Session 1/9

Approximate Solution of the Progress of Consolidation

in a Sediment
Solution approximative du processus de consolidation dans un sédiment

by R . G r a n O l s s o n , Dr. Ing., Professor o f M echanics, N orw egian Institute o f Technology, Trondheim, N orw ay, V isiting Pro­
fessor at Brown U niversity, Providence, R .I., U .S .A .

Summary Sommaire

In a sediment the vertical pressure depends on the depth below La pression verticale dans un sédiment dépend de la profondeur
ground level and the time which has elapsed since the sedimentation au-dessous du niveau du sol et du temps écoulé depuis le début de
started. In this paper the vertical effective grain pressure is expressed la sédimentation.
as a product o f a function o f depth and time, respectively. Cet article introduit une équation exprimant la pression verticale
Using the two-dimensional continuity equation for a compressible réelle entre les particules en fonction de la profondeur et du temps.
soil and Darcy’s law for seepage through porous materials, one Partant de la supposition que la distribution de la pression est une
arrives at a differential equation of the first order for the time func­ fonction exponentielle, le résultat est donné par un graphique. Une
tion, ’Q(t), the solution of which contains the incomplete Gamma- solution de la distribution linéaire de la pression est également ex­
function. posée.
Assuming the pressure distribution, with respect to depth, to be L’article comprend en plus une méthode pour le calcul de la distri­
given by an exponential function, a complete numerical calculation bution de la pression pendant la consolidation, une fois la sédi­
is carried out and plotted on arithmetic and semi-logarithmic papers. mentation achevée.
An approximate solution, given by the exponential integral, is added
on both plots. It is also demonstrated how one can obtain the solu­
tion for linear pressure distribution as a special case, in which the
more general solution is reduced to contain the error integral as the
main part.
In this paper it is also shown how one can calculate the pressure
distribution during the consolidation which will take place after
sedimentation is completed. In this case the time function is obtained
by integration of Darcy’s equation, leading to exponential functions.
From the origin 'Q(/j), located on the £-curve, the corresponding cur­
ves are drawn tangential to the asymptote t ( 0 )/( I ) = 1. By means
of this set of curves the effective grain pressure distribution is de­
termined both during and rfter the sedimentation process.

In tro d u ctio n

K. Terzaghi (1924), in one o f the earlier editions o f his book
on soil m echanics, has given a treatment o f the progress o f
consolidation in a sediment which is being deposited at a co n ­
stant rate per unit o f time. For the mathematical treatment o f
this problem , Terzaghi established a linear differential equation
o f the first order, the solution o f which was given in integral
form . In a paper published som e years ago the writer presented
a solution given by the error integral and elementary functions,
satisfying the boundary conditions at any tim e (Gran Olsson,
1949).
K . Terzaghi made the assum ption that the pressure trans­
mitted from grain to grain in the sedim ent is at any given time
constant, is similar and modified only by the increasing height
o f the sedim ent and a tim e factor £ which is a function o f time
only. For this condition it will be show n that the differential
equation o f the problem m ay be solved by the incom plete

38
Gam m a-function being tabulated in an extension satisfying all
yx vdx (4 a)
practical purposes. The p ro o f o f this im portant property o f
- ’ - Ï
the equation o f Terzaghi will be given in this paper.
in which the discharge velocity formula, Equation (3 a), must
be introduced.
Symbols and N otations
For the effective grain pressure p , Terzaghi (1924) assumes
In the follow ing investigation the sym bols com m only em ­ an exponential function, where the power exponent contains
ployed in soil m echanics are used. Let the depth x as well as a time function t. The latter is found
by solving a differential equation for f.
c =
The expression for the grain pressure p may, with som e
C = constant o f integration
m odification, be written as follow s:
e =
h = p = y h ( \ — e~^xlh) ........................................................ (5)
i = in which e is the base o f the Napierian logarithm.
k = Particularly, when taking only the two first terms into ac­
mv = count in a series for e ~ ^ h, Equation (5) yields
n = P = Z vx ................................................................................... (5)
o f soil For this grain pressure variation, which is linear with respect
P = effective norm al pressure per unit o f area
to depth, solutions are given by Terzaghi (1924) and Gran Olsson
<1 = (1949).
If The follow ing derivations will be based on the assumption
u = that the grain pressure profiles are similar with respect to depth
V =
and time. In terms o f mathematics:
=
a ,ß = P = yh £{t) m ................................................................. (5b)
in which equation the depth x is represented by the dim ension-
ri =
— submerged unit weigth (y = y l — y w) less free variable f = x/h.
y
= The change in grain pressure is obtained from Equation (5 b)
yw
r dp dp 3C df
•- dp = — d t + — dx = y h ( 6)
; = dt dx at at;
f x /h = ratio between depth x and head o f sediment, h. dx q
and hence, since — = — ,
dt y
Derivation o f the Differential Equation for the Time
ac 3 /,
Function f -^ -d x = y h
d (6 a)
dt dt y
The requirement for continuous flow o f water leads to the Since / ( f ) and f ( t ) are functions o f f and t , respectively, the
follow ing equation between the discharge velocity v , and the partial derivatives 3 may be replaced by the usual sym bols d ,
pore water pressure u ( Terzaghi, 1946) leading to
9v a dv du j (i) dp ,
— dx ■y h dx + d m c ( t ) - (6 b)
T x ix m vJ 7 dt dt y
The sum o f the change in pore water pressure and the change The com bination o f Equations (3 a) and (6 b) yields
in grain pressure p m ust be zero, hence l1
dp du (3 b)
(2) v = m vy h j ^ .f ^ d x + d n m « )1
d1 dt dt y_
The com bination o f Equations (1) and (2) yields:
Water taô/e
, dp
dv = m .. — d x (3)
dt
from which
h

