One-dimensionalconsolidation theory: unsaturated soils

DELWYN
G . FREDLUND
AND JAMSHEDU. HASAN
Departmetzt of Civil Etrgitieeritig, Utiiversiiy of Suskatclre~vatz,Saskrrtoori, Sask., Catiada S7N 0 WO
Received October 4, 1978
Accepted March 28, 1979
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

A one-dimensional consolidation theory is presented for unsaturated soils. The assumptions

made are in keeping with those used in the conventional theory of consolidation for saturated
soils, with the additional assumption that the air phase is continuous. Two partial differential
equations are derived to describe the transient processes taking place as a result of the applica-
tion of a total load to an unsaturated soil.
After a load has been applied to the soil, air and water flow simultaneously from the soil
until equilibrium conditions are achieved. The simultaneous solution of the two partial
differential equations gives the pore-air and pore-water pressures at any time and any depth
throughout the soil. Two families of dimensionless curves are generated to show the pore-air
and pore-water dissipation curves for various soil properties.
For the case of an applied total load, two equations are also derived to predict the initial
pore-air and pore-water pressure boundary conditions. An example problem demonstrates the
nature of the results.

L'article prksente une thiorie de consolidation unidimensionnelle pour les sols non saturks.
Les hypothkses faites sont conformes h celles utilisees dans la thCorie conventionnelle pour les
sols saturCs, avec I'hypothkse additionnelle de la continuit6 de la phase gazeuse. Deux equations
aux dCrivees partielles sont Ctablies pour dCcrire les phCnomknes transitoires qui rCsultent de
For personal use only.

l'application d'une charge h un sol non saturC.

Aprks l'application d'une charge au sol, l'eau et l'air s'echappent simultanCment du sol
jusqu'h ce que des conditions d'kquilibre soient atteintes. La solution simultanCe des deux
equations aux dCriv6es partielles donne les pressions interstitielles de I'air et de l'eau en tout
temps et toute profondeur dans le sol. Deux familles de courbes adimensionnelles sont
gCnCrCes pour donner les courbes de dissipation des pressions d'air et d'eau pour differentes
propriCtCs du sol.
Pour le cas d'une charge totale appliquCe, deux equations sont Cgalement Ctablies pour
ridi ire les oressions initiales dans I'air et dans l'eau. Un cas type dkmontre la nature des

[Traduit par la revue]

Can. Geotech. J., 16,521-531 (1979)

