BDU, BiT, FCWRE

CHAPTER FOUR

TWO-DIMENSIONAL FLOW OF WATER THROUGH SOILS

4.1 Introduction

Many catastrophic failures in geotechnical engineering result from instability of soil masses due to
groundwater flow. Lives are lost, infrastructures are damaged or destroyed, and major economic
losses occur. In this chapter, you will study the basic principles of two-dimensional flow of water
through soils. The topics covered here will help you to avoid pitfalls in the analyses and design of
geotechnical systems where groundwater flow can lead to instability. The emphasis in this chapter is
on gaining an understanding of the forces that provoke failures resulting from groundwater flow. You
will learn methods to calculate flow, pore water pressure distribution, uplift forces, and seepage
stresses.

4.2 Two-dimensional Flow of Water Through Porous Media

The flow of water through soils is described by Laplace’s equation. The popular form of Laplace’s
equation for two-dimensional flow of water through soils is

∂2H ∂2H
kx 2
+ kz =0 (4.1)
∂x ∂2z
where H is the total head and kx and kz are the coefficients of permeability in the X and Z directions.
Laplace’s equation expresses the condition that the changes of hydraulic gradient in one direction
are balanced by the changes in the other directions.
The assumptions in Laplace’s equation are:
• Darcy’s law is valid.
• The soil is homogeneous and saturated.
• The soil and water are incompressible.
• No volume change occurs.
If the soil were an isotropic material then kx = kz and Laplace’s equation becomes

∂2H ∂2H
+ =0 (4.2)
∂x 2 ∂2z
In this chapter, we are going to emphasis in an approximate (graphical) solution technique for
Laplace’s equation called the flow net sketching. The flow net sketching technique is simple and
flexible and conveys a picture of the flow regime. It is the method of choice among geotechnical
engineers. But before we delve into this solution technique, we will establish some key conditions
that are needed to understand two-dimensional flow.
The solution of Eq. (4.1) depends only on the values of the total head within the flow field in the xz

plane. Let us introduce a velocity potential ( ξ ), which describes the variation of total head in a soil
mass as
ξ = kH (4.3)
where k is a generic coefficient of permeability. The velocities of flow in the X and Z directions are

Soil Mechanics I: Lecture Notes 1

BDU, FOE, Department of Civil Engineering
∂H ∂ξ
vx = k x = (4.4)
∂x ∂x

∂H ∂ξ
vz = k z = (4.5)
∂x ∂z
The inference from Eqs. (4.4) and (4.5) is that the velocity of flow (v) is normal to lines of constant
total head (also called constant piezometric head or equipotential lines) as illustrated in Fig. 4.1.
The direction of v is in the direction of decreasing total head. The head difference between two
equipotential lines is called a potential drop or head loss.
If lines are drawn that are tangent to the velocity of flow at every point in the flow field in the XZ
plane, we will get a series of lines that are normal to the equipotential lines. These tangential lines
are called streamlines or flow lines (Fig. 4.1). A flow line represents the flow path that a particle
of water is expected to take in steady state flow.

Figure 4.1: Illustration of flow terms.

Since flow lines are normal to equipotential lines, there can be no flow across flow lines. The rate of
flow between any two flow lines is constant. The area between two flow lines is called a flow channel
(fig. 4.1). Therefore, the rate of flow is constant in a flow channel.
4.3 Flow Net Sketching

4.3.1 Criteria for Sketching Flow Nets

A flow net is a graphical representation of a flow field that satisfies Laplace’s equation and comprises
a family of flow lines and equipotential lines.
A flow net must meet the following criteria:
1. The boundary conditions must be satisfied.
2. Flow lines must intersect equipotential lines at right angles.
3. The area between flow lines and equipotential lines must be curvilinear squares. A curvilinear
square has the property that an inscribed circle can be drawn to touch each side of the square
and continuous bisection results, in the limit, in a point.
4. The quantity of flow through each flow channel is constant.
5. The head loss between each consecutive equipotential line is constant.
6. A flow line cannot intersect another flow line.
7. An equipotential line cannot intersect another equipotential line.