J
f dPA
m v ——d x
dt
(3 a)

In this equation h designates the thickness o f the entire layer

o f sediment, while x represents an arbitrary depth below the
ground surface at a certain tim e t , see Fig. 1.
A ccording to D arcv's law the discharge velocity v may be
expressed as follow s:
k d ( y x — p)
(4)
Y ,„ dx Jmpermeaò/e ¿ose
where (y x — p) is the pore water pressure, since y x represents ôa se /mpermécrô/e
the total pressure at depth jc while p is the effective grain pres­ Fig. 1 O n e-D im en sio n a l F low in a Sedim en t
sure. Integration o f Equation (4) leads to: F iltration un iaxiale dans un séd im ent

39
or, when integrated The com bination o f Equations (10a) and (11) offers the com ­
plete solution
v = m vy h ^ {F ( 1) - F (f)} /) + - £ (0 { / ( l ) - / ( I ) } (3c)
at y £ 0 0 = [C 0 — / ( I ) " 1 / z~ae~z d z ] zaez .. ( 1 0 b)
o
where F ( l) — F(£) = f /( £ ) d $ . The integral is known as the incom plete G am m a-function, the
i numerical values o f which may be found in Pearson (1946).
The com bination o f Equations (4 a), (5 a) and (3 c) leads to In order to determine C 0 one m ust require that C is finite
the follow ing for / = 0 , that is for z = oo, or
OO

G (f) + G(0)] A» + Co = m ) - 1 f z - ae - * d z = / ( l ) - i r ( l - a ) . . . . ( 1 1 a)
o
where T (1 — a) denotes the com plete G am m a-function, hence
+ - C ( 0 [ / ( D f - F ( { ) + F(0)]A (4b)
Y C0 0 = / ( l ) - 1 { r ( l — a) — j z - ae - z d z ) z aez . . . . ( 1 0 c)
f o
where G(£) — G (0) = f F (() d'Q. The ordinary sym bols for the /"-function is obtained by intro­
o
ducing n — 1 = — a, after which
For x = h = — , that is £ = 1, Equation (4b) is reduced to
7 C(z) = / d ) - 1 [^ (« ) — y ( n , z ) ] z - ”+ 1e * ............................. (lO d)
y wm v g2t dC 1 where y (n ,z ) designates the incom plete / ’- function, as defined
[1 — ¿ ( 0 / ( 1 )] = 1 -f- + - a ^ C (4c)
ky dt t by the follow ing relationship,
Z
where y ( n ,z ) = j z n~'i e~z d z .
o
a ? = / ( I ) - F ( 1) + F (0) i Equation (10 d) represents the general solution o f the differen­
P = F ( 1 ) - G ( 1 ) + G(0) f
tial Equation (8 a).
The follow ing abbreviation for the “ time con stan t” is intro­
duced
Example with Exponential Pressure Distribution
/(l) ky2
(7) According to the pressure distribution curves, which are
P Y w m v ‘f
given by Terzaghi (1924, p. 171, Fig. 30), it appears likely that
by means o f which E quation (4 c) becomes the distribution can be sufficiently approxim ated by the follow ­
dZ (a c\ c 1 ing expression,
c = ( 8)
\Tt + \ l + 7>) /o o i* p = y h U 0 (1 - e - f ) (12 )
This is the general differential equation in t, provided that the hence,
effective grain pressure distribution is given by Equation (5b).
m = 1 - ¿-i; i.e. / ( I ) = 1 - -
The boundary condition requires that the pore water pressure
y x — P = y h [f — £ (0 / ( f ) ] F(f) = f + F (0) = 1
be equal to zero for t = 0 and x = h, i.e. t = 1. T his require­
m ent is seen to be satisfied when C(0 ) / ( 1 ) = 1 , from which G(S) = H 2 - ^ ; i.e. G (l) = J -------; G (0) = - 1
relationship one obtains the only constant o f integration which e
w ill appear in the solution o f the differential Equation ( 8). w hereof

