fig. 45 Variation of Pore Pressures with
75% at the edge of the unit cell for q*o increasing from Radial Distance—Effect of Time
0.5 to 4. The excess pore pressure at any radial
distance increases with increase of load intensity as it
leads to a corresponding increase of the stress influence. However, the pore pressure variation
increment ratio, σ’f /σ’o. becomes as expected relatively negligible in the later
stages of consolidation (Th >0.6), the variation being
Figure 45 shows the effect of time on excess only from 0 to 12 % at Th=0.6. The pore pressure
pore pressure, U*(r/rw, z) distribution along a radial line distributions along the radial line at mid depth of the
at mid depth of the layer for q*o=1 and n=15. The pore layer for qo*=1, at Th=0.20 and for various n values are
pressure variation along the radial line is relatively shown in Figure 46. The maximum excess pore pressure
considerable in the initial stages of consolidation. At a at the end of zone of influence increases with increase
time factor of 0.10 the pore pressure varies from zero at of n as the farther points take relatively longer times to
the drain to 83.25% at the farthest end of zone of dissipate the pressures.
225
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
q = ku
M Footing Point of
Rotation M
x0 L‐x0
X
Point of q = kc
Winkler’s Spring ku or kc
Rotation
Compression due to Moment
(a) (b) (c)
Fig. 47 (a) Winkler Model of Soil-Structure Response
(b) Moment Loading on Foundation (c) Contact Pressure due to Moment on Rectangular Footing
226
Indian Geotechnical Journal 40(4), 2010
Fig. 50 (a) Stress-Deformation Response for Fig. 50 (b) Contact Pressure Distribution
Nonlinear Winkler Model below Footing
The contact pressure distribution is shown in Figure 50(b). The force and moment equilibrium equations on
simplification are
x02
1 x0 ln 1 1 x0 0 (19)
Rk 2
2
1 1 x
2 0
2 1 1 x ln 1 1 x 0.5 2 Rk x0 M
3
2
0
0 (20)
2 3 L3kcB
2 2 2 3 k ci BL
3
(25) 0.50
0.40
Circular Footing with Moment Only
0.35
X0
Circular Footing
Moment - rotation relationship for circular footing 0.30 Rectangular Footing
(Figure 48) on nonlinear foundation [Figure 50(c)] is For any L/B
obtained from the corresponding force and moment 0.25
equilibrium equations (Madhav et al. 2009). The terms
0.20
are integrated numerically to determine X0 and .
Table 1 lists the parametric values used in the present 0.15
analysis. 1 10 20 30
Rk (ku/kc)
Results and Discussion Fig. 51 X0 verses Rk for Rectangular and
Circular Footings – Linear Winkler Model
The effect of inelastic stiffness ratio, Rk, on the
moment-rotation relationships of rigid footings
subjected to moment loading are obtained based on 22
Winkler model. Figure 51 presents the variation of X0,
the distance to the axis of rotation for rectangular and
circular footings with Rk. The axis of rotation passes
through the midpoint of the length of rectangular or 16
diameter of the circular footing, i.e. X0=0.5 for Rk=1. The
axis of rotation shifts towards the zone of stress I
reduction with increasing inelasticity or Rk values. The
rate at which the axis of rotation shifts with increase in 10 Circular Footing
Rk is quite high for Rk<5 for both rectangular and circular
footings. The rate of decrease of X0 with Rk decreases
gradually for Rk>5. Though the rate of shift of the axis of Rectangular footing
rotation is nearly the same for both types of footings for 4
Rk<2, the rate of decrease is somewhat faster
1 10 Rk = (k u/k c ) 20 30
for rectangular footing for higher Rk. For example, at
Rk=20, X0 shifts by about 32% of its length for
Fig. 52 I verses Rk for Rectangular and
rectangular footing, while it is only 25% of the diameter
Circular Footings – Linear Winkler Model
for circular footing.
229
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
almost the same for all annular ratios for Rk <2.5. The
1.0
rate decreases slightly with increase in annular ratio for
0.9 Rk>2.5. I reduces to 35% from its value at Rk =1 for a
circular footing i.e., annular ratio=0, while for annular
0.8 ratio of 0.9 the reduction is to about 40% for Rk=20.
I (Rk) /I (Rk=1)
0.7
The moment - rotation relationships for
0.6 Rectangular Footing rectangular footings for different Rk based on nonlinear
(hyperbolic response for loading and linear response for
0.5 unloading) are presented in Figure 56 for β=0.005. The
Circular Footing
0.4 conventional analysis using Rk=1 yields least moment
for a given rotation because of least subgrade modulus
0.3 for unloading. The normalized moment increases for any
1 10 20 30 rotation with increase in Rk. The rate of increase of
Rk (ku/kc) moment for a given rotation gets reduced with increase
in Rk for values beyond 2.
