Determination of Organic Soils Permeability using the Piezocone

Dissipation Test

VAN BAARS, S.
Delft University of Technology, The Netherlands

VAN DE GRAAF H.C.

Lankelma Geotechniek-Zuid, The Netherlands

Key Terms: cone penetration test, dissipation, hydraulic conductivity, permeability,

piezocone, site investigation

ABSTRACT

The hydraulic conductivity, which is among other parameters, needed to predict time-
dependant settlement, can be estimated using a piezocone dissipation test. This test is
commonly used together with cone penetration tests (CPTs), soil borings and laboratory
tests in foundation site investigations. The piezocone dissipation test is based on the fact
that the rate of decay of the large excess pore water pressures generated during
penetration of the piezocone through saturated clays and silts depends on the hydraulic
conductivity of the material. However, interpretation of the dissipation curve is often
problematic as existing methods of analysis assume a continuous decrease of pore
pressure with time, whereas actual dissipation curves often exhibit non-standard
behavior, the interpretation of which is more complex.

This paper presents a new method of interpretation that can be used to estimate the
hydraulic conductivity regardless of the shape of the dissipation curve. Examples of
results using the new analysis method are compared with results obtained using
laboratory oedometer testing.

INTRODUCTION

In a piezocone test, the surrounding soil is compressed by the penetration of the cone tip,
which create excess pore water pressures. At the same time, the shearing of the soil by the
cone tip may also result in dilatant behavior, which results in a negative excess pore pressure.
Depending on the relative magnitudes of the pore pressures created by the compression and
shearing actions, the resultant excess pore pressure may be less than, equal to, or greater than
the initial hydrostatic pressure. In clean sands and gravels, essentially drained response is
observed and measured pore pressures are hydrostatic. In most other cases, an initial
undrained response occurs that is followed by drainage. Once penetration has stopped, the
excess pressures will dissipate with time and eventually reach the pre-test hydrostatic value.
Since the rate at which the pressures dissipate depends on the hydraulic conductivity of the

1
surrounding medium, the pressure dissipation curve can be used to estimate the hydraulic
conductivity of the medium.

The shape of the pressure dissipation versus time curve, however, depends on several factors
including the geometry of the cone tip, the location of the pore-pressure sensor on the cone
(Figure 1) and the dilatancy of the soil. Consequently, a wide range of dissipation curves can
be obtained as shown in Figure 2.

Figure 1. Piezocone with mid-face and shoulder pore-pressure sensors

time [s] time [s]

pore-water pressure [kPa]
pore-water pressure [kPa]

a) b)
time [s]
time [s]
pore-water pressure [kPa]
pore-water pressure [kPa]

c) d)
Figure 2. Different types of pore water pressure dissipation curves

2
pore-water pressure [MPa]

u0

time [s]

Figure 3. Difference between the measured pressure of the mid-face and shoulder
elements (after Shiyo et al., 1994)

The response also depends on whether the pore-pressure sensor is located on the mid-face of
the cone (u1), or on the shoulder behind the cone (u2). The excess pore-pressure measured at
the shoulder is typically less than that measured at the mid-face location (Figure 3). This
observation can be explained as follows:

 The soil is more compressed at the cone tip than at other locations
 Due to symmetry, soil below the cone tip is not subject to shear, which does occur
adjacent to the rods.
 As the distance from the cone tip increases, the dissipation of the pressure and
deformations induced by the cone tip increases.

CALCULATION OF THE HYDRAULIC CONDUCTIVITY

The problem is to understand how all these different results can lead to a hydraulic
conductivity k or a consolidation coefficient cv. As shown in Figure 3, the pressure can vary
significantly over a small distance in the beginning of the measurement. That is why the
pressures at the beginning of the curve should not be used for the calculation of the
coefficients. It is better to use the data obtained as the pore pressure nears final equilibrium,
whereby the pressures are the same for all measuring points around the cone.

Occasionally the measurements are stopped before final equilibrium is reached, especially in
the case of a low permeability soil where the dissipation time may exceed 24 hours, making a
complete dissipation test economically unattractive. Figure 2b and Figure 3 are examples of
unfinished dissipation tests. For these cases an approximation of the equilibrium pore
pressure u0 must be obtained by other tests or calculations (see Figure 3).