Soil Mechanics I: Lecture Notes 2

BDU, FOE, Department of Civil Engineering
An infinite number of flow lines and equipotential lines can be drawn to satisfy Laplace’s
equation. However, only a few are required to obtain an accurate solution. The procedure for
constructing a flow net is described next. A few examples of flow nets are shown in Figs.4.2 to 4.3.

Figure 4.2: Flow net for a sheet pile.

4.3.2 Procedure for Sketching Flow Nets

1 Draw the structure and soil mass to a suitable scale.

2 Identify impermeable and permeable boundaries. The soil-impermeable boundary interfaces are
flow lines because water can flow along these interfaces. The soil-permeable boundary interfaces
are equipotential lines because the total head is constant along these interfaces.
3 Sketch a series of flow lines (four or five) and then sketch an appropriate number of equipotential
lines such that the area between a pair of flow lines and a pair of equipotential lines (cell) is
approximately a curvilinear square. You would have to adjust the flow lines and equipotential lines
to make curvilinear squares. You should check that the average width and the average length of
a cell are approximately equal. You should also sketch the entire flow net before making
adjustments.

Figure 4.3: Flow net under a dam with a cutoff curtain (sheet pile) on the upstream end.

Soil Mechanics I: Lecture Notes 3

BDU, FOE, Department of Civil Engineering

Figure 4.4: Flow net in the backfill of a retaining wall with a vertical drainage blanket.

4.4 Interpretation of Flow Nets

4.4.1 Flow Rate

The head loss ( ∆h) between each consecutive pair of equipotential lines is:

∆H
∆h = (4.6)
Nd

Where ∆H is the total head loss across the flow domain and Nd is the number of equipotential drops
(number of equipotential lines minus one). From Darcy’s law, the flow through each flow channel for
an isotropic soil is
∆h B ∆H B
∆q = Aki = (B ×1)k = k∆h = k (4.7)
L L Nd L
where B and L are sides of the curvilinear square. By construction B/L = 1, and therefore the total
flow is
Nf
q = N f ∆q = k∆H = k∆hN f (4.8)
Nd

where Nf is the number of flow channels (number of flow lines minus one). The ratio Nf/Nd is called
the shape factor. Both Nf and Nd can be fractional. In the case of anisotropic soils (different
permeabilities in X and Z directions), the quantity of flow is
Nf
q = ∆H kxkz (4.9)
Nd

4.4.2 Hydraulic Gradient

You can find the hydraulic gradient over each square by dividing the head loss by the length, L, of the
cell, that is

∆h
i= (4.10)
L
You should notice that L is not constant. Therefore the hydraulic gradient is not constant. The
maximum hydraulic gradient occurs where L is a minimum, that is,

Soil Mechanics I: Lecture Notes 4

BDU, FOE, Department of Civil Engineering
∆h
i max = (4.11)
Lmin

Where Lmin is the minimum length of the cells within the flow domain. Usually, Lmin occurs at exit
points or around corners.

4.4.3 Static Liquefaction

Let us consider an element of soil at a depth z near the downstream end of a sheet pile wall
structure subjected to seepage forces with upward flow. The vertical effective stress is:

σ z' = γ ' z − iγ w z (4.12)

If the effective stress becomes zero, the soil loses its strength and behaves like a viscous fluid. The
soil state at which the effective stress is zero is called static liquefaction. Various other names such
as boiling, quicksand, piping, and heaving are used to describe specific events connected to the
static liquefaction state. A structure founded on a soil that statically liquefies will collapse.
Liquefaction can also be produced by dynamic events such as earthquake.