= F ( l) - (7(1) + G(0) = •
2e
Solution o f the Differential Equation (8)
A general solution o f Equation ( 8 ) may be established for *P = / ( I ) - F ( l) + F (0) = 1 - -
any values o f the parameters a, c and / ( l ) , by substituting yielding

z = — i.e., d z = — c t-^ d t (9) /(I ) ky2 2 (e — 1 ) ky2

t
P y wm vq 4 — e y w m vqi
after which E quation (8 ) reads
2 (e — 2 )
dt (a \ = 1,1208
I C -------/ ( I ) “1 (8 a)
dz \z + ) '
8 — 3e
n = 1 _ a = ----------= — 0,1208
N eglecting the term on the right hand side o f this equation one 4— e
arrives at the hom ogeneous solution
Introducing these values o f n and / ( l ) into Equation (1 0 d)
CoOO = c0z aez ...................................... ( 10 ) one can evaluate f for all values o f z = c /i. H ence the grain
The com plete solution is then obtained by varying the constant pressure p is found for any time, t, and depth, x = i h , by
o f integration means o f Equation (12).
For the numerical com putation, however, it is preferable to
C(z) = [C 0 + Cfi(z)] z aez ........................................................ (10a)
substitute the W hittaker function (1935) defined as
where Cp (z) is found by introducing Equation (10a) into Equa­
tion (8 a), yielding e 2z k —J + m
Cp(z) = — / ( I ) - 1 / z~ae~z d z .. (11)
Wk, m(z) - k —4 + m
K ) p dtp

40
The relationship between the W hittaker- and the Gam ma-
function is the follow ing
71— 1 Z
r { n ) — y (n, z) = z~2~ e 2 W ”- i » (z)
2 ’2
In our case, the W hittaker parameters, k and m, are seen to be
n— 1 n
m ■ 2

after which OO

_£ r/ <p\n _i
W<n- 1 n (z) = e 2 Z 2 I I 1 H-----1 e ~ <p dtp

since F ( l ) = 1. H ence, the com plete solution for 'Q(z) becomes

Fig. 3 T im e F u n ction s £ (t) in L ogarithm ic T im e Scale
F o n ctio n s du tem ps £ (t) en éch elle logarith m iqu e
= / ( D - ij | i + ^ y V f« /?

Since the integral, E i( — z), is tabulated, this approxim ate

U sing this expression t( z ) is evaluated for various values o f c
c solution is readily evaluated for various values o f z = —, see
z = The arithm etic and sem i-logarithm ic plots are found
Figs. 2 and 3. From these plots it is seen that the approxim a­
in Figs. 2 and 3, respectively. tion is fairly good at the beginning o f the process o f sedimenta­
(a) A n approxim ate solution for 'Q(z) is obtained by intro­ tion, but the discrepancy between the tw o curves (a = 1 and
ducing a = 1 (instead o f 1.1208) that is n = 0. Hence a = 1.1208) increases as the sedim entation progresses.
(6 ) In case the pressure distribution is linear, Equations (5 a)
F (0) — y ( 0 ,z ) = z ~ i e 2 W - li0 (z) = — E i( — z)
and (5 b),
where E i( — z) denotes the exponential integral, defined by
f(S ) = f ; i.e. / ( I ) = 1
W hittaker and W atson (1935, p. 341) as
OO F ( 0 = i f 2; i.e. F ( l) = i , F (0) = 0
C e -v
— E i{ — z) = ----- d<p G (f) = I S 3 ; i.e. G (l) - 1 , G(0) = 0
J f from which
Z
Therefore, the approxim ate solution for n = 0 becomes
OO P = i — s + 0 = I
aft = 1 — i + 0 = i
Uz) = / ( l ) _1z e zj*^— dtp
a = § or n = — \
while
w
, ky2
c = 3 y----------
wm vq 22
0.9 '\ - .

