Chapter Five

5. Permeability and Seepage

Contents of chapter 5
5.1 Introduction .................................................................................................................................................65
5.2 Hydraulic heads: ..........................................................................................................................................66
5.3 Seepage velocity vs ......................................................................................................................................67
5.4 Indirect methods to determine k: ...............................................................................................................68
5.5 Determination of coefficient of permeability k in the laboratory ...............................................................69
5.5.1 Constant head permeability test ......................................................................................................69
5.5.2 Falling head permeability test ..........................................................................................................69
5.6 Determination of coefficient of permeability k in the field.........................................................................69
5.7 Coefficient of permeability (hydraulic conductivity) k in layered soil .........................................................70
5.7.1 Flow parallel to soil layers ....................................................................................................................70
5.7.2 Flow normal to soil layers .....................................................................................................................70
5.8 Flow net .......................................................................................................................................................71
5.8.1 One-Dimensional Flow Net...................................................................................................................71
5.8.2 Flow Net for Two-Dimensional Problems with Isotropic Soils .............................................................72
5.9 Design of filters ............................................................................................................................................76
5.10 Applications of filters .................................................................................................................................78
5.10.1 Clay core .............................................................................................................................................78
5.10.2 Retaining wall .....................................................................................................................................78
5.11 Additional Problems ..................................................................................................................................79

5.1 Introduction
Permeability is the property of soil which permits flow of water from points of high energy to points of
low energy, due to existence of interconnected voids. It can be measured by coefficient of permeability
k (cm/s or m/s).
It is possible to measure k in the laboratory (by constant head method and falling head method, see
laboratory hand-outs) and in the field (well pumping test, see Craig’s Soil Mechanics)
The importance of permeability
1- Settlement prediction (preloading).
2- Seepage through and beneath earth structures such as earth dams and retaining walls.
3- In designing of filters in which protect hydraulic structure from piping.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 65

 4- Discharge of wells.
5- To determine the amount of water that return into shallow and deep excavations during
construction of a project.

5.2 Hydraulic heads:

The total head (ht) at a point in a submerged soil mass (see Figure 5.1) is expressed by Bernoulli
Equation as shown below:

(Total head) = (pore water head)] + (elevation head) + (velocity head)

The velocity of water in the soil is very small, therefore it is possible to ignore the last part of the
above equation resulting in Equation 5.1:

According to the position of datum, the values of and could be positive or negative.

Figure 5.1 Hydraulic heads (Water flow through a pipe)

Heads hz, hp and ht at various points in Figure 5.1

hz:elevation head is a distance from a point to datum

(+) if the point up the datum, (-) if the point down the datum
hp: pressure head is a distance from a point to Water Level
(+) if the point down the W.L (-) if the point up the W. L
ht: total head ht may be + or -.

Head loss is an energy loss. When water flows in soils, it must flow through many small passages in

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 66

void sections of soils, as illustrated in Figure 5.2. This creates frictional resistance on the surfaces of
particles. Flow energy is transmitted to frictional resistance on particle surfaces and then may be lost
in heat generation, although it may not be easy to measure the temperature rise due to this energy
transfer.

Figure 5.2 Frictional energy loss around particles due to water flow.

Darcy’s law: the velocity of flow of water through a soil is proportional to the hydraulic gradient (i).

where q is the rate of flow of water (i.e. discharge of water in, for example, volume/time),
k is coefficient of permeability of a soil,
i is hydraulic gradient = h/L,
h is loss in total head (ht) (difference in total head (i.e. hA – hB , see Figure 5.3)) ,
L= length between points A and B of the soil mass (see Figure 5.3).

Figure 5.3 Pressure, elevation and total heads for flow of water through a soil

5.3 Seepage velocity vs

The seepage velocity vs is the average velocity of water flowing through pores. It can be estimated
by dividing the flow rate q by the average area of voids Av on a cross section normal to the
direction of flow (see Figure 5.4):

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 67

Multiplying Eq. 5.4 by L resulting in:

Figure 5.4 derivation of Equation 3.5

Typical permeability ranges

Soils exhibit a very wide range of permeability and while particle size may vary by about 3-4 orders of
magnitude, permeability may vary by about 10 orders of magnitude.