3
Pore water pressure dissipation consists of two phenomena, the groundwater flow and the
conservation or storage of water. The groundwater flow can be described by Darcy’s Law:

q  ki where: q  Q / A and: i  h / x (1)

(q = specific discharge, i = hydraulic gradient, Q = total discharge, A = cross section, h =

hydraulic head, x = flow distance).
The hydraulic head depends on the pressure:

h  p /  w  z so: h  p /  w (2)

(  w = unit weight of water, z = vertical coordinate of hydraulic head location)

Combining equations (1) and (2), we obtain the relation between pressure and specific
discharge:

k p
q (3)
 w x

The increase of the water volume in the incremental volume V is for a small time step:

I  q  V  t (4)

In the same area, storage of water can be obtained by the compressibility of water:

B  n  V    p (5)

The storage capacity depends on the change in pressure p, the compressibility of water ,
and the ratio of water in the soil, which is the porosity n (at saturation).
Considering the Law of Conservation of Mass and assuming soil solids are incompressible,
the sum of (4) and (5) must be zero:

I  B  0
(6)
  q  V  t  n  V    p  0

For one-dimensional groundwater flow ( q  q / x ) the incremental volume element V

can be defined as:

V  x  A (7)

By inserting equation (7) in (6) we find:

q
 x  A  t  n  x  A    p  0 (8)
x

By combining equations (3) and (8), we obtain the following diffusion equation:

4
 k 2 p p
 n 0 (9)
 w x 2
t

or:

dp 2 p k
 cd 2 where : cd  (10)
dt x n w

The dissipation constant cd is a diffusion coefficient which determines the rate of the excess
pore water dissipation.

For two-dimensional axisymmetric (radial) groundwater flow (r = x), the relation is identical,
except that the flow cross-section A depends on the distance r from the origin, the angle  (in
radians, see Figure 4) and the height b:
A    b  r so: A    b  r (   0 ) (11)
For a constant specific discharge q and a small time step t, the difference between the
discharge entering the area and leaving the area is q  A  t . This gives for axisymmetric
groundwater flow:

q
 r  A  t  n  r  A    p  q  A  t  0
r
(12)
dp   2 p 1 p  k
 cd  2   where: cd 
dt  r r r  n  w

For three-dimensional groundwater flow around the cone tip, the flow cross-section is:

A   r 2 so : A    2rr  (13)

This gives the following diffusion equation:

dp   2 p 2 p  k
 cd  2   where: cd  (14)
dt  r r r  n w

5
Figure 4. Two-dimensional and three-dimensional groundwater flow

Near the shoulder element at the shaft (u 2) the dissipation process will be approximately the
average of the 2-D and 3-D processes, as shown in Figure 4. This average is:

dp   2 p 1.5 p  k
 cd  2   where: cd  (15)
dt  r r r  n w

The solution to this differential equation consists of Bessel functions. Because of the
complexity of the analytical solution it is customary to evaluate the dissipation constant cd by
plotting the excess pore pressure versus the logarithm of time. If the relative pore pressure is
plotted for one specific distance from the origin (r = 2.5 cm as in Figure 5), for several
dissipation constants, then it becomes clear that the distribution in time depends linearly on
the dissipation constant.

0.25
relative pore pressure (u / umax ) [-]

cd = 210-6 m2/s
0.20

0.15

0.10
1cd
2cd
0.05
4cd

0.00
1 10 100 1000

t (s) 163 s 325 s 650 s

Figure 5. Relative pore pressure as a function of the dissipation constant (r = 2.5 cm)
6
This means that if the distance from the origin is known, theoretically speaking each clear
point of the curve can be used to calculate the dissipation constant, for example the top of the
curve or the tangent of the ascending or descending parts of the curve.
As the rate of pore pressure dissipation is a function of the radius from the cone r, it is found
that, especially at the start of the dissipation, the results depend strongly on the position of
the pore pressure elements, as shown in Figure 6.

1.00
relative pore pressure (u / umax ) [-]

0.90
0.80
0.70
0.60 r = 0.0 cm
0.50 r = 1.5 cm
0.40
r = 2.5 cm
0.30
r = 4.0 cm
0.20
0.10
0.00
1 10 100 1000

t (s)

Figure 6. Relative pore pressure as a function of the distance from the origin

As the distance from the origin increases, the pore pressure change decreases and the
maximum pressure change occurs at a progressively longer delay from the onset of the test.
To ensure consistent behavior, regardless of the distance from the origin, it is best to
calculate the dissipation constant from the tangent to the descending portion of the curves.
An example of the calculation procedure is now described with the aid of Figure 7. The
dissipation constant cd is related to the point of intersection of the diagonal tangential line
with the horizontal line of the pore water pressure at equilibrium u0. Please note that the
corresponding time of this intersection point, which is called t100%, does not actually
represent 100% dissipation, as the pore pressure approaches u0 asymptotically.

t [s]

t100% = 1514 s
pore pressure [kPa]

uo

Figure 7. Determination of the dissipation constant

7
As Figure 5 shows, the reciprocal value of this time t100% depends linearly on the dissipation
constant. Hence:

cd  X / t100% in which: X  0.0013 m2 (16)

In this case, the value of the constant X was derived from equation (15) numerically (shown
in Figure 5 and 6). This equation is based on a relatively thin cone in respect to the
dissipation area. With a calibration of the CPT cone the value of X can be determined more
precisely.

