CE-837

Topic 3

Seepage Analysis

Dr. Turab Jafri

1
  Introduction

Contents  Flow Nets

 Seepage through Earth Dam

 Casagrande’s method
 Seepage through CFRD

2
Introduction

3
3.1 Introduction

Seepage is considered to be all movement of water from the reservoir through the embankment,
abutments, and foundation. This includes porous media (intergranular) flow, flow in fractures, and
concentrated flow through “defects” such as cracks, loose lifts, etc. Soil is a porous media; depending on
the pore size and other contributing factors, water can flow through it under different gradients. Water
in the reservoir retained by an embankment (dam) is expected to flow (or seep) through the
embankment as well as from the foundation soils. The seeping water, in addition to causing a loss in
storage, can also be detrimental to the stability of the dam embankment and/or the foundation soils.
Flow of water through the soil may be one, two or three dimensional. For earth dams with
impermeable abutments, the flow is mostly two dimensional.

4
 3.1 Introduction

Flow of water through the soil may either be a:

• Steady State Flow
If the structure of soil remains unchanged as the water flows through it, the flow is termed as a steady
state (laminar) flow, or it indicates that the soil and flowing water are in a state of equilibrium and no
change is being affected in flow pattern or in soil structure. The quantity of water entering an element of
soil equals the amount leaving the element.
• Transient Flow
During Transient (or unsteady) flow, the structure of soil does not remains the same, i.e., the volume of
water changes during transient flow. The quantity of water entering a soil element is not equal to the
volume of water leaving the soil element.

5
Flow Nets

6
3.2 Flow Nets
3.2.1 Concept of Flow Nets
A flow line is a line along which a water particle will travel from upstream to the downstream side in the
permeable soil medium.
An equipotential line is a line along which the potential head at all points is equal. Thus, if piezometers
are placed at different points along an equipotential line, the water level will rise to the same elevation
in all of them.

7
3.2 Flow Nets
3.2.1 Concept of Flow Nets
A combination of a number of flow lines and equipotential lines is called a flow net. Flow nets are
constructed for the calculation of groundwater flow and the evaluation of heads in the media. To
complete the graphic construction of a flow net, one must draw the flow and equipotential lines in such
a way that:
1. The equipotential lines intersect the flow lines at right angles.
2. The flow elements formed are approximate squares.

In any flow net, the strip between any two adjacent flow lines is called a flow channel.

8
3.2 Flow Nets
3.2.1 Concept of Flow Nets

9
3.2 Flow Nets
3.2.1 Concept of Flow Nets
The following four boundary conditions apply:
1. The upstream and downstream surfaces of the permeable layer (lines ab and de) are equipotential
lines.
2. Because ab and de are equipotential lines, all the flow lines intersect them at right angles.
3. The boundary of the impervious layer—that is, line fg—is a flow line, and so is the surface of the
impervious sheet pile, line acd.
4. The equipotential lines intersect acd and fg at right angles.

10
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
A flow channel with the equipotential lines forming square elements is shown in the figure given below.
Let h1, h2, h3, h4, . . ., hn be the piezometric levels corresponding to the equipotential lines. The rate of
seepage through the flow channel per unit length (perpendicular to the section through the permeable
layer), considering steady flow, can be calculated as follows.

From Darcy’s law, the flow rate (considering unit thickness) is equal to kiA. Thus,

11
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
If the flow elements are drawn as approximate squares, the drop in the piezometric level between any
two adjacent equipotential lines is the same. This is called the potential drop. Thus,

12
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
If the number of flow channels in a flow net is equal to Nf , the total rate of flow through all the
channels per unit length can be given by:

13
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
Although drawing square elements for a flow net is convenient, it is not always necessary. Alternatively,
one can draw a rectangular mesh for a flow channel, provided that the width-to-length ratios for all the
rectangular elements in the flow net are the same. In this case, the rate of flow through a single channel
can be modified to:

14
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
If b1/l1=b2/l2=b3/l3=, . . . .,=n (i.e., the elements are not square), then