Fig. 53 Normalised I verses Rk for Rectangular
and Circular Footings – Linear Winkler Model
0.0025
60
R k=10
Inner Radius =0.005
n= 0.0020 R k=4
Outer Radius
R k=2
40 0.0015
M/(KcL3B)
R k=1
n=0.9 0.0010
I
20
n=0.8 0.0005
n=0.5
n=0.0 0.0000
0 0 0.04 0.08 0.12
1 10 20 25 Rotation, Radian
Rk = (ku/kc)
Fig. 56 Effect of Rk on Moment-Rotation
Fig. 54 Effect of Rk on I for Annular Footing Relationship for Rectangular Footing –
Subjected to Moment – Linear Winkler Model Nonlinear Winkler Model
230
Indian Geotechnical Journal 40(4), 2010
Figure 57 shows the effect of β (=qult/kcL) on particular applied moment as the vertical pressure
moment-rotation relationship for rectangular footings for applied generated prior to the moment application. This
Rk=2. β equal to infinity implies linear response and pressure is expressed as a fraction of ultimate stress
decreasing values of β correspond to lesser values of that the soil can undergo and is designated by q*. Thus
ultimate stress, qult. Moment value increases for any the curves become increasingly flatter at higher values
rotation with increase in β corresponding to stronger of q*.
sub-grade. The asymptotic value of moment is reached
at much lesser rotation for lower values of β (weaker The moment-rotation relationship for circular
sub-grade) than that for higher values of β (stronger sub- footing for different Rk based on nonlinear analysis for β
grade). (=qult/(KcR)) of 0.01 (Figure 59) is similar to that for
rectangular footing. For a particular applied moment,
The effect of axial load prior to application of rotation decreases markedly with increase in Rk from 1.
moment loading on moment-rotation relationship for The rate of reduction of moment with increasing values
rectangular footing based on non-linear model for of Rk decreases with further increase in Rk. The effect of
β=0.005 and Rk=2 is shown in Figure 58. Moment- β on moment-rotation relationship for circular footing
rotation relationship in case of linear response is not based on non-linear analysis for Rk=2 is similar (Figure
affected by the axial load applied before moment 60) to that for a rectangular footing (Figure 58). The
loading. But in case of nonlinear response, axial load resisting moment increases for any rotation with
softens the soil so that moment loading acts on increase in β, though the rate of increase with β gets
softened soil. As a result, rotation increases for a reduced with further increase of β (linear response).
0.006 0.0016
=0.01 6.0
R k=10.0
=0.0275 4.0
R k=2
0.0012 2.0
0.004 0.0125
M/(Kcd4)
M/(KcL3B)
0.0075 0.0008
R k=1.0
0.002
0.0004
=0.0025
0.000 0
0 0.04 0.08 0.12 0.00 0.02 0.04 0.06 0.08 0.10
Rotation, Radian Rotation, Radian
0.0020 0.0035
=0.005 0.2 =0.045
Rk = 2 q*=0 0.4 0.0030
R k=2.0
0.0016 0.035
0.0025
0.6
0.025
M/ (Kc d4)
0.0012 0.0020
M/(KcL3B)
0.0015 0.015
0.0008
q*=0.8 0.0010
0.0004 =0.005
0.0005
0.0000 0.0000
0 0.04 0.08 0.12 0.00 0.02 0.04 0.06 0.08 0.10
Rotation, Radian Rotation, Radian
Fig. 58 Effect of Axial Load Prior to Moment Loading Fig. 60 Effect of on Moment-Rotation
on Moment-Rotation Relationship for Relationship for Circular Footing –
Rectangular Footing - Nonlinear Response Nonlinear Response
231
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
0.3
developed by Pasternak (1954), Filonenko and Borodich
(1945) and Hetenyi (1946) were the extensions of 0.2
Winkler’s model and eliminated the discontinuous
behaviour of the model by introducing interaction
between Winkler springs through elastic layers, smooth 0.1
membrane and elastic plate or beam respectively 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(Selvadurai, 1979). On the other hand, Vlasov (1949) H/L
and Reissner (1958) developed models assuming the (b)
soil medium as elastic continuum but introducing
Fig. 61 Variation of k with H/L for
constraints on displacements and stresses that simplify
different B/H ratios (a) = 0.2 (b) = 0.3
the basic equations for a linear elastic isotropic
after (Noonan and Nixon, 1972)
continuum (Selvadurai, 1979).
232
Indian Geotechnical Journal 40(4), 2010
(b)
constrained modulus, D, and shear modulus, G, kPa . The analysis is carried out for Poisson’s ratio of =
respectively and - the Poisson’s ratio of the sample. 0.25. Typical shear stress – horizontal displacement
The parameters ED and EG are obtained by conducting plots for clayey silt at OMC and medium dense Ganga
direct shear tests on Ganga sand at different relative sand at relative density of 60% are given in Figure 62.
densities - 35%, 60% and 85% and on clayey silt at The variations of moduli computed from shear modulus
different moisture contents - OMC, 2% dry of OMC and (EG) and constrained modulus (ED) and shear modulus
2% wet of OMC. Dimensions of the sample are 663 (G) with normal stress are depicted in Figure 63 and the
cm and the gap ‘B’ between the shear boxes is close to results tabulated in Table 2. The ratio EG/ED (Table 2) for
zero and the normal stress is in the range of 50 to 200 sands in dense state is very close to 1 (as predicted by
200
200 kPa
200
Shear Stress, kPa
150 150 kPa
150 200 kPa
Shear Stress, kPa
100 100 kPa
150 kPa
100
100 kPa
50 50 = σ=50 kPa
σ = 50 kPa
t (a) (b)
0 0
0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5
Horizontal Displacement, cm Horizontal Displacement, cm
Fig. 62 Typical Results from Direct Shear Tests on (a) Clayey Silt at OMC and dmax = 18.4 kPa
(b) Medium Dense Sand of Relative Density = 60 %
25000
E from D 30000
20000 G
25000
E and G in kPa
E from G
E, G in kPa
15000 20000
10000 15000
10000
5000
(a) 5000 (b)
0 0
0 50 100 150 200 0
250 0 50 100 150 200 250
Normal Stress, kPa Normal Stress, kPa
Fig. 63 Typical Results of Variation of Deformation and Shear Moduli with Normal Stress
(a) Clayey Silt at OMC and dmax = 18.4 kPa (b) Medium Dense Sand of Relative Density = 60 %
Table 2 Moduli from Direct Shear Tests in kPa for Normal Stresses of 50 to 200 kPa
classical theory of elasticity) but significantly different symmetric element A (Figure 64a), one gets the
from 1 for sands in loose to medium dense state governing equation in normalized form as,
indicating that the parameters E and G are independent
of each other. The ratios EG/ED are significantly high (2.3 2W 2W 1 W
to 4.7) for all moisture contents for clayey silt indicating 2 0 (26)
Z 2
R R R
that the parameters E and G once again are uncoupled.