Table 5.1 typical values of k (hydraulic conductivity) for common soil types
Soil type k (cm/s) Description Drainage
Clean gravel (GW, GP) > 1.0 High Very good
Clean sands, clean sand and gravel mixtures (SW, SP) 1.0 to 10-3 Medium Good
Fine sand, silts, mixtures comprising sands, silts and 10-3 to 10-5 Low Poor
clays (SM-SC)
Weathered and fissured clays
Silt, slity clay (ML, MH) 10-5 to 10-7 Very low Poor
Homogeneous clays (CL, CH) < 10-7 Practically Very poor
impervious

5.4 Indirect methods to determine k:

Permeability is often estimated from correlations with particle size. For example

where D10 is effective diameter of soil particle (mm) for finer 10%. In Equation 5.6, k has to be in
(cm/s) where C = 1.0 - 1.5. This expression was first proposed by Hazen in 1893. It is satisfactory for
sandy soils but is less reliable for well graded soils and soils with a large fines fraction.

Another similar empirical type of correlation is given by Chapuis (2004):

(5.7)

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 68

where
k: coefficient of permeability (cm/s)
D10: particle size for which 10% of the soil is finer (or effective grain size) (mm)
e: void ratio of soil

5.5 Determination of coefficient of permeability k in the laboratory

There are two main permeability tests in the laboratory to determine the value of k as follow:

5.5.1 Constant head permeability test

5.5.2 Falling head permeability test

Details of these two tests are available in the laboratory hand outs.

5.6 Determination of coefficient of permeability k in the field

Pumping test

Pumping water at a constant flow rate from a well and measuring the decrease in groundwater level at
observation wells (see Figure 5.5) is a common method of determining the coefficient of permeability
k in the field. The following assumptions are made to derive a simple equation for k (for more details
about this method, read Das, 2014).

Why pump test is only practical for coarse-grained soils?

Pumping tests lower the groundwater, which then causes stress changes in the soil. Since the
groundwater is not lowered uniformly, as shown by the drawdown curve in Figure 5.5, the stress
changes in the soil will not be even. Consequently, pumping tests near existing structures can cause
these structures to settle unevenly. You should consider the possibility of differential settlement on
existing structures when you plan a pumping test.

Figure 5.5 Pumping test layout

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 69

5.7 Coefficient of permeability (hydraulic conductivity) k in layered soil
5.7.1 Flow parallel to soil layers
In a layered soil, when the flow is parallel to the soil layers (see Figure 5.6), the hydraulic gradient is
the same at all points. The flow through the soil mass as a whole is equal to the sum of the flow
through each of the layers. If it is considered a unit width (in the y direction) of flow and use Equation
(5.3), we obtain

where Ho is the total thickness of the soil mass, kx(eq) is the equivalent permeability in the horizontal
(x) direction, z1 to zn are the thicknesses of the first to the nth layers, and kx1 to kxn are the horizontal
hydraulic conductivities of the first to the nth layer. Solving Equation (5.14) for kx(eq), the result is

Figure 5.6 Flow through layered soil

5.7.2 Flow normal to soil layers

For flow normal to the soil layers, the head loss in the soil mass is the sum of the head losses in each
layer:

where ΔH is the total head loss, and Δh1 to Δhn are the head losses in each of the n layers. The velocity
in each layer is the same.
From Darcy’s law, we obtain

where kz(eq) is the equivalent hydraulic conductivity in the vertical (z) direction and kz1 to kzn are the
vertical hydraulic conductivities of the first to the nth layer. Solving Equations (5.16) and (5.17) leads
to

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 70

Values of kz(eq) are generally less than kx(eq).

The equivalent hydraulic conductivity for flow parallel and normal to soil layers is

5.8 Flow net

Flow net is a convenient graphical tool to compute hydraulic properties such as the amount of water
flow, water pressure on flow boundaries, etc. for two-dimensional flow problems with complex
geometries. A simple one-dimensional model is first introduced here to help understand the principle
of the flow net.

5.8.1 One-Dimensional Flow Net

Figure 5.7 shows a water flow through soil in a vertical cylinder with length L and cross-sectional area
A. The flow is downward due to the total head h. The cylinder is equally divided into three (Nf) flow
channels, which are parallel to the direction of the flow and through which equal amounts of water
flow. These imaginary boundaries of the flow channels are called flow lines, and water flow never
crosses the flow lines. The specimen length L is equally divided by four (Nd) as seen with dotted
horizontal lines. Since the head loss occurs linearly through the specimen depth in this case, the head
losses between adjacent horizontal lines are all h/4 (h/Nd). The total heads on individual horizontal
lines are constant since they have the same elevation heads and the same pressure heads; thus, these
lines are called equipotential lines. Flow lines and equipotential lines make a net (mesh) geometry,
which is called flow net.