15
3.2 Flow Nets
3.2.2 Seepage calculation from a Flow Net
Given that kx = kz = k = 5 x 10-3 cm/sec, determine
a. How high (above the ground surface) the water will rise if piezometers are placed at points a and b
b. The total rate of seepage through the permeable layer per unit length
c. The approximate average hydraulic gradient at c

16
3.2 Flow Nets
3.2.3 Flow Nets in Anisotropic Soil
Most soils exhibit some degree of anisotropy. To account for soil anisotropy with respect to hydraulic
conductivity, flow net construction must be modified. For anisotropic soils,
The rate of seepage per unit length can be calculated by

17
3.2 Flow Nets
3.2.3 Flow Nets in Anisotropic Soil
A dam section is shown in Figure. The hydraulic conductivity of the permeable layer in the vertical and
horizontal directions are 2 x 10-2 mm/s and 4 x 10-2 mm/s, respectively. Calculate the seepage loss of the
dam in m3/day/m.

18
3.2 Flow Nets
3.2.4 Uplift Pressure under Hydraulic Structures

19
3.2 Flow Nets
3.2.4 Uplift Pressure under Hydraulic Structures

20
Seepage through Earth Dam

21
3.3 Seepage through Earth Dam
3.3.1 Schaffernak’s method
Figure shows a homogeneous earth dam resting on an impervious base. Let the hydraulic conductivity of
the compacted material of which the earth dam is made be equal to k. The free surface of the water
passing through the dam is given by abcd. It is assumed that abc is parabolic. The slope of the free
surface can be assumed to be equal to the hydraulic gradient. It is also is assumed that, because this
hydraulic gradient is nearly constant with depth (Dupuit, 1863), (discharge proportional to aquifer
thickness)

22
3.3 Seepage through Earth Dam
3.3.1 Schaffernak’s method
Considering the triangle cde, we can obtain the rate of seepage per unit length of the dam as,

23
3.3 Seepage through Earth Dam
3.3.1 Schaffernak’s method
The rate of seepage (per unit length of the dam) through the section bf is,

24
3.3 Seepage through Earth Dam
3.3.1 Schaffernak’s method
The rate of seepage (per unit length of the dam) through the section bf is,

25
3.3 Seepage through Earth Dam
3.3.2 Schaffernak’s method - Casagrande’s Correction
Casagrande experimentally showed that the parabolic free surface starts from a’ , not a
Calculate 0.3Δ

26
3.3 Seepage through Earth Dam
3.3.2 Schaffernak’s method - Casagrande’s Correction
Given that β = 45°, α = 30°, B = 10 m, H = 60 m, height of dam = 76 m, and k = 61 x 10-6 m/min,
calculate the seepage rate, q, in m3/day/m length.

27
Casagrande’s method for Seep
age through Earth Dam

28
3.4 Casagrande’s method

It was shown by Casagrande (1932) that, when the downstream slope angle α in becomes greater than
30°, deviations from Dupuit’s assumption become more noticeable.

29
Seepage through CFRD

30
 3.5 Seepage through CFRD
3.5.1 Theories on flow through rockfills
• If the permeability of the rockfill is 10-3 cmls or less, it is to be analyzed according to the theories of
soil mechanics. Thus, the flow is laminar and controlled by Darcy's law expressed by v = ki.

• To ensure a permeability of 10-3 cmls, the fine fraction of rockfill comprised of rock powder of sand
size, rock fragments, and even soil would have to fill the voids between the rock blocks and therefore
control the flow.

• At Itaúba Dam (Brazil - rockfill with clay core), it was common to see tiny water pools on top of the
rockfill. A practical test to decide whether to approve or reject the rockfill dam design consisted of
digging a shallow trench in the downstream rockfill with the blade of a scraper, filling it with water
and taking a walk. If soon after returning from that walk the water was still there, the rockfill would
have to be removed; but if the water was not there, the rockfill could stay. Impoundment of water
on the surface in the form of pools could result in increased overall weight of the dam and a
decrease in shear strength of the material which can destabilize the dam body.