Introduction saturation of the soil. He states the problem is

Since the inception of modern soil mechanics,
Terzaghi's theory of consolidation for saturated soils
has formed an extremely useful conceptual frarne-
work in geotechnical engineering. Unfortunately, the
study of the behavior of unsaturated soils has taken
place in the absence of a similar theoretical frame-
work. As a result, it has been difficult to envisage the
trailsitions in theory when going from a saturated
soil to an unsaturated soil.
As far back as 1941, Biot presented an analysis of
the transient flow problem in saturated soils. He
suggested two constitutive relations for the soil and
solved for changes in the pore-water pressure with
time. Biot considered the air to be in an occluded
state with no flow of air during the consolidation
process.
In 1965, Barden presented an analysis of the one-
dimensional consolidation of compacted, unsatu-
rated clay. He subdivides the consolidation problem
into various categories depending upon the degree of
0008-3674/79/030521-11$01 .OO/O
@ 1979 National Research Council of Canada/Conseil national de recherches du Canada
indeterminate and. therefore. various assum~tions
are made in order to complete the analysis. Bishop's
equation is used to describe the stress conditions in
the unsaturated soil.
Partial differential eauations have been develo~ed
in the soil science discipline to describe unsteady
moisture movement. More recently these equations
have received increasing acceptance in the soil
mechanics field (Aitchison et crl. 1965). These equa-
tions should be considered as a special case since the
compressibility of the soil structure and the escape of
air are not taken into consideration. They are
generally not applied to transient processes associated
with the application of an external load.
Fredlund and Morgenstern (1977) proposed stress
state variables for unsaturated soils on the basis of
the equilibrium equations for a multiphase system.
~hese-werealso verified experimentally. The element
of unsaturated soil was considered as a four phase
system with two phases that come to equilibrium
 CAN. GEOTECH. J. VOL. 16, 1979
-
AV
SOIL STRUCTURE
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13
For personal use only.
PHASE
FIG. 1. Constitutive surfaces for the various phases of an
unsaturated soil.
under applied stress gradients (i.e., soil particles and
contractile skin or air-water interface) and two
phases that flow under applied stress gradients (i.e.,
air and water). Fredlund and Morgenstern (1976)
also proposed and experimentally tested constitutive
relations for volume change in unsaturated soils. In
addition, the continuity requirements for an element
of unsaturated soil were outlined. The description of
the stress, continuity, and constitutive relations, and
suitable flow and compressibility laws for air and
water, provide the necessary physical requirements
for a more rigorous formulation of transient pro-
cesses in unsaturated soils.
This paper presents a general one-dimensional
consolidation (or swelling) theory for unsaturated
soils within a theoretical framework similar to that
for saturated soils. As well, equations are derived to
predict the initial air and water boundary conditions
associated with the application of an externally
applied load. An example problem is included to
demonstrate the solution of the above equations.
Physical Requirements for the Formulation
The state of stress in an unsaturated soil can be
described by any two of a possible three stress state
variables (Fredlund and Morgenstern 1977). Accept-
able combinations are: (1) ( o - 11,) and (u, - u,");
(2) (o - u,") and (u, - 11,); and (3) (o - un) and
(o - u,~).The stress variables selected to derive the
consolidation equations in this paper are (o- u,) and
(ua - u,), where o = total stress; u, = pore-air
pressure; and u, = pore-water pressure. Continuity
of an unsaturated soil element requires that the over-
all volume change of the element must equal the sum
of the volume changes associated with the com-
ponent phases (Fredlund 1973). If the soil particles
are considered incompressible and the volume change
of the contractile skin (i.e., air-water interface) is
considered as internal to the element, the continuity
requirement can be written:
[ I ] AV/V = av,p + A V , ~
where V = overall volume of the soil element;
Vw = volume of water in the soil element; and V , =
volume of air in the soil element.
If any two of the volume changes are known, the
third can be computed. In other words, it is necessary
to have two constitutive equations to define volume
change behavior in unsaturated soils.
Fredlund and Morgenstern (1976) proposed and
tested constitutive relations to link the stress and
deformation state variables. The proposed constitu-
tive relationship for the soil structure is given by [2]
and the relationship for the water phase is given by [3]
(Fig. 1).
[2] +
A V / V = mlsd(o - Z I ~ ) m Z ~ L I -
( Uilw)
~
[3] A V,/ V = V Z ~ ~ C-/ ( C + m 2w~l(~,
- u,")
where in: = compressibility of the soil structure
when cl(u, - u,) is zero; mz" compressibility of the
soil structure when d(o - 11,) is zero; mlw = slope of
the (o - 11,) plot when d(u, - 11,~)is zero; and mzw=
slope of the (u, - 11,) plot when d(o - u,) is zero.
The constitutive relationship for the air phase is the
difference between [2] and [3] because of the con-
tinuity requirement.
[4] AV,/ V = nzlncl(o - u,) + mzad(ua - u,)
where mla = slope of the (o - u,) plot when
&u, - uw)is zero, and r?zza = slope of the (u, - tr,)
plot when d(o - u,) is zero.
 FREDLUND AND HASAN 523

Flow of the water phase is described by Darcy's MASS LEAVING

law (Childs and Collis-George 1950).

where v = water velocity; kw = coefficient of perme-

ability with respect to the water phase; yw = density
of water; and y = depth in the y-direction.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