Equipotential lines and the flow lines intersect each other at 90°.

h = difference in total head , Δh = drop in head, Nd = number of drops, Nf = number of channels.

For one channel

It is assumed that a=b (where a (A/Nf) is width of one flow line and b (L/Nd)), then Equation 5.21
becomes

The total amount of water flow q through the entire cross section of the soil is

Equation 5.23 is equally applied for one-dimensional flow and two-dimensional flow.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 71

 Δh=h/Nd

Δh=h/Nd
h

Δh=h/Nd

Δh=h/Nd

Figure 5.7 One-dimensional flow net principles

5.8.2 Flow Net for Two-Dimensional Problems with Isotropic Soils

The following steps are recommended for flow-net constructions:
1. Draw the geometry of the structure correctly on the paper. It implies that the horizontal and vertical
scales should be the same. Otherwise, the square shape requirement (a = b) cannot be met.

2. Select a proper Nf value. Normally, Nf of three or four is adequate for the first trial.

3. Identify the boundary flow lines and boundary equipotential lines in the drawing. In an example in
Figure 5.8, the upstream ground surface and downstream ground surface are the initial and the final
equipotential lines, respectively. The front and back sides of the sheet pile and the surface of the
impervious layer are the boundary flow lines.

4. First, draw trial flow lines with selected Nf for the entire earth structure (Figure 5.8). It should be
noted that there are equal amounts of water flow through all flow channels.

5. Starting from the upstream site, draw the first equipotential line to have all net openings squares or
near-squares with 90° intersections. In two-dimensional problems, however, it is impossible to have all
net openings be exactly squares. Figure 5.9 shows examples of acceptable near-squares in the flow net.
At the corners of the structural boundaries cases, margining of two equipotential lines is allowed as
seen in the triangle case. Make the necessary corrections on originally drawn flow lines to satisfy near-
square and 90° intersection requirements as closely as possible.

6. Draw the second and third equipotential lines, so on, until it reaches the downstream exit as seen in
Figure 5.9.

7. At the downstream exit point, it may not get to full squares with the last equipotential line. In such a
case, draw an imaginary equipotential line beyond the last physical equipotential line to have full near-
square sections.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 72

In Figure 5.9, Nf =3 is chosen and Nd = 5.4 is obtained, and thus the shape factor Nf/Nd of this geometry
is 0.556. If a larger number of Nf is chosen, a proportionally larger Nd value will be obtained when the
square rule is carefully followed, and thus the similar shape factor should be obtained.

Figure 5.8 Two-dimensional flow net drawing

Total heads at any points on the same equipotential line should be the same (see Figure 5.9), if
standpipes are placed all along the first equipotential line, the water levels in the pipes are the same.

Figure 5.9 Complete construction of flow net for Figure 5.8 and total head pressure in flow net
In Figure 5.10, several examples of flow net under concrete dams and through earth dams are shown.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 73

Figure 5.10 Examples of flow net for dams

Example 5.1:
Figure below shows the flow net under a sheet pile wall. Considering 1 meter length of the wall:
1- calculate the rate of flow (in m3/year),
2- calculate the value of pore water pressure at point A,
3- Determine the value of z at point B.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 74

Solution:

Example 5.2:
Figure below shows a flow net of a sheet pile. Based on the available information on the Figure,
determine:
1) the water head at upstream (h1) and downstream (h2),

2) the rate of flow of water in m3/day.

Solution:

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 75

5.9 Design of filters
When seepage water flows from a soil with relatively fine grains into a coarser material (see, for
example, Figure 5.11), there is a danger that the fine soil particles may wash away into the coarse
material. Over a period of time, this process may clog the void spaces in the coarser material. Such a
situation can be prevented by the use of a filter or protective filter between the two soils.