where W = w/a, R = r/a, Z = z/H, and =
New Two Parameter Model with Uncoupled (G/E).(H/a)2. Eq. (26) is solved numerically for the
Moduli following cases.
5a
A simplified 2-parameter continuum model is
proposed (Madhav and Raja Sekhar 2010) considering a 1 2 3 ……nj
NR = 5 (nj-1)
the elastic constants E (deformation modulus) and G 1 r
(shear modulus) to be independent of each other unlike 2
in the classic theory of elasticity. A uniformly loaded 3
circular footing of diameter 2a, with load of intensity p, i . w= 0
(Figure 64a) on the surface of a finite layer of soil mass .
H .
underlain by a rigid base is considered. Analysis of the .
problem is available for semi-infinite elastic layer and for
finite elastic layer underlain by a rough, rigid base
(Burmister 1956; Poulos 1967, Dempsey and Li 1989)
based on the elastic continuum approach. In the
present study, the soil mass below the footing is w= 0
assumed as homogeneous and isotropic continuum but
characterized by the parameters – deformation
modulus, E, and shear modulus, G, independent of each Fig. 64 (b) Definition Sketch Discretisation
other. As the horizontal displacements, u, are of less
engineering interest and are considered to be small for Rigid Footing
symmetrical vertical loading compared to vertical
displacements, w, they are assumed to be zero. The A uniform displacement, wo, is imposed at all
vertical displacements, w, at a radial distance r = 5a are nodes beneath the rigid footing and the normal stress
considered to be negligible (Figure 64b). Considering on the other nodes on the surface are zero. Eqs. (26),
vertical and horizontal force equilibrium of an axi- are solved iteratively, for the following boundary
conditions: W1,j = Wo, Wn,j= 0 and Wi,5nj-4= 0. Normalizing
stress, z, with deformation modulus, E, one gets *z = -
2a
[a/H].W/Z where z* = z , W = w and z = Z . Once
E a H
the displacements are determined, contact pressures
r beneath the rigid footing are determined.
A Uniform Loading
rz+(rz /r) r R
W1, j 1 1
H 1 2R j
r+(r/r) r 2W2, j p 1
(a) a NH W R
z+ (z/z) z 1
1, j 1 2R j
Fig. 64 (a) Definition Sketch - W1, j (27)
Stresses on Element ‘A’ 2 1 1
234
Indian Geotechnical Journal 40(4), 2010
W/Wo
W/Wo
0 0.5 1
0
20
10
z/a
5
2
z /H
0.5 H/a = 1
Boussinesq’s curve
G/E = 0.5
1
(a)
1
(a)
Fig. 65 (a) Normalized Displacements versus
Depth Effect of Shear Stiffness
at Centre of Rigid Footing
W/Wo
/
0 0.5 1
W/Wo 0
20
10
2
5
/H H/a = 1
z /H
0.5
G/E = 0.5
z/a
W 1
Boussinesq’s curve (b)
1
(b)
q/qavg
1
5 2
2
2 1
Wo = 1
3 Wo = 1 5, 10, 20
(a) (b)
3 4
(a) (b)
W/Wo W/Wo
W/Wo
0 0.5 1 0 0.5 1
0 0
10.0 10.0
5.0
5.0
2.0
H/a = 20
2.0 p* = 1
H/a = 20 1.0
z/a
z/a
z/a
5 5
z/a
1.0 *
1 0.5
Boussinesq’s curve
0.5
Boussinesq’s curve 0.2
0.2
G/E = 0.1 G/E = 0.1
(a) (b)
10 10
(a) (b)
solution are slightly more than those for G/E = 0.1, for are almost the same for G/E = 0.5 (Figure 70b).
z/a ≤ 1. The displacements lie between those for soils
with G/E = 0.1 and 0.2 for 4 < z/a ≤ 7 (Figure 68b). Vlasov and Leontiev (1966) proposed a two
Boussinesq’s solution overestimates the settlements, parameter model assuming exponential decrease of
both beneath the centre and the edge of the footing for displacements with depth. The proposed exponential
stiff soils. The variation of normalized displacements function is
with normalized depth for deposits of thicknesses, H/a
= 1 to 20, are depicted in Figure 69. sinh ' (1 z / H)
h(z) (28)
sinh '
The vertical surface displacements beneath the
footing decrease with increase in stiffness and are
nearly constant along the footing for soils with G/E = 0.1 where ’ = H - a function of G/E, H/a and r/a. The
to 10 for r/a ≤ 0.5 (Figure 70a). The displacements parameter, ’, determines the rate of decrease of
decrease gradually for r/a > 0.5. The decrease in displacements with depth (Scott 1981). Jones and
displacements is more for soils with G/E = 0.1 to 2 Xenophontos (1977) found Eq. (28), proposed by Vlasov
compared to those for soils with G/E = 5 to 10. The and Leontiev (1966), to be fairly accurate for finite layer
vertical surface displacements under uniformly loaded problems (Scott 1981).
