Em - 1110-2-1901 Seepage Analysis and PDF

CECW-EG Department of the Army EM 1110-2-1901
U.S. Army Corps of Engineers

Washington, DC 20314-1000
Engineer Manual 30 September 1986
1110-2-1901 (original)
30 April 1993
(change 1)
Engineering and Design
SEEPAGE ANALYSIS AND

CONTROL FOR DAMS
Distribution Restriction Statement

Approved for public release;
distribution is unlimited.
DEPARTMENT OF THE ARMY EM 1110-2-1901
U.S. Army Corps of Engineers Change 1
CECW-EG-8 Washington, DC 20314-1000
Engineer Manual
No. 1110-2-1901 30 April 1993

SEEPAGE ANALYSIS AND CONTROL FOR DAMS
1. This change replaces Appendix D, "Filter and Drain Design and Construction" of EM 1110-2-1901,
dated 30 September 1986.
2. File this change sheet in front of the publication for reference purposes.
FOR TilE COMMANDER:
WILUAM D. BROWN
Colonel, Corps of Engineers
Chief of Staff
EM 1110-2-1901
ENGINEER MANUAL 30 September 1986
ENGINEERING AND DESIGN
SEEPAGE ANALYSIS AND

CONTROL FOR DAMS
DEPARTMENT OF THE ARMY

CORPS OF ENGINEERS
OFFICE OF THE CHIEF OF ENGINEERS
DAEN-ECE-G

No. 1110-2-1901

1. Purpose. This manual presents the fundamental design principles and

guidance concerning seepage considerations for design of new dams and the
evaluation of existing projects.
2. Applicability. This manual is applicable to all HQUSACE/OCE elements and

field operating activities having responsibility for the design and
construction of civil works projects.
3. Discussion. All earth and rock-fill dams are subject to seepage through
the embankment, foundation, and abutments. Concrete gravity and arch dams are
subject to seepage through the foundation and abutments. Seepage control is
necessary to prevent excessive uplift pressures, sloughing of the downstream
slope, piping through the embankment and foundation, and erosion of material
by loss into open joints in the foundation and abutments. The purpose of the
project, i.e., long-term storage, flood control, etc., may impose limitations
on the allowable quantity of seepage.
FOR THE COMMANDER:
Chief of Staff
This manual supersedes EM 1110-2-1901 dated February 1952

DEPARTMENT OF THE ARMY EM-1110-2-1901
US Army Corps of Engineers
DAEN-ECE-G Washington, D. C. 20314-1000

No. 1110-2-1901

Table of Contents
Subject Paragraph Page
CHAPTER 1. INTRODUCTION
Purpose------------------------------------ 1-1 1-1
Applicability ------------------------------ 1-2 1-1
References--------------------------------- 1-3 1-1
Objective and Scope------------------------ 1-4 1-1
General Considerations--------------------- 1-5 1-1
CHAPTER 2. DETERMINATION OF PERMEABILITY OF SOIL

AND CHEMICAL COMPOSITION OF WATER
Darcy's Law-------------------------------- 2-1 2-1
Range of Validity of Darcy's Law----------- 2-2 2-4
Coefficient of Permeability---------------- 2-3 2-6
Factors Influencing Permeability----------- 2-4 2-9
Indirect Methods for Determining
Permeability----------------------------- 2-5 2-20
Laboratory Methods for Determining
Permeability----------------------------- 2-6 2-28
Origin, Occurrence, and Movement of
Ground Water ----------------------------- 2-7 2-30
Field Methods for Determining
permeability----------------------------- 2-8 2-33
Chemical Composition of Ground Water and
River (or Reservoir) Water--------------- 2-9 2-34
CHAPTER 3. DETERMINATION OF PERMEABILITY OF ROCK
Permeabilities of Rock Masses-------------- 3-1 3-1

Flow Characteristics in Rock Masses-------- 3-2 3-1
Methods for Determining Rock Mass
Permeability----------------------------- 3-3 3-3
Applications of Rock Mass Permeability----- 3-4 3-11
This manual supersedes EM 1110-2-1901, Feb 52
i
EM 1110-2-1901
30 Sep 86
CHAPTER 4. SEEPAGE PRINCIPLES
General Considerations---------------------- 4-1 4-1