The diffusion process of a dissipation test looks very similar to a one-dimensional

compression test. However, the dissipation constant cd is different from the consolidation
coefficient cv:

k
cd 
n  w
(17)
k
cv 
 mv  n   w
Note that in soil mechanics text books it is commonly assumed that the pore space
compressibility term n is small compared to mv. The (mv+n) term represents the storativity
of the porous medium (Bear, 1979) .
The similarity is that both coefficients depend linearly on the permeability and the unit
weight of water. The greatest difference is that the one-dimensional compression test
depends mainly on the compressibility of the soil mv. The dissipation process does not,
because during this test the cone is stationary and the soil is not being compressed. The
dissipation process depends linearly on the porosity n, and the compressibility of the pore
water . Substituting equation 16 into 17 we obtain:

 X 
k   n  w
 t100% 
(18)
 X  n
cv   
 t100%   mv  n  

From comparison of the two relationships in equation 18, it is clear that the consolidation
coefficient can also be estimated if the compressibility of the soil mv is known. This can be
found for example with a one-dimensional compression test, which is unfortunately not an in
situ test. However, the compressibility of the soil can be determined in situ using the Menard
pressuremeter test or empirical CPT-correlations.

In the calculation of the permeability from the dissipation time, there can be some
inaccuracy. First, the in situ porosity is not known, although one can assume roughly for sand
n  0.4 , for clay n  0.6 , and for peat n  0.8 . This can lead to errors of 10% to 20% in k,
which is generally acceptable. Second, the in situ compressibility of the pore water is not

8
known. The compressibility  depends on the degree of saturation S and the pressure p
(Verruijt, 2005):

  S 0 
1  S  (19)
p

If there is no air (or methane for peat) in the soil then the compressibility of the pore water is
 0  0.5 109 m2 /N . If there is 1% air or methane in the water then the compressibility is
  1.0 107 m2 /N , for shallow samples close to atmospheric conditions: p = 100 kPa. This
means that a small amount of (non-dissolved) air or methane can change the compressibility
and the permeability of the soil by a factor of 200.

This is not a severe limitation because the amount of non-dissolved air and methane in
organic materials is primarily below 1%. Non-organic materials (pure clay and sand) have
almost no air in the pore water, which results in a compressibility approaching 0.
The validity of the approach is demonstrated in the following three examples from The
Netherlands that concern organic silty clays.

EXAMPLE 1: PANNERDEN

Figure 8. shows a dissipation curve obtained near Pannerden in The Netherlands for a new
train line. Borings show an organic silty clay layer at a depth of 18 m below mean sea level.
Although no actual measurement data is available, the amount of non-dissolved air and
methane in the silty organic clay is estimated to be low, about 0.2%, which results in a
compressibility of about   2.0 108 m2 /N . According to a one-dimensional compression
test (log-t method) the permeability of this layer is about k  5.9 108 cm/s (boring 28). A
nearby dissipation test showed a reduced hydrostatic pressure due to dilitant behavior,
resulting in a dissipation time of t100%  210 s (see Figure 8). Using equation (18) yields an
estimated hydraulic conductivity of:

X 0.0013
k n w   0.5  2 108  9.8 103
t100% 210
k  6.1108 cm/s

This value is comparable to the value of the one-dimensional compression test, even though
the amount of non-dissolved air and methane was estimated.

9
pore pressure u1 [kPa]

uo

t100% = 210 s
t [s]

Figure 8. Dissipation test in organic silty clay near Pannerden

EXAMPLE 2: RIJSWIJK

Figure 9. shows a dissipation curve obtained near Rijswijk. Both u 1 and u2 dissipation tests
were carried out on the same organic silty clay. Both tests resulted in a hydraulic
conductivity of:

X 0.0013
k n w   0.6  2 108  9.8 103
t100% 300
k  5.1108 cm/s

This value falls between the values from the one-dimensional compression test obtained
using the log-t method ( k  3.6 108 cm/s ) and the -t method ( k  6.7 108 cm/s ).