footing resting on the deposits of thicknesses, H > 5a,
W/Wo W/Wo
0.0 0.5 1.0 0.0 0.5 1.0
0.0 0.0
o 5
20 20
1 10
0 5 5 2
Z/a
2
H/a = 1
H/a = 1
Z/a
0.5 0.5
G/E = 0.5 G/E = 0.5
p* = 1 p * = 1
r/a r/a
0 0.5 1
0 0.5 1
0 0
H/a = 1
p* = 1 10 G/E = 0.5
p* = 1
5
0.5
w 2 H/a = 1
1 w
0.6
0.5 2
0.2 1
G/E = 0.1 5
10, 20
(a) (b)
1.2 1.5
(a) (b)
N ormalized Depth, z/H
loaded circular footing for G/E = 1.0, H/a = 2.0 and ’ = 0.2
2.3 for settlements beneath the center and the edge.
Similarly, the variations of displacements with depth 0.4
under the rigid circular footing are determined for G/E =
0.5, H/a = 2.0 and ’ = 1.26 beneath the center and the 0.6
edge. The variations of displacements with depth, under Vlasov & Leontiev
the rigid and flexible circular footings, obtained from Centre
0.8 Edge
the proposed new uncoupled model compare well (a)
(Figure 72) with those obtained from Vlasov and
1
Leontiev (1966). (a)
Normalized Displacement, W
0 0.2 0.4 0.6 0.8 1
0
18
15 Rigid
H/a = 10 N ormalized Depth, z/H 0.2
Flexible
12 0.4
H/a = 5
9
0.6
Vlasov & Leontiev
6 H/a = 2 Centre
0.8 Edge
3 (b)
H/a = 1
1
0
0 2 4 6 8 10 (b)
G/E Fig. 72 Comparison of Vertical Displacements with Depth
from Proposed 2-Parameter Model under Circular Footing
Fig. 71 Variation of Parameter with
with those from Vlasov and Leontiev (1966)
Relative Stiffness, G/E, - Effect of Thickness, H/a (a) Uniform Loading (b) Rigid Footing
Strain Induced Non-homogeneity of Soils soil, Es, varies with depth as measured through e.g.,
CPT) based on the assumed variation of strain influence
Stresses and deformations due to applied loads factor with depth.
are estimated based on linear elastic and homogeneous
conditions. The underlying assumption is that at working A typical plot (Figure 73, next page) of variation
loads the soil mass is linearly elastic and that the of secant modulus with deviatoric strain (Fahey and
stress/strain changes in the soil are close to those given Carter 1993) illustrates the strong decrease of the
by linear elasticity even though the soil response may be modulus even in pre-failure deformation or strain range.
nonlinear. Boussinesq solution for the distribution of The variation of modulus of deformation, E, with strain,
stresses in a half-space resulting from surface loads is , can be expressed in the form,
based on the assumption of linear elastic homogeneous
isotropic half-space for the soil media. While some E = Emax/{1 + ( (29)
writers reported acceptable agreement between
experimental results and those from elastic theory where Emax is the small strain modulus and and are
(Siddharthan et al. 1996), others observed significant parameters.
differences which in some cases exceed 200% (Ullidtz et
al. 1996). The disparity between the experimental Analysis
results and the elastic theory could be attributed to
nonlinear soil behavior, anisotropy, stress or strain A circular footing, 4.0 m in dia. resting on the
induced non-homogeneity, etc. Schmertmann’s method surface of cohesionless soil with angle of shearing
(Schmertmann et al., 1978) is ideally suited to estimate resistance of 32o, deformation modulus of 20 MPa and
the settlement of footings for the cases with material Poisson’s ratio: 0.3, subjected to a uniform nominal load
non-homogeneity (i.e., the deformation modulus of the of 10 kPa is analyzed using PLAXIS. Radial deformation
238
Indian Geotechnical Journal 40(4), 2010
Vertical Strain, %
0.00 0.03 0.06
0.0
E/Emax
0.5 H/D=4
a=0.25, b=0.70
0.5
Z/H
a=0.25, b=0.90
1.0
0.0
1.E-06 1.E-02 1.E+02
H/D=0.5
Strain, %
Fig. 73 Normalised Deformation Modulus vs Strain 1.0
(after Fahay and Carter 1993) Fig. 75 Variation of Strain with Depth
0.5H for H/D of 0.50 and 0.1H for H/D of 4.0. The
variations of final or converged deformation moduli with
depth for H/D of 0.50, 1.0, 2.0 and 4.0 are presented in
Figure 76. It should be noted that the medium is 1.0
homogenous initially with the deformation modulus
equal to 20 MPa everywhere. The deformation modulus
decreases with depth near the top close to the surface H/D=0.5
to attain a minimum value and increases thereafter to 1.0
all depth to diameter ratios. The modulus of deformation
is of the order of 16 to 17 MPa for a thin layer of H/D = Fig. 76 Variation of Final Deformation Moduli with Depth
239
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
0.50. For larger thickness of the ground, the modulus of of dissipation of pore pressures are sensitive to the
deformation increases with depth from the minimum magnitude of loading in the case of a thick layer while
value and in case of a very thick deposit attains the these are independent of loading in the thin layer
initial value of 20 MPa where the strains are negligibly theory.