Flow of the air phase is described by Fick's law

(Blight 1971).

where v, = mass rate of air flow; D = a transmission

constant having the same units as coefficient of
permeability; p = absolute air pressure (i.e., u, +
tr,~,); and u,tm = atmospheric air pressure.
The isothermal compressibility equation for the
air phase, Pa, is (Fredlund 1976):

and the isothermal compressibility equation of an

air-water mixture (with no diffusion) in the presence AIR I I WATER
of a particulate mass, P,, is:
For personal use only.

where S = initial degree of saturation, and B,, =
pore-pressure coefficient equal to A~lu,/Au,. The
above listed physical relationships are sufficient to
derive the one-dimensional consolidation (or swell-
ing) equations and the pressure boundary condition
equations for an unsaturated soil.
Derivation of the Consolidation Equations
The one-dimensional consolidation equations for
unsaturated soils are derived using the conventional
assumptions for Terzaghi's consolidation theory
with the following additions:
(i) The air phase is continuous.
(ii) The coefficients of permeability with respect to
water and air, and the volume change moduli remain
constant during the transient processes.
(iii) The effects of air diffusing through water and
the movement of water vapor are ignored.
The above assumptions are not completely accu-
rate for all cases; however, they are reasonable for a
first attempt to derive a general consolidation theory
for unsaturated soils.
After applying a load to an unsaturated soil, there
will be a dissipation of the excess pore-air and pore-
water pressures. In order to compute these values as
a function of time, it is necessary to have two
equations. This is accomplished by independently
considering the continuity of the water and air
phases. Then the derived equations are solved
MASS ENTERING
FIG. 2. A referential element in the soil mass.
simultaneously to give the water and air pressures at
any elapsed time.
Water Phase Pcrrtial D~fferentialEquation
Let us consider a referential soil element as shown
in Fig. 2. The water phase is assumed incompressible.
For the consolidation process, water flows out of the
element with time. The constitutive relationship for
the water phase defines the volume of water in the
element for any combination of total, air, and water
pressures. The volume of water entering and leaving
the element in the y-direction is described by Darcy's
law as:
[9] Volume entering = (-k,/y,)(au,/dy) dx dz
The net flux of water in the element is:
Equation [lo] can be equated to the constitutive
relationship for the water phase ([3]) in accordance
with the continuity requirement ([I]).
For the consolidation process the change in total
stress with respect to time is set to zero. Simplifying
and rearranging [I I], the water phase partial differen-
tial equation can be written:
[12] auw/at = - cW(au,/at) +~,~(a~~,/ap)
 524 CAN. GEOTECH. J. VOL. 16, 1979

where C , = (1 - ~ n ~ ~ / m ~ ~ ) / ( r n z " /and

t ? zis~ 'called
~) the air phase (i.e., [4]), gives:
the interactive constant associated with the water
phase equation. This equation may be further
simplified by defining R , as n?z"/tnlW.When the soil
is saturated, Rw = 1. cVw= (l/R,v)(k,/y,tn,") and
is called the coefficient of consolidation with respect
to the water phase.
As a soil becomes saturated, the interactive con-
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

stant approaches zero and [12] reverts to Terzaghi's The change in total stress with respect to time can
one-dimensional consolidation equation. The fact be set to zero for the consolidation process. Simpli-
that the air phase cannot remain continuous as a soil fying and rearranging [19], the air phase partial
goes towards saturation is a physical restriction; differential equation can be written as follows.
however, it does not pose a mathematical problem.
Equation [12] also reverts to Terzaghi's equation
whenever the air pressure induced is small.