4.4.4 Critical Hydraulic Gradient

We can determine the hydraulic gradient that brings a soil mass (essentially, coarse grained soils) to
static liquefaction. Solving for i in Eq. (4.12) when σ z' = 0 , we get
γ '  Gs − 1  γ w G s − 1
i = i cr = =  = (4.13)
γ w  1+ e  γ w 1+ e
where icr is the critical hydraulic gradient, Gs is specific gravity, and e is the void ratio. Since Gs is
constant the critical hydraulic gradient is solely a function of the void ratio of the soil. In designing
structures that are subjected to steady state seepage, it is absolutely essential to ensure that the
critical hydraulic gradient cannot develop.

4.4.5 Pore Water Pressure Distribution

The pore water pressure at any point j is calculated as follows:

1. Select a datum (for example, choose the downstream water level as the datum.)
2. Determine the total head at j: H j = ∆H − ( N d ) j ∆h, where ( N d ) j is the number of equipotential
drops at point j.
3. Subtract the elevation head at point j from the total head Hj to get the pressure head. For point
j below the datum (recall the datum is assumed to be the downstream water level), let the
elevation head hz is –z. The pressure head is then
(h p ) j = ∆H − ( N d ) j ∆h − h z (4.14)
4. The pore water pressure is
u j = (h p ) j γ w (4.15)

4.4.6 Uplift Forces

Lateral and uplift forces due to groundwater flow can adversely affect the stability of structures such
as dams and weirs. The uplift force per unit length (length is normal to the XZ plane) is found by

Soil Mechanics I: Lecture Notes 5

BDU, FOE, Department of Civil Engineering
calculating the pore water pressure at discrete points along the base (in the X direction) and then
finding the area under the pore water pressure distribution diagram, that is,
n
Pw = ∑u
i =1
j ∆x j (4.16)

where Pw is the uplift force per unit length, uj is the average pore water pressure over an interval ∆x j ,
and n is the number of intervals. It is convenient to use Simpson’s rule to calculate Pw:
 
∆x  
n n
Pw =
3  ∑ ∑
u1 + u n + 2 u i + 4 u i 

(4.17)
 i =3 i =4 
 odd even 
EXAMPLE 4.1

A dam, shown in Fig.E4.1a, retains 10 m of water. A sheet pile wall (cutoff curtain) on the upstream
side, which is used to reduce seepage under the dam, penetrates 7 m into a 20.3 m thick silty sand
stratum. Below the silty sand is a thick deposit of clay. The average coefficient of permeability of the
silty sand is 2.0 ×10 -4
cm/s. Assume that the silty sand is homogeneous and isotropic.
(a) Draw the flow net under the dam.
(b) Calculate the flow, q.
(c) Calculate and draw the pore water pressure distribution at the base of the dam.

(d) Determine the uplift force.

(e) Determine and draw the pore water pressure distribution on the upstream and downstream
faces of the sheet pile wall.
(f) Determine the resultant lateral force on the sheet pile wall due to the pore water.
(g) Determine the maximum hydraulic gradient.
(h) Will piping occur if the void ratio of the silty sand is 0.8.
(i) What is the effect of reducing the depth of penetration of the sheet pile wall?

Strategy: Follow the procedures described in section 4.2 to draw the flow net and calculate the
required parameters.

FigureE4.1a

Soil Mechanics I: Lecture Notes 6

BDU, FOE, Department of Civil Engineering
4.5 Flow Through Earth Dams
Flow through earth dams is an important design consideration. We need to ensure that the pore
water pressure at the downstream end of the dam will not lead to instability and the exit hydraulic
gradient does not lead to piping. The major exercise is to find the top flow line called the phreatic
surface (Fig.4.2). The pressure head on the phreatic surface is zero. Casagrande (1937) showed
that the phreatic surface can be approximated by a parabola with corrections at the points of
entry and exit. The focus of the parabola is at the toe of the dam, point F (Fig.4.2).

Figure 4.5: Phreatic surface within an earth dam.