It can be shows that, if three perfect spheres have diameters greater than 6.5 times the diameter of a
smaller sphere, the small sphere can move through the void spaces of the larger ones (Figure 5.12a).
Generally speaking, in a given soil, the sizes of the grains vary over a wide range. If the pore spaces in
a filter are small enough to hold D85 of the soil to be protected, then the finer soil particles also will
be protected (Figure 5.12b). This means that the effective diameter of the pore spaces in the filter
should be less than D85 of the soil to be protected. The effective pore diameter is about of the filter.

For the proper selection of the filter material, two conditions should satisfy.

1. The size of the voids in the filter material should be small enough to hold the larger particles of the
protected material in place.

2. The filter material should have a high permeability to prevent build up of large seepage forces and
hydrostatic pressures in the filters.

Based on the experimental investigation of protective filters, Terzaghi and Peck (1948) provided the
following criteria to satisfy the above conditions:

to satisfy condition 1 to satisfy condition 2

Figure 5.11 Toe filter

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 76

Figure 5.12 (a) Large spheres with diameters of 6.5 times the diameter of the small sphere;
(b) boundary between a filter and the soil to be protected

The proper use of Equation 5.13 to determine the grain-size distribution of soils used as filters is
shown in Figure 5.13. Consider the soil used for the construction of the earth dam shown in Figure 5.8.
Let the grain-size distribution of this soil be given by curve a in Figure 5.13. We can now determine
5D85(S) and 5D15(S) and plot them as shown in Figure 5.13. The acceptable grain-size distribution of the
filter material will have to lie in the shaded zone. Note that the shape of curves b and c are
approximately the same as curve a.

Figure 5.13 Determination of grain-size distribution of filter using Equation 3.14

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 77

5.10 Applications of filters
5.10.1 Clay core
There are slight differences for application to new construction and modification to existing dams. For
new construction, the chimney would be placed near the center line of the dam for central core
designs, whereas the addition of a chimney to an existing dam would require removal of a large
portion of the existing embankment to obtain this location. The central location is desirable to
maximize the confining stress on the chimney as well as to minimize hydrostatic pressure in the
downstream shell. There-fore, modifications to existing dams will typically locate the chimney further
downstream than what would be used for new construction. When chimneys are located downstream,
sufficient overburden must be provided for stability against a potential full reservoir head. In a similar
manner, the blanket added during an existing dam modification would be shorter since the chimney it
connects to is further downstream. Examples of the two arrangements are shown in Figures 5.14 and
5.15.

Figure 5.14 Simple cross section showing a chimney used in a new dam.

Figure 5.15 Simple cross section showing a chimney added to an existing dam.

5.10.2 Retaining wall

Figure 5.16 shows a drainage zone being constructed next to a battered concrete wall that is part of a
spillway chute. In this application, perforated pipes in a gravel backfill are used to provide drainage
behind the wall. Since the gravel drain is not filter compatible with the foundation, an intervening sand
layer is used to provide filter protection. This is a two-stage system used to protect the foundation
while providing drainage for the wall.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 78

 Figure 5.16 drainage zone being constructed next to a battered concrete wall that is part of a spillway
chute

5.11 Additional Problems

P5.1:
The particle size distribution (PSD) curves for soils A and B are shown in the figure below, estimate
the value of k of the soils using Hazen’s and Chapuis’s Equations (Equation 5.6 and 5.7), use C=1.2
and void ratio e equals to 0.95 and 0.6, respectively.
Answers:
Use Hazen’s Equation : Soil A k = 0.32 cm/sec, Soil B k = 3.48x10-6 cm/sec
Use Chapuis Equation :
Soil A k = 0.465 cm/s, Soil B k = 2.38x10-5 cm/sec

Soil B

Soil A

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 79

P5.2:
Soil’s gradation data are given in the following table. The void ratio, e, was 0.550 at the site, and its
particle shape was found to be round.

Estimate the approximate values of coefficient of permeability by

(a) Hazen’s formula
(b) Chapuis’s formula
Answers: (a) k = 0.04 cm/s, (b) k = 0.0346 cm/s,

P5.3:
Figure below shows water flow though the soil specimen in a cylinder. The specimen’s k value is
3.4×10−4 cm/s.
(a) Calculate pressure heads hp at Points A, B, C, and D and draw the levels of water height in
standpipes.
(b) Compute the amount of water flow q through the specimen.