The general solution is obtained from Equation (10b), intro­

ducing a = I, / ( l ) = 1
0.8

t(z ) = [c„ — J z 2 e~z d z j\ z 2 ez

0.7 1
Substituting z = r2 the above equation is written as follow s
0.6 C(r) = [C 0 + 2»—1 e ~ r* + 2 n i& (r )]r 3 er‘
rM ,
V 2c+t where 0 ( r ) denotes the error integral. The constant o f inte­
os gration becomes C0 = — I n 1, after which
^ ^ w f t i i - [ r r - ^ H ) - f ( ÿ a e ^ j ( ÿ ] ( ÿ e c/t £(/•) = [r - 1 e - r’ + n * 0 ( r ) - l ] 2 r3 er‘
OA
f \
W This particular solution is identical with the equation given by
Gran Olsson (1949, p. 340, last equation), when r2 is replaced
0.3
Vj' V by z.
^

W fM - a t)

0.2 The Consolidation o f the Soil after the Sedimentation is

1/

Completed
0.1 The sedim entation process is assumed to be com pleted at
a time t = tv In the follow ing an attem pt will be made in
t/c -
order to establish an approxim ate solution for the consolida­
10 12 n 16
tion’ o f the soil at a time t > tx.
Fig. 2 T im e F u nctions~£ (t) in Linear T im e Scale A s soon as the depositing on the surface ceases, i.e. q = 0,
G raphique représentant les fo n ctio n s du tem ps £ (/) par rapport dx
à un e éch elle lin éaire des tem ps one can introduce — = 0. H ence, from Equation (6 a),
at
41
 between this decrease in pressure and the discharge velocity

m vy h l ^ [ F ( \ ) - F ( 0 ) ] [1 ■ £ /(!)] •• (4d)
dt yw
The tim e t = h corresponds to the m om ent at which the sed i­
m entation process is com pleted, and the value £ (O m ay be
extrapolated from Figs. 2 or 3.
W hen solving Equation (4d) with respect to d t, and inte­
grating one obtains

1 - £ ( 0 / ( D = [1 - £ ( 0 / ( 1 ) ] e x p J - ^ ^ 3 { t— td [ F ( 1 ) - F ( 0 )
{ y wm vhi
Introducing the “ time co n sta n t” , defined in Equation (7), one
obtains

1 — £ ( 0 /( 1 ) - [1— £ ( 0 / ( 1 ) ] e x p .jc ( f ~ fl) [ / ( l ) — a/?]-1! (13)

P t\
where a and P have the sam e m eaning as before.
A pplying the pressure distribution given by Equation (12),
E quation (13) becomes

1 - £ ( 0 / ( D = [1 - £ ( 0 / 0 )] exp. (13a)
H H ( H ]
or in a m ore symmetric way,

[1 — £ ( i ) /( l) ] e x p . ^ - 2 ^ = [ l - £ ( 0 /(l)] e x p . ^ - 2^

Since 2 — e l 2 = 0.641 the above equation

ion m ay alsc
also be written
Fig. 4 T im e F u n ctio n s £ (/) after th e Sedim en tation is C om pleted
F o n ctio n s du tem ps £ (0 un e fo is la séd im en ta tio n achevée 1 — £ ( 0 / ( 0 = [ 1 — £ ( 0 ] exp. 0.641 (13 b)
K H I
By means o f this equation, £ ( 0 /( 1 ) m ay be evaluated for each
^ d x = y h ^ f(Ç )d x .. ( 6 c) chosen point (tiCi) on the curve £(<)• C hoosing several initial
at at points (tlt Ci) one will obtain a set o f curves as show n in Fig. 4,
A ccording to Equation (3 b) the discharge velocity o f the water defining the consolidation o f the soil after the com pletion o f
percolating through the sedim entated soil becomes the sedimentation.
i
References
/ ( f ) dS = m vy h \ j - ( [F (l) - F (0)] (3d)
Gran O lsson, R . (1949): Q uarterly o f A pp lied M ath em atics, v o l. V II,
pp. 3 3 8 -3 4 2 .
The decrease in pressure per unit area, causing the water to P earson , K . (1946): T ab les o f th e In com p lete r -F u n c tio n . C am bridge
flow upwards, is U n iversity Press.
T erzagh i, K . (1924): E rdbaum echanik a u f b o d en p h ysik alisch er G rund-
lage. L eipzig und W ien, p. 175.
A {i-£/(i)}J=r{i-£/(D } Terzagh i, K. (1946): T h eoretical S o il M echan ics. N e w Y o rk , p. 270.
«i
W h itta k er, E. T. and W atson, G. N . (1935): A C ou rse o f M od ern A n a ­
A ccording to D arcy's law the follow ing relationship exists lysis. C am bridge, 4. ed ., p. 339.

42