Boundary Conditions------------------------- 4-2 4-1
Confined and Unconfined Flow Problems------- 4-3 4-2
Laplace’s Equation------------------------ 4-4 4-3
Methods for Solution of Laplace's
Equation------------------------------ 4-5 4-5
Graphical Method for Flow Net
Construction----------------------------- 4-6 4-9
Flow Net for Anisotropic Soil--------------- 4-7 4-14
Flow Net for Composite Sections------------- 4-8 4-14
Determination of Seepage Quantities,
Escape Gradients, Seepage Forces, and
Uplift Pressures-------------------------- 4-9 4-21
CHAPTER 5. CONFINED FLOW PROBLEMS

Gravity Dam on Pervious Foundation of
Finite Depth------------------------------ 5-2 5-1
Gravity Dams on Infinitely Deep
Pervious Foundations------------------------ 5-3 5-1
CHAPTER 6. UNCONFINED FLOW PROBLEMS
Introduction -------------------------------- 6-1 6-1

Homogeneous Earth Dam on Impervious
Foundation -------------------------------- 6-2 6-1
Earth Dam with Horizontal Drain on
Impervious Foundation--------------------- 6-3 6-10
Earth Dam with Toe Drain on Impervious
Foundation -------------------------------- 6-4 6-10
Earth Dam with Vertical or near Vertical
Horizontal Drains on Impervious
Foundation-------------------------------- 6-5 6-10
Flow Net for a Composite Zoned Dam---------- 6-6 6-10
Zoned Earth Dam on Pervious Foundation------ 6-7 6-17
CHAPTER 7. SEEPAGE TOWARD WELLS
Use of Wells -------------------------------- 7-1 7-1

Analysis of Well Problems------------------- 7-2 7-1
Basic Well Equations for Steady State
Flow-------------------------------------- 7-3 7-2
Special Conditions-------------------------- 7-4 7-11
Nonsteady State Flow------------------------ 7-5 7-21
ii
EM 1110-2-1901
30 Sep 86
CHAPTER 8. SEEPAGE CONTROL IN EMBANKMENTS
General------------------------------------- 8-1 8-1

Methods for Seepage Control----------------- 8-2 8-1
Flat Slopes Without Drains------------------ 8-3 8-1
Zoning Embankments-------------------------- 8-4 8-1
Vertical (or Inclined) and Horizontal
Drains------------------------------------ 8-5 8-8
Seepage Control Against Earthquake
Effects----------------------------------- 8-6 8-18
CHAPTER 9. SEEPAGE CONTROL IN EARTH FOUNDATIONS
General------------------------------------- 9-1 9-1

Selection of Method for Seepage
Control ----------------------------------- 9-2 9-1
Horizontal Drains--------------------------- 9-3 9-1
Cutoffs------------------------------------- 9-4 9-1
Upstream Impervious Blanket----------------- 9-5 9-48
Downstream Seepage Berms-------------------- 9-6 9-58
Relief Wells -------------------------------- 9-7 9-62
Trench Drain-------------------------------- 9-8 9-83
Concrete Galleries -------------------------- 9-9 9-84
CHAPTER 10. SEEPAGE CONTROL THROUGH EARTH ABUTMENTS

ADJACENT TO STRUCTURES AND BENEATH
SPILLWAYS AND STILLING BASINS
Through Earth Abutments--------------------- 10-1 10-1

Adjacent to Outlet Conduits----------------- 10-2 10-1
Beneath Spillways and Stilling Basins------- 10-3 10-3
CHAPTER 11. SEEPAGE CONTROL IN ROCK FOUNDATIONS AND

ABUTMENTS

Cutoff Trenches----------------------------- 11-2 11-1
Abutment Impervious Blankets---------------- 11-3 11-2
Drainage and Grouting Galleries and
Tunnels----------------------------------- 11-4 11-2
Grouting of Foundations and Abutments------- 11-5 11-3
Surface Treatment of Foundations and
Abutments--------------------------------- 11-6 11-3
CHAPTER 12. REMEDIAL SEEPAGE CONTROL

Remedial Methods ---------------------------- 12-2 12-1
Storage Restriction------