Answers: (a)

(b) q = 4.27 × 10−3 cm3/s

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 80

P5.4:
A canal is cut into a soil with a stratigraphy shown in Figure below. Assuming flow takes place
laterally and vertically through the sides of the canal and vertically below the canal:
1- Determine the equivalent hydraulic conductivity in the horizontal and vertical directions. The
vertical and horizontal hydraulic conductivities for each layer are assumed to be the same.
2- Calculate the ratio of the equivalent horizontal hydraulic conductivity to the equivalent vertical
hydraulic conductivity for flow through the sides of the canal.
3- The equivalent hydraulic conductivity.
4- Find equivalent vertical hydraulic conductivity below the bottom of the canal.

0.023

Answers:
1: kx(eq) = 3x10-6cm/s and kz(eq) = 0.61x10-6cm/s , 2: kx(eq) / kz(eq) = 4.9. ,
3: keq = 1.355x10-6cm/s , 4: kz(eq) = 7.2x10-6cm/s

P5.4:
For the same dam and the same soil conditions as in Problem 5.3, a 4-meter-long vertical sheet pile is
added at the left end of the dam base:
(a) Draw the flow net by using Nf = 3 and Nd=6.5.
(b) Compute the flow rate q under the dam.
(c) Compute the water pressures at the heel (A) and toe (B) sections of
the base of the dam based on the drawn flow net.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 81

Answer: (b) q = 0.401 cm3/s/cm

Example 5.5:
A sand sample of 35cm2 cross sectional area and 20cm long was tested in a constant head
permeameter. Under a head of 60cm, the discharge was 120ml in 6min. The dry weight of sand used
for the test was 1120g, and Gs=2.68. Determine (a) the hydraulic conductivity in cm/s, (b) the
discharge velocity, and (c) the seepage velocity.

Answers: a) 3.174 x 10-3 cm/s, b) 9.52 x 10-3 cm/s, c) 2.36 x 10-2 cm/s
Example 5.6:
A cofferdam of two lines of sheet piles penetrated into a seabed by 6m ,the soil inside the sheet
piles is excavated by 2m down the bed of average k=2x10-5m/s and sat=20 kN/m3, the water level is
fixed with excavation level by pumping. The requirements are to find the rate of flow of water and u
at point P.
Answers: q =4.666 m3/ day, u = 47.53 kN/m2

5.5m

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 82

P5.7:
Two soil of same cross sectional and length , a constant head test is performed on the soils as
shown in fig below if the k of soil 2 is 4 time k of soil 1 ,find the
 the total heads (ht) at points
A,B,C,
 the hydraulic gradient (i) for both
soils ,
 the flow rate q in cm3/sec .
Answer: ht(B) = 98 cm

P5.8:
In a falling head permeameter, the sample used is 20cm long having a cross-sectional area of 24cm2.
Calculate the time required for a drop of head from 25 to 12cm if the cross-sectional area of the stand
pipe is 2cm2. The sample of soil is made of three layers. The thickness of the first layer from the top is
8cm and has a value of k1 = 2 x 10-4 cm/s, the second layer of thickness 8 cm has k2 = 5 x 10-4 cm/s
and the bottom layer of thickness 4cm has k3 = 7 x 10-4 cm/s. Assume that the flow is taking place
perpendicular to the layers.

Answer: t = 3771 s = 62.9 minutes

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 83

P5.9:
A constant head test was carried out on two soils shown in the figure below.
Determine:
1- the total heads (ht) at points A, B and C
2- the hydraulic gradient (i) for both soils,
3- the rate of flow (q) of water through the soils,
4- seepage (vs) velocity for each soil.

W.L.

Δh=50cm
(between points
A and C)

W.L.

P5.10:
The figure below shows a completed flow net around a sheet pile. Points a, b, c, d, e, f, g, h, i, j and k
are labeled along the sheet pile. Determination of water pressures at those points is required.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 84

 h

P5.11:

A flow net under a concrete dam is drawn in the figure below.

(a) Calculate and plot the water pressure distribution along the base of the dam.
(b) Compute the resultant uplift force against the base of the dam.
(c) Calculate the point of application of the resultant uplift force.

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 85

 Figure I

Answers:

(a).

(b)
Resultant uplift force P = 926.2 kN/m (per dam length).
(c) Point of application of P = = 6.99 m from Point “a” .

Ahmed M. Hasan, PhD, Geotechnical Specialist, College of Engineering – Salahaddin University 86