31
3.5 Seepage through CFRD
3.5.1 Theories on flow through rockfills
A simple calculation can be made to estimate the rockfill permeability. If v = ki and the flow is vertical,
i = 1. In one hour the travelled distance would be d = 3600 v (= 3600 k), and for a water level of 50 cm,
the required permeability would be k = 0.0138 cm/s = 1.38x10-2 cm/s.
This analysis suggests that rockfills must have a minimum k in the order of 10-2 cmls to avoid the
occurrence of water impounded on the surface.

Whenever rockfills with fines are used in CFRDs, it is recommended they are placed in the central area,
because if the amount of fines is excessive the shear strength will drop and the stability of the slopes
could be jeopardized.

32
3.5 Seepage through CFRD
3.5.1 Theories on flow through rockfills
The research developed on rockfills is aimed at two issues.
• An attempt to establish an equation for the water flow velocity ( v = ki ), or even to define an
equivalent to "permeability coefficient" or discharge velocity for rockfills.
• An attempt to define the maximum permissible flow through a rockfill that would not endanger the
downstream slope stability.

Marulanda and Pinto (2000) claim that adequate transitions are necessary in order to avoid migration of
fines to the upstream and downstream shells of the dam.

33
3.5 Seepage through CFRD
3.5.1 Theories on flow through rockfills
The research developed on rockfills is aimed at two issues.
• An attempt to establish an equation for the water flow velocity ( v = ki ), or even to define an
equivalent to "permeability coefficient" or discharge velocity for rockfills.
• An attempt to define the maximum permissible flow through a rockfill that would not endanger the
downstream slope stability.

Back in 1933, Lindquist (cited in Cruz, 1979) noted that in seepage tests on glass beads of 2.7 cm
diameter placed in a transparent tube flow was laminar only up to a gradient of 0.7.

34
3.5 Seepage through CFRD
3.5.2 Critical aspects for stability – (a) Flows
Rockfill structures might fail if submitted to a throughflow above a critical value, called the critical flow.

Olivier (1967) established the inclination of stable slopes for rock blocks dumped in running water.

35
3.5 Seepage through CFRD
3.5.2 Critical aspects for stability – (a) Flows
Leps (1973), reviewing the work of Olivier, points out the characteristics and conditions that govern the
stability of the water discharging slopes:
• Rock properties: specific weight, dominant diameter, gradation and shape of the blocks;
• Relative density of the rockfill;
• Maximum hydraulic gradient;
• Inclination of downstream slope.

36
3.5 Seepage through CFRD
3.5.2 Critical aspects for stability – (a) Flows
Thomas (1976) computed the maximum permissible flows through a compacted rockfill considering
three equivalent (or dominant) diameters and obtained the values shown in Table.

A formula to compute the critical flow was proposed by Olivier. The metric version was developed by
Collet (cited in Thomas, 1976) and is:

37
3.5 Seepage through CFRD
3.5.2 Critical aspects for stability – (b) Downstream slope stabilit
Factors that seem to influence either the stabilityyor the instability of rockfills are:

• Placement conditions, e.g. whether the rockfill was dumped or compacted;

• The average gradient in the rockfill;
• The exit height of phreatic line;
• The discharge rate.
• Rock block specific gravity;
• Rockfill dominant size;
• Rockfill block gradation and shape;
• Downstream rockfill slope inclination.

In fact any description of the failures observed in rockfills during throughflows always mentions that
dams failed by progressive sliding and removal of rock blocks from the toe up to the crest. The problem,
therefore, is clearly the seepage forces that develop during throughflow and that are strong enough to
uplifit and remove rock blocks from the slope, which leads to progressive failure.

38
3.5 Seepage through CFRD
3.5.2 Critical aspects for stability – (b) Downstream slope stabilit
y of the upstream water level and a constant friction
A simplified analysis considering different positions
angle of the rockfill is shown in the Table below for a dam 140 m high.
The safety factors were computed using Hoek and Bray (1974) stability charts, for a slope of
1.3(H):1.0(V).Data from Table show that for a loose rockfill (ϕ= 50°), ∆H/H that leads to instability is in
the order of 0.10. A compacted rockfill (ϕ = 60°) would support ∆H/H up to 0.25 or more.

39