small. Thus non-homogeneity is induced in the medium
because of strain attained at that particular point. One of the major findings of the non–linear
theory of consolidation for both vertical and radial flows
is that while the degree of settlement is independent of
Conclusions the final to initial stress ratio, the degree of dissipation
of pore pressure is very much dependent on the stress
Geotechnical Engineers are highly indebted to ratio. The residual excess pore pressures are
Karl Terzaghi for founding the discipline in mid 1920s. underestimated in the conventional linear theory. The
The subject has grown by leaps and bounds due to the present theory substantiates the actual in-situ slower
efforts succeeding academics and practitioners. SM & rate of degree of dissipation of excess pore pressures
GE is celebrating the Platinum Jubilee in November compared to that of the degree of settlement. The
2010. It is an appropriate time to review some of the degree of dissipation of excess pore pressure decreases
basic tenets introduced by Terzaghi nearly eighty years with increase of the stress ratio. This ratio varies with
ago. We now have a much better understanding of the depth, and is relatively large at shallow depths where
complexity of soil as an engineering material and ground initial stress is very less. Hence the non-linear
as an entity. This paper examines some of those basic consolidation effect is pronounced at shallow depths
assumptions, viz., rigid plasticity and non-consideration compared to the effect at greater depths. A major
of the structure in estimating bearing capacity of finding of the present work is that the excess pore
foundations, the extension of theory of consolidation for pressures due to radial drainage vary not only with time
thin layers based on linear void ratio – effective stress and radial distance but also with depth in contrast to
relation to thicker deposits and soils that follow linear depth-independent pore pressures from the
void ratio – log effective stress relation, inelasticy of conventional theory for radial flow. That is, while pure
ground and its effect on moment – rotation radial flow consolidation may be valid for thin layers, no
relationships, uncoupling deformation and shear moduli such assumption can be made for thicker layers. The
based on experimental evidence and its incorporation in effect of non-linear consolidation is more pronounced
a simple foundation model, etc., made in the early during the intermediate stages of consolidation. The
formative years. practical significance of the proposed theory is that it
can explain failure of high embankments constructed
Conventional approaches for the estimation of rapidly on thick deposits of fine grained soils as the
the ultimate capacities of both shallow and deep stress ratio at shallow depths would be significantly high
foundations use limit equilibrium methods and are leading to slower rate of dissipation of pore pressures
based on only the strength parameters apart from the and higher residual pore pressures.
geometry of the foundation element. Ground/soil being The inelastic response of the ground needs to be
a much more complex material than metals from which considered for the analysis of rigid foundations
the original theories have been developed, requires the subjected to moment loading. The inelastic response is
consideration of stiffness or compressibility as well as characterized by the stiffness ratio, Rk, defined as the
the strength parameters for the estimation of ultimate ratio of subgrade moduli in unloading to that in
loads. The results for the bearing capacity of shallow compression. The ratio, Rk, and the non-linear stress –
displacement response have very significant effect on
foundations and the ultimate axial capacity of piles
the overall response of footing to applied moment
illustrate their dependence on the stiffness of the
loading. The axis of rotation shifts towards the unloading
ground. Consideration of height of a structure rather
side from the center of the contact area with increase of
than just only the foundation leads a very interesting
Rk. The rate at which the axis of rotation gets shifted
and unique failure state, the leaning instability of tall
reduces gradually with increase of Rk. The moment
structures that is governed by the stiffness of the
ground rather than the strength. influence coefficient, I, decreases with increase in Rk
implying that the conventional approach over-estimates
A simple and approximate non-linear theory of the rotation of footings due to moment loading. The rate
one-dimensional consolidation for thick clay layer is of decrease of I with Rk reduces gradually. The
developed. The conventional thin layer theory under- normalized values of I for any Rk with respect to the
estimates the degree of consolidation but over- corresponding value at Rk=1 are nearly the same for
estimates the degree of dissipation of excess pore rectangular and circular foundations. I for annular
pressures. The isochrones in the case of a thick layer footings for any Rk varies marginally with increase in
with PTPB condition are slightly unsymmetrical about annular ration for n < 0.5 but increases markedly for n >
the mid depth in contrast to symmetrical isochrones in 0.5. The variation (decrease) of normalized I with Rk is
thin layer theory. The degree of settlement and degree almost independent of annular ratio, n. Results based
240
Indian Geotechnical Journal 40(4), 2010
on nonlinear response of the ground are similar to the graduate students, Saha, Rajasekhar, Manoj,
response based on linear models. The rotation of footing Padmavathi, Ayub Khan, Lakshmana Rao, Suresh,
caused by a moment loading on a nonlinear model is Vidyaranya, etc., who have worked hard over several
more if an axial load is acting prior to the application of years to concretise the ideas enunciated in this paper. I
moment loading. am always indebted to my students who have made me.