Air Phase Partial Differential Equation m za/m I"

C, =
The air phase is compressible and flows indepen-
dent of the water phase when subjected to an air
( 1 - in za/nz la) + (&(1f-~iatrn)t721"
S)n

pressure gradient. As well, the constitutive relation-

ship for the air phase defines the volume of air in the and is called the interactive constant associated with
element for any combination of the total, air, and the air phase equation. This equation may be further
water pressures. According to Fick's law the mass of defining R, as r?z2a/n21a.
For personal use only.

air entering the element in the y-direction is:

[13] Mass entering = - D(dp/dy) d x dz
The net mass flux of air in the element is:
where nz = mass of air in the element.
The mass rate of change is written in terms of a
volume rate of change by differentiating the relation-
ship between mass and volume.
For isothermal conditions the density of air, y,, is:
where w = molecular weight of the mass of air;
R = universal gas constant; and 0 = absolute
temperature.
The mass of air is written in terms of the density of
air, the degree of saturation, S, and the porosity of
the soil, n.
Substituting [14], [16], and [17] into [15] gives:
cv- DRB 1
= -
w (1 - Ra) (u, + u,t,)nzla + (1 - S)n
and is called the coefficient of consolidation with
r e s ~ e c to
t the air ~ h a s e .
As a soil becomes completely dry, the interactive
constant approaches zero and [20] reverts to the form
presented by Blight (197 1).
The dissipation of the excess pressure of the pore-
air and pore-water phases is obtained by a simul-
taneous solution of [12] and [20] using the finite
difference technique described in the Appendix.
The results can be ex~ressedin a dimensionless
form by defining an average degree of consolidation
and time factor for each of the fluid phases. The
average degree of consolidation for the water phase
is:
where Uw = average degree of consolidation with
respect to the water phase; uwi = initial water
pressure; u, = water pressure at any time; and H =
length of drainage path.
The time factor for the water phase is:

[18]
d(Va/V) - - DRB d2tla
dt wp dy2
+ (1 P S)n at
- alla
- [22] Tw = cvwt/H2

where t = elapsed time.

Equating [la] to the constitutive relationship for Similarly, the average degree of consolidation and
 FREDLUND AND HASAN 525

I I 1 1 1 1 1 1 1 I I I 1 1 1 1 1 1 I I 1 1 1 1 1 1 1

0.1

0.2
Cw
C,
RANGING
RANGING
FROM
FROM
0.0 TO
0.0 TO
5.0
1.0 -1
0.3
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

0.4

0.5

0.6

0.7
ALL Ca VALUES
0.8

0.9

I I 1 1 1 1 1 1 1 I
0.01 0.I 1.o 10
TIME FACTOR ( WATER PHASE
For personal use only.

FIG. 3. Dimensionless time factor versus degree of consolidation curves for the water phase.

I
I the dimensionless time factor with respect to the air during the process. The derivations follow a form
phase are defined as: similar to that presented above but the coefficients of
permeability are treated as variables during differen-
tiation. The pore-water partial differential equation
now becomes:
and

where U, = average degree of consolidation with where c,," = I/R,,.y,rn~".

respect to the air phase; T, = time factor with The pore-air partial differential equation becomes
respect to the air phase ; u,i = initial air pressure ; and
u, = air pressure at any time.
Figure 3 shows the water phase degree of con-
solidation versus time factor curves for various air-
water interaction constants. Similar curves for the
air phase are shown in Fig. 4. The interactive
constant in the air phase partial differential equation
was assumed to be constant for the calculation of
pore-air pressure dissipation. The curves cover
anticipated reasonable ranges for the soil moduli.
They also show a smooth transition towards the case
of a completely saturated soil (Terzaghi 1936) and These equations can readily be solved using a
the case of a completely dry soil (Blight 1971). finite difference technique; however, it is difficult to
present the solutions in dinlensionless form due to
Varying Permeabilities During Consolidation nonlinearity.
The partial differential equation for the pore-water
and pore-air phases can also be derived for the case Pore Pressure Boundary Condition Equations
where the coefficients of permeability are variables When an external load is applied to an element of
 CAN. GEOTECH. J. VOL. 16, 1979