140
u1
and u2 [kPa]

120
U1 & U2u1[kPa].....

100
u2
80
pore pressure

60

40 uo
20

0
1 10 100 1000
t100%=300 s
t [s]
[s]

Figure 9. Dissipation test in organic silty clay near Rijswijk

EXAMPLE 3: ROTTERDAM

10
In Rotterdam in The Netherlands a u2 dissipation test was carried out on organic silty clay by
Kort (2002). Figure 10 is a plot of the results. The water content w was measured to be about
50%, which results in a porosity of n = 0.57. There are about 14 m of peat and organic clay,
producing methane, below the depth tested. During excavation of a building pit, methane was
even bubbling up from the ground. Therefore the amount of non-dissolved gas was estimated
at around 1%. Equation (19) gives an estimation of the pore water compressibility of:

1 S 1  0.99
  S 0   0.99  0.5 109 
p0 (100  13) 103
  8.85 108

Based on this assumed compressibility and based on the dissipation time of 1210 s, the
dissipation test results in a permeability of:

X 0.0013
k n w   0.57  8.85 108  9.8 103
t100% 1210
k  5.3 108 cm/s

This value agrees well with the values obtained from the one-dimensional compression test
using both the log-t and the -t method ( k  5.2 107 mm/s ), despite the rough estimation of
the amount of gas.

60

50
u2
2 [kPa]

40
U2 u[kPa]

30
pore pressure

20

10

0
1 10 100 1000 10000

tt [s]
[s] t100%=1210 s

Figure 10. Dissipation test in organic silty clay in Rotterdam

COMPARISON BETWEEN THE NEW INTERPRETATION METHOD AND THE

CLASSICAL INTERPRETATION METHODS

The new interpretation method differs fundamentally from classical methods, which are
presented by Lunne, et all. (1997). The new method is based on the concept of dissipation of
the excess pore pressure for a constant soil volume. The soil volume is constant because the
penetration has ceased at the beginning of the dissipation test. In this way, the dissipation
time depends on the permeability. The classical methods are all based on an assumed
changing volume during the test (which is incorrect) in such a way that the dissipation time

11
depends on the consolidation coefficient. The hydraulic conductivity and the consolidation
coefficient are fundamentally different parameters. Therefore it is impossible to compare
directly the results of the two different approaches.

Additionally, the classical methods can only directly interpret curves that continuously
descend in time. To evaluate other curve shapes requires a trial-and-error approach (Mayne,
2002). The new method however can be used to directly obtain the hydraulic conductivity for
all dissipation curve shapes.

CONCLUSIONS

The new interpretation method for the dissipation test, presented in this article, can be used to
determine the hydraulic conductivity for all dissipation curve shapes, which is not the case
for the classical interpretation methods. Another point is that the new method results in a
hydraulic conductivity instead of a consolidation coefficient, because during the cone
penetration test the cone is stationary and the soil is not being compressed as during
consolidation.

The values of the hydraulic conductivity obtained with this new method correspond well with
values obtained from one-dimensional compression tests, for the limited number of tests on
organic soils for which data was available. However, caution should be exercised when using
this method for non-organic soils.

REFERENCES

Bear, J., 1979, Hydraulics of groundwater, McGraw-Hill, New-York

Kort, D.A., 2002, Rotterdam sheet pile wall field test, CUR Publication 207, Gouda: CUR,
ISBN 90 376 0192 8
Lunne, T., Robertson, P.K. & Powell, J.J.M., 1997, Cone penetration testing in geotechnical
practice, Blackie Academic and Professional, ISBN 0-751-40393-8
Mayne, P.W., 2002, Flow properties from piezocone dissipation tests.
www.ce.gatech.edu/~geosys
Shiyo Chen, B. & Mayne, P. W., 1994, Profiling the overconsolidation ratio of clays by
piezocone tests, Rep. no. GIT-CEEGEO-94-1, National Science Foundation, August
1994, Arlington Virginia
Van Baars, S. & Van de Graaf, H.C., 2005, L’utilisation du piézocône pour la determination
de la perméabilité des sols compressibles, International Symposium Geoline, Lyon,
France, ISBN 2-7159-2982-X, May 2005
Van De Graaf, H.C., 2002, Le piézocône pour l'interprétation de paramètres de sol: faut-t-il
mesurer u1 ou u2?. PARAM 2002, Magnan (ed.), presses de l'ENPC/LCPC, Paris
Verruijt, A., 2005, Grondmechanica, VSSD, Delft, The Netherlands, ISBN 90-71301-45-1

12