The contribution of Dr Umashankar in formatting the
A new two parameter continuum model with paper is gratefully appreciated.
uncoupled moduli is proposed to estimate the vertical
displacements and contact pressures under rigid and References
flexible circular footings. The proposed uncoupled model
is validated by comparing the variations of
Almeida, M.S.S., Santa Maria, P.E.L., Martins, I.S.M., Spotti,
displacements with depth with those from Vlasov and
A.P. and Coelho, L.B.M. (2000): 'Consolidation of very
Leontiev (1966). It is observed that in case of rigid soft clay with vertical drains', Geotechnique, 50(6),
footing, the Boussinesq’s curve at centre of the footing 633–643.
falls between G/E = 0.2 and 0.5 and at edge the
displacements are more near the top than those Ayub Khan, P., Madhav, M.R. and Saibaba Reddy, E.
obtained from 2-parameter model for all G/E values. In (2009): 'Non-linear Theory of Consolidation of Thick
case of flexible footing, the displacements from Layer Clay Layer by PVD. Int. Symp. on Ground
Boussinesq’s solution at centre of the footing matches Improvement Technologies and Case Histories',
closely with those for G/E = 0.2 till z/a = 4 and Singapore, 337-344.
thereafter the Boussinesq’s curve falls between G/E = Ayub Khan, P., Madhav, M.R. and Saibaba Reddy, E.
0.1 and 0.2. The Boussinesq’s curve almost matches (2010): 'Simplified Non-Linear Theory of Vertical
closely with G/E = 0.1 at the edge till z/a =1.5 and Consolidation of Thick Clay Layers', IGC 2010, Mumbai.
thereafter falls between G/E = 0.1 and 0.2.
Boussinesq’s solution over predicts the vertical Barron, R.A. (1948): 'Consolidation of fine grained soils by
displacements in case of stiff soils (G/E = 5 to 10) under drain wells', Trans. ASCE, 113, 718–754.
rigid and flexible footings. Basak, P. and Madhav, M.R. (1978): 'Analytical solution of
sand drain problems', J of Geotech. Engrg, ASCE,
The contact pressures beneath the rigid footing
104(1), 129–135.
increase very sharply near the edge of the footing for
soils with G/E = 0.1 to 0.5 and are same for soil layers, Bergado, D.T., Balasubramaniam, A.S., Fannin, R.J. and
with same stiffness, G/E = 0.2, of thickness greater that Holtz, R.D. (2002): 'Prefabricated vertical drains (PVDs)
5 a. The vertical surface displacements beneath the in soft Bangkok clay: a case study of the new Bangkok
flexible footing decrease with increase in stiffness of soil International Airport project', Can.Geotech, J., 39, 304-
and are almost constant along the footing for stiff soils 315.
(G/E = 5 to 10). But for soils with G/E = 0.1 to 2 the
Borowicka, H. (1943): ‚Uber ausmittig belaste starre platen
vertical surface displacements decrease along the auf elastic-isotropem Unter grund‘. Ingenier-Archiv,
footing and the rate of decrease is more near the edge. Berlin, 1, 1-8.
The vertical surface displacements beneath the flexible
footing increase with increase in thickness of the soil Burmister, D.M. (1956): ‘Stress and Displacement
layer but are almost the same for layers of thickness, H, Characteristics in Two-Layer Rigid Base soil System:
greater than 10.a. Influence Diagrams and Practical Applications’. Proc.
HRB, 35, 773.
The stresses and strains in soil mass depend on
Cheney J.A., Abghari A. and Kutter B.L. (1991): ‘Stability of
the stress-deformation characteristics, anisotropy and
Leaning Towers’. J Geotech. Eng., Proc. ASCE, 117( 2),
non-homogeneity of the medium. Recent research has 297-318.
established the variation of deformation modulus with
strain level. The problem of deformation response of a Conte, E. and Troncone, A. (2009): ‘Radial consolidation
uniformly loaded circular footing on an initially with vertical drains and general time-dependent
homogeneous medium of finite thickness is analyzed loading’, Can. Geotech. J., 46, 25-36.
incorporating the variation of deformation modulus with
Davis, E.H. and Raymond, G.P. (1965): A nonlinear theory
strain. The medium becomes non-homogeneous
of consolidation, Geotechnique, 15(2), 161–173.
because of strain induced non-homogeneity. The degree
of non-homogeneity depends on the relative thickness Dempsey, J.P. and Li, H. (1989): ‘A Rigid Rectangular
of the deposit. Footing on an Elastic Layer’, Geotechnique, 39, 147-
152.
Acknowledgments Egorov, K.E., Barvashobv, V.A. and Fedorovsky, V.G. (1973):
Some Application of Elasticity Theory to Design of
This work would not have day-lighted without the Foundations. Proc. 6th Int. Conf. SMFE, Moscow, 1(3),
contributions of a large number of former and current 60-74.
241
Grundbaumechanique/Geomechanics Revisited
A New Paradigm for Geotechnical Engineering
Madhira R Madhav
Egorov, K.E., Kitaikina, O.V. and Zinovyev, A.V. (1979): Jones, R and Xenophontos, J. (1977): ‘The Vlasov
Deformation of Ring Foundation Base. Proc. 6th Asian Foundation Model’. Int. Journal of Mech. Sc., 19, 317-
Regional Conf. SMFE, Singapore, 1, 281-284. 323.