Ca RANGING FROM 0.0 TO 1.0

\
\
\ C, RANGING FROM 0.0 TO 5 . 0
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

TIME FACTOR (AIR PHASE

FIG.4. Dimensionless time factor versus degree of consolidation curves for the air phase.
For personal use only.

unsaturated soil, instantaneous compression occurs
under undrained conditions and excess pressures are
induced in the air and water phases. Two equations
are necessary to predict the relative magnitudes of
the excess pore-air and pore-water pressures. The
pore pressures depend upon the compressibility of
the soil structure, the air and water phases. In addi-
tion, the contractile skin has an effect on the relative
changes in the pore-air and pore-water pressures.
The induced pore pressures form the boundary con-
ditions for the consolidation process.
The overall continuity requirement for the soil
requires that the compression of the soil structure
must equal the compression associated with the
pore-fluid phases. This is satisfied by equating [2]and
[8] and using the water phase as the reference phase
for the pore-fluid pressure (Fredlund 1976).
[27] tnlsA(o - u,) + m?8A(ua - u,) =
Equation [28]involves two unknowns (i.e., Au, and
Au,) and therefore another equation is required for
its solution. A second equation is also logical since
only one of two necessary constitutive relations has
been used in formulating [28].The second constitutive
relation can be incorporated by considering the
continuity of the air phase. Volume change described
by the compression of the air phase must equal the
volume change defined by the air phase constitutive
relationship.
+
[29] mlaA(o - 11,) tn2"A(u, - 11,) =
+
[(I - S)n/(lL ~~,trn>lAli,
Simplifying [29] and solving for the change in
pore-air pressure gives:

Defining Rs = m2/1nlg,[27]can be solved for the

change in pore-water pressure. Equations [28] and [30] can be solved for the
changes in pore-air and pore-water pressures result-
ing from a change in the applied load. Let us simplify
these equations by defining the following variables.
 FREDLUND AND HASAN

I N I T I A L STRESS
POINT

INSTANT AFTER
INITIAL CONDITIONS
LOADING \ .
'RESS
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

I CHANGE IN PRESSURES I I /

FIG.6. Initial and final stress points on a constitutive

surface for the consolidation process. * kPa.

Example Problem
An example problem is solved to demonstrate the
For personal use only.

solution of boundary condition equations (i.e., [37]

and [38]) and the partial differential equations for the
I pore-air and pore-water phases (i.e., [I21 and [20]).
FIG. 5. Initial and final boundary conditions for a con- Let us assume that a compacted soil layer overlying
solidation process. an impervious surface is initially in equilibrium with
the stresses shown in Fig. 5. The initial equilibrium
conditions are altered by changing the total stress
applied to the top of the soil. The instantaneous
change in total stress of 689 kPa (100 psi) produced
a corresponding pore-air and pore-water pressure
change of 41.4 kPa (6 psi) and 303 kPa (44 psi),
respectively. The initial and final stresses (and stress
state variables) with the assumed soil properties are
shown in Table I. The corresponding stress state
variable changes on the constitutive surface are
shown in Fig. 6. The dissipation of the excess pore-
Therefore, [28] and [30] can be written: water and pore-air pressures is obtained by simul-
taneous solution of [I21 and [20].
[35] Au, = RlAu, + R~AG The average degree of consolidation with respect to
R3Allw - R4Ac the water phase (U,) is plotted against the time
[36] Au, =
factor for water phase (T,) in Fig. 7. It shows that
Equations [35] and [36] are combined and solved the U, versus T, plot is only slightly different than
to give two pore pressure coefficients that can be the conventional Terzaghi consolidation plot for
used to compute the boundary conditions. saturated soils. This deviation of the U, versus T,
plot from the conventional Terzaghi plot can be
[37] B, = Au,/AG = (Rz - RlR4)/(1 - R1R3) attributed to the interaction effect in the simul-
taneous solution of the two partial differential
equations.
Equations [37] and [38] require an iterative tech- The average degree of consolidation with respect
nique in their solution since RI, R3, and R4 contain to the air phase (U,) plotted against a time factor
the absolute pore-air pressure value. with respect to the air phase (T,) is presented in
 528 CAN. GEOTECH. J. VOL. 16, 1979