The procedure to draw a phreatic surface within an earth dam, with reference to Fig. 5.5, is
as follows.
1. Draw the structure to scale.
2. Locate a point A at the intersection of a vertical line from the bottom of the upstream face and the
water surface, and a point B where the waterline intersects the upstream face.
3. Locate point C, such that BC=0.3AB.
4. Project a vertical line from C to intersect the base of the dam at D.
5. Locate the focus of the basic parabola. The focus is located at the toe of the dam.

6. Calculate the focal distance, f = ( b 2 + H 2 − b) / 2 , where b is the distance FD and H is the height

of water at the upstream face.

7. Construct the basic parabola from z = 2 f ( f + x) .

8. Sketch in a transition section BE.

9. Calculate the length of the discharge face, a, using
b
a= − cos β b 2 − H 2 cot 2 β ; β ≤ 30 0 .
cos β

For β > 30 0 , use Fig.5.7 and (a) measure the distance TF, where T is the intersection of the basic
parabola with the downstream face; (b) for the known angle β , read the corresponding factor
∆a / L from the chart; and (c) find the distance a = TF (1 − ∆a / L) .
10. Measure the distance a from the toe of the dam along the downstream face to point G.
11. Sketch in a transition section, GK.

Soil Mechanics I: Lecture Notes 7

BDU, FOE, Department of Civil Engineering
12. Calculate the flow using q = ak sin β tan β , where k is the coefficient of permeability. If the
downstream slope has a horizontal drainage blanket as shown in Fig. 5.3, the flow is calculated
using q = 2 fk .

Figure 4.6: A horizontal drainage blanket at the toe of an earth dam.

∆a
L

β (degrees)

Figure 4.7: Correction factor for downstream curve.

5.6 Design of Filters
Hydraulic structures such as weirs & dams built on pervious foundations sometimes fail by
formation of a pipe-shaped channel in its foundation or in the earth dam itself, known as piping
failures. The failure occurs when water flowing through the foundation has a very high hydraulic
gradient and carries fine soil particles with it.

Retaining wall

Earth Dam

Perforated pipe
Piping thrrough dam

piping in the foundation

Soil Mechanics I: Lecture Notes 8

BDU, FOE, Department of Civil Engineering
Seepage flows under dam & spillway aprons usually develops greater uplift pressure which causes
the failure of the dam (Collapse) if not properly designed or the uplift is reduced.

Dam

uplift presuure

Upsrteam curtain

In water draining structures (perforated pipes) and construction of earth dams, it is often
necessary to provide a layer (or layers) of filter material to prevent fine particles being carried in to
pipes or into the void space of courses materials and also to avoid the problem of uplift.
A graded filter material consists of layers of pervious material, which permit flow of water but
prevent the movement of soil particles. The soil particles in particular layer are courser than that in
the preceding layer, however, the difference of sizes of the particles in the two layers should not be
excessive otherwise the particles of the preceding layer will be carried into the next layer.

An effective filter material may be designed using a few simple rules and the grading characteristics
of the soil to be protected.

1. The filter material should be coarse enough so that the percolating water moves easily without
any build up of water pressure in the filter.
- For a filter to provide free drainage.
i.e The d15 size of the filter should lie b/n four times d15 for the soil & four times d85
for the soil.
i.e 4 x d15 (soil) < d15 (filter) < 4 x d85 (soil).
2. The filter material should be fine enough so that the soil particles of the base material are not
washed through the filter.
d15 (filter) < 4 x d85 (soil)
3. The filter should have a fines content (particle size < 75µm) of not more than 5%.
4. The grading curve of the filter should have the same approximate shape as that of the base soil
(parallel Grading curve).
5. The filter should not contain material of particle grater than 80mm. (70 avoid segregation).
6. The base soil content exceeding a grain size of 19mm should be discounted.
7. The thickness & area of the filter should be sufficient to carry the seepage discharge safely.
8. The d85 size of the filter should be not less than twice the inside pipe diameter or screen mesh
size (where applicable).

Soil Mechanics I: Lecture Notes 9