Egorov, K.E., Vronsky, A.V., Finseva, T.V. and Illynykh, V.A. Indraratna, B., Bamunawita, C., Redana, I.W and McIntosh,
(1980): Study of the Stress and Strain State of a G. (2003): Modeling of prefabricated vertical drains in
Subgrade under an Eccentric Load. Proc. 8th Danube- soft clay and evaluation of their effectiveness in
European Conf.SMFE, Varna, Bulgeria, 2, 67-78. practice, Gr. Impr., 7(3), 127–137.
Egorov, K.E. and Vronsky, A.V.(1980): Study of Stress and Indraratna, B., Rujikiatkamjorn, C. and Sathananthan, I.
Strain Rate of a Subgrade under an Eccentric Load. (2005a): ‘Radial consolidation of clay using
Proc. 8th Danube-European Conf. SMEF, Varna, 2, 67- compressibility indices and varying horizontal
78. permeability,’ Can. Geotech. J., 42(5), 1330–1341.
Fahey M and Carter J.P. (1993): “A finite element study of Indraratna, B., Sathananthan, I., Rujikiatkamjorn, C. and
the pressuremeter in sand using a nonlinear elastic Balasubramaniam, A.S. (2005b): ‘Analytical and
plastic model” Canadian Geotechnical J. 3, 348-362. numerical modeling of soft soil stabilized by
prefabricated vertical drains incorporating vacuum
Filonenko-Borodich, M. M. (1945): A Very Simple Model of preloading’, I J of Geomechanics, 5(2)114–124.
an Elastic Foundation capable of Spreading the Load.
Sb. Tr. Mosk. Elektro. Inst. Inzh. Trans. No: 53 Indraratna, B., Rujikiatkamjorn, C. and Chu, J. (2007): ‘Soft
Transzheldorizdat (in Russian). clay stabilization with geosynthetic vertical drains
beneath road and railway embankments: A critical
Gazetas, G. (1998): ‘Elastic Formulae for Lateral review of analytical solutions and numerical analysis’,
Displacement and Rotation of Arbitrarily-shaped Advances in Measurement and Modeling of Soil
Embedded Foundations’. Geotechnique, 38(3), 439- Behavior (GSP173), 236, 7.
444.
Indraratna, B., Aljorany, A. and Rujikiatkamjorn, C. (2008):
Georgiadis, M. and Butterfield, M. (1988): ‘Displacements ‘Analytical and numerical modeling of consolidation by
of Footings on Sand under Eccentric and Inclined vertical drain beneath a circular embankment’, I J of
Loads’. Canadian Geotech. J., 25, 199-211. Geomech., 8(3), 119–206.
Gibson, R. E. and Anderson, W. F. (1961): ‘In-situ Lambe, T.W and Whitman, R.V. (1969): Soil Mechanics,
measurements of soil properties with the John Wiley & Sons, New York, 151-161.
pressuremeter’. Civ. Engng Public Wks. Rev., 56, 615-
618. Lancelotta, R. (1993): ‘Stability of a Rigid Column with Non-
linear Restraint’. Geotechnique, 43(2), 331-332.
Gibson, R.E., Schiffman, R.L. and Cargill, K.W. (1981): ‘The
theory of one-dimensional consolidation of saturated Lekha, K.R., Krishnaswamy, N.R. and Basak, P. (1998):
clays: II. Finite nonlinear consolidation of thick ‘Consolidation of clay by sand drain under time–
homogeneous layers’. Can. Geotech J, 18(2)280-293. dependent loading’, J of Geotechnical and
Geoenvironmental Engineering, 124(1), 91-94.
Hambly, E.C. (1985): ‘Soil Buckling and Leaning Instability
of Tall Structures’. Structural Engr. 63A(3), 77-85. Madhav, M.R. and Miura, N. (2004): ‘Phase transition
effect on preconsolidation stress of soil’, Indian
H Hambly, E.C. (1990): ‘Overturning Instability’. J. of Geotechnical Journal, 34(1), 80-95.
Geotechnical Engineering, ASCE, 116(4), 704-709.
Madhav, M.R. and Padmavathi, V. (2008): ‘Effect of
Hanrahan, E.T and Mitchell, J.A. (1969): ‘The Importance of Stiffness of Ground on Ultimate Capacity of
Shear in Consolidation.’ Proc. of 7th ICSMFE, Mexico, 1, Foundations’, Proc. IGC 2008, Bangalore.
183-190.
Madhav, M.R., Padmavathi, V., and Saha, S. (2009):
Hanrahan, E.T. (1973): ‘The Stress-Strain Relationship in ‘Inelastic Response and Moment – Rotation
Soil Mechanics.” Trans., Inst. of Engrs of Ireland, 97. Relationship of Foundations’, IGC 2009, Guntur.
Hanrahan, E.T. (1971): ‘The eg and ek Parameters’. Roscoe Madhav, M.R. and Raja Sekhar, M. (2010): ‘New Two
Symp., Cambridge, 362-367. Parameter Model with Uncoupled Moduli for Ground’,
IGC 2010, Mumbai.
Hansbo, S. (1979): ‘Consolidation of clay by band shaped
prefabricated drains’, Gr. Engg, 12(5)16-25. Menard, L. (1957d): ‘Mesures in situ des propriétés
physiques des sol's’, Annales des Ponts et Chaussées,
Hansbo, S. (1981): ‘Consolidation of fine grained soils by
Paris, 14, 357-377.
prefabricated drains’, Proceedings of the 10thICSM&GE,
Stockholm, Sweden, 3, 667-682. Noonan, D.K.J and Nixon, J.F. (1972): ‘The Determination of
Young’s Modulus from the Direct Shear Test’. Canadian
Hetenyi, M. (1946): Beams on Elastic Foundations.