TABLEI. Initial and final stress conditions in the soil

(r L{n k v ((r - l d (14, - ~ 1 , ~ )

Stresses (psi (kPa)) (psi* (kPa)) (psi (kPa)) (psi (kPa)) (psi (kPa))
Initial condition 100.0 (690) 0.0 (0.0) -60.0 (-414) 100.0 (690) 60.0 (414)
Instant after loading
= $6 psi (41 kPa); An, = +44 psi (303 kPa) 200.0 (1380) 6 . 0 (41) -16.0 (-110) 194.0 (1338) 22.0 (152)
Final conditions 200.0 (1380) 0.0 (0.0) -60.0 (-414) 200.0 (1380) 60.0 (414)
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

NOTES:S = 50%; ,I = 50%; m l a = 0.0008 in.l/lb (1.16 X lo-' m2/kN); ntlS = 0.001 in."lb (1.45 X m2/kN); R. = -0.01 ; Rw = 0.5; and Rs =0.5.
*Gauge.

' O r A C ~ = I O O P( 6S9IO k P a ) SOLUTION

A u W = 4 4PSI ( 3 0 3 k ~ a )
Aua 6 PSI (41 k P a )
For personal use only.

70 rn; = 0.001 I N ~ / L B ( 1.45 x 1 6M~

2 / k~ )

TIME FACTOR ( WATER PHASE 1

FIG. 7. Average degree of consolidation versus time factor.

I A u = 100 PSI ( 6 9 0 k P a )

\
A u w - 4 4 PSI ( 3 0 3 kPa)
A u =~6 PSI ( 4 1 kPa
rn: = 0.001 IN.^ / L B ( 1.45 x M2 / k N
70 S=50%

TlME FACTOR ( A I R PHASE 1

FIG. 8. Average degree of consolidation versus time factor.
 FREDLUND AND HASAN
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13
TIME
FIG.9. Percent mid-plane pore-air and pore-water pressure versus time factor.
For personal use only.
FIG. 10. Dimensionless isochrones for the water phase.
Fig. 8. It also shows a slight deviation from the
Terzaghi solution for saturated soils. It should be
noted that the c," term has been assumed constant
during the consolidation process. To do this, the
absolute air pressure has been set to a constant equal
to the average air pressure during the consolidation
process. The error involved for this example problem
is negligible.
Figure 9 shows the percent mid-plane pore-water
529
FACTOR
FIG. 11. Dimensionless isochrones for the air phase.
and pore-air pressures plotted against the dimen-
sionless time factors for both of the fluid phases.
Dimensionless isochrones for both the water and the
air phases are shown in Figs. 10 and 11 respectively.
The plots show only slight deviations from the con-
ventional Terzaghi plots. Other example problems
have shown the interaction effects between the air
and water phase partial differential equations to be
more extreme.
 CAN. GEOTECH. J. VOL. 16, 1979
t iAz
-I6'
-16 A
I
U",(Itl,j)
I
1 -I6
u," ( I , ] )
*
U,"(I,J+l)
I ,
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13
WATER
FIG. Al. Finite difference mesh for the consolidation equations.
Summarv
For personal use only.
All the necessary physical relationships are avail-
able for a complete formulation of the one-dimen-
sional consolidation problem in unsaturated soils.
The formulation proceeds along lines similar to that
of Terzaghi's conventional theory for saturated soils.
However, it is necessary to have one partial differen-
tial equation for the air phase and another partial
differential equation for the water phase (i.e., [I21 and
[20]). They must be solved simultaneously.
For the completely dry and saturated cases, there
is a smooth transition to the conventional solutions.
At intermediate degrees of saturation there are
varying amounts of interaction between the air and
water phases during the dissipation process.
The necessary physical relationships are also
available for a rigorous formulation of the pore-air
and pore-water pressures generated during undrained
loading. Once again two equations are formulated
(i.e., [37] and [38]) that must be solved for the pore-air
and pore-water boundary conditions to the con-
solidation problems.
Acknowledgements
The authors would like to acknowledge the
Department of Highways and The Saskatchewan
Research Council, Government of Saskatchewan, for
their interest and financial support of research into
transient,processes in unsaturated soils.
AITCHISON, G. D., RUSSAM, K., and RICHARDS, B. G. 1965.
Engineering concepts of moisture equilibria and moisture
changes in soils. Itz Moisture equilibria and moisture changes
PHASE
in soils beneath covered areas. Review Panel Presentation.
Butterworth, Sydney, Australia, pp. 7-21.
BARDEN, L. 1965. Consolidation of compacted and unsaturated
clays. GCotechnique, 15(3), pp. 267-286.
BIOT,M. A. 1941. General theory of three-dimensional con-
solidation. Journal of Applied Physics, 12(2), p. 155.
BLIGHT,G. E. 1971. Flow of air through soils. ASCE Journal
of the Soil Mechanics and Foundations Division, 97(SM4),
pp. 607-624.
CHILDS,E. C., and COLLIS-GEORGE, N. 1950. The permeability
of porous material. Proceedings of the Royal Society, 201A.
FREDLUND, D. G. 1973. Volume change behavior of unsatu-
rated soils. Ph.D. thesis, University of Alberta, Edmonton,
Alta.
-- 1976. Density and compressibility characteristics of
air-water mixtures. Canadian Geotechnical Journal, 13,
pp. 386-396.
FREDLUND, D. G., and MORGENSTERN, N. R. 1976. Constitu-
tive relations for volume change in unsaturated soils.
Canadian Geotechnical Journal, 13, pp. 261-276.
1977. Stress state variables for unsaturated soils.
ASCE Journal of the Geotechnical Engineering Division, 103
(GT5), pp. 447-466.
HASAN,J. U. 1977. Transient flow processes in unsaturated
soils. MSc. thesis, University of Saskatchewan, Saskatoon,
Sask.
SCOTT, R. F. 1963. Principles of soil mechanics. Addison-
Wesley Publishing Co. Inc., London, England.
TERZAGHI, K. 1936. The shearing resistance of saturated soils.
Proceedings, 1st International Conference on Soil Mech-
anics, Vol. 1.
Appendix
An explicit finite difference technique is used t o
solve each of the partial differential equations. Since
both equations must be solved simultaneously, the
solution iterates between the two equations. In
addition, one of the equations is nonlinear.
 FREDLUND AND HASAN 531

Figure A1 shows the finite difference grids for the The ~ , w A t / ( A y )term
~ has been set to 0.25 for the
. problem and defines the related variables.
example solution of the example - -problem.
Following are the finite difference forms for the
Finite Difference For177 of'tlze Air Phase Partial
partial differential equations.
Differential Equation
Finite D~fferenceFor117 of the water P / ? ( I sPartial
~ [A21 Lla(i,j + 1) = Ll,(i,j ) - C,[u,(i, j + 1)
DifJerential Equation - + +
uJi, j)] cVaAt/(Ay)'[ua(i 1,j )
+ + +
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by University of Queensland on 10/02/13

[All u,(i, j 1) = u J i , j ) - Cw[ua(i,

j 1) ua(i - 1 , j ) - 2ua(i,j)]
+
- udi,j)] C , ~ A ~ / ( A . Y ) ~3-
[ L1~, j,)~ ( ~ The cVaAt/(Ay)' term has again been set to 0.25.
+ u,(i - 1,j ) - 21lW(i,j)]
For personal use only.