Geotechnical Journal, 9, 504-507.
University of Michigan Press, Ann Arbor, Michigan.
242
Indian Geotechnical Journal 40(42), 2010
Pasternak, P.L. (1954): ‘On a New Method of Analysis of an Terzaghi, K. and Peck, R.B. (1967): Soil Mechanics in
Elastic Foundation by means of Two Foundation Engineering Practice. Wiley, New York.
Constants’. Gosudarstvennoe Izdatelstro Liberaturipo
Stroitelstvui Arkhitekture, Moscow (in Russian). Ullidtz, P., Askegaard, V., and Sjolin, F.O. (1996): Normal
stresses in a granular material under FWD loading.
Potts, D. M. (2003): ‘Numerical Analysis: A Virtual Dream or TRB, National Research Counsil, Washington, D.C.
Practical Reality’. Rankine Lecture, Geotechnique,
53(6) 535-572. Vaid, Y.P. (1985): ‘Constant rate of loading nonlinear
consolidation’, Soils and Foundations, 25(1), 105–108.
Poulos, H.G. and Davis, E.H. (1974): Elastic Solution for Soil
and Rock Mechanics. John Wiley & Sons, NY, 43-54. Vesic, A. S. (1972): ‘Expansion of Cavities in infinite soil
mass’. J. Soil Mech. And Found. Div., ASCE, 98(3), 265-
Poulos, H.G. (1967): ‘Stresses and Displacements in an 290.
Elastics Layer Underlain by a Rough Rigid Base’.
Geotechnique, 17, 378-410. Vesic, A.S. (1973): ‘Analysis of Ultimate Loads on Shallow
Foundations’. J. Soil Mech. And Found. Div., ASCE,
Reissner, E. (1958): ‘Deflection of Plates on Viscoelastic 99(1), 45-73.
Foundation’. J Appd Mech., 80, 144-145.
Vesic, A.S. (1977): ‘Design of Pile Foundations’, National
Richart, F.E. (1959): ‘Review of the theories for sand Cooperative Highway Research Program, Synthesis of
drains’, Trans. of ASCE, 124(2999), 709–736. Practice, No. 42, Transportation Research Board,
Washington, DC.
Rujikiatkamjorn, C. and Indraratna, B. (2006): ‘Three-
dimensional numerical modeling of soft soil Vlasov, V.Z. (1949a): ‘Structural Mechanics of Thin-Walled
consolidation improved by prefabricated vertical Three-Dimensional Systems’. Gosstroiizdat, Moscow,
drains’, Ground Modification and Seismic Mitigation Leningard (in Russian).
(GSP 152), 196, 22, 161–168.
Vlasov, V.Z. (1949b): ‘General Theory of Shells and its
Rujikiatkamjorn, C., Indraratna, B. and Chu, J. (2008): ‘2D Application in Engineering’. Gosterkhizdat, Moscow,
and 3D numerical modeling of combined surcharge and Leningard (in Russian).
vacuum preloading with vertical drains’, I J of
Vlasov, V.Z and Leontiev, U.N. (1966): ‘Beams, Plates and
Geomechanics, 8(2), 114–156.
Shells on Elastic Foundations’, Israel Program for
Sathananthan, I. and Indraratna, B. (2006): ‘Plane-strain Scientific Translations, Jerusalem (Translated from
lateral consolidation with non-Darcian flow’, Russian).
Can.Geotech, J., 43, 119-133. Walker, R., Indraratna, B. and Sivakugan, N. (2009):
Scott, R.F. (1981): Foundation Analysis. Prentice Hall, NJ, ‘Vertical and radial consolidation analysis of
85-118. multilayered soil using the spectral method’, J.
Geotech. Geoenv. Eng., ASCE, 135(5), 657– 663.
Schmertmann, J. H. and Hartman, J.P. (1978): ‘Improved
Strain Influence Factor Diagrams,’ J. of Geotechnical Xie, K.H. and Leo, C.J. (2004): ‘Analytical solutions of one-
Engrg. Div., ASCE, 104(GT8), 113-1135. dimensional large strain consolidation of saturated and
homogeneous clays’. Computers and Geotechnics, 31,
Selvadurai, A.P.S (1979): ‘Elastic Analysis of Soil-Structure 301-314.
Interaction’, Developments in Geotechnical Engineering
Vol 17, Elsevier Scientific Publishing Company. Xie, K.H., Xie, X.Y. and Jiang, W. (2002): ‘A study on one-
dimensional nonlinear consolidation of double-layered
Siddharthan, R.V., Yao, J., and Sebaaly, P. (1996): Field soil’. Computers and Geotechnics, 29, 151-168.
Verification of a Moving Load Model for Pavement
Response. TRB, National Research Council, Zhuang, Y.C., Xie, K.H. and Li, X.B. (2005): ‘Nonlinear
Washington, D.C. analysis of consolidation with variable compressibility
and permeability’, J. of Zhejiang University (Engineering
Teh, C.I. and Nie, X. (2002): ‘Coupled consolidation theory Science), 6A(3), 181–187
with non-Darcian flow’, Computers and Geotechnics,
29(3), 169-209.
Terzaghi, K. (1943): Theoretical Soil Mechanics. John Wiley
& Sons, New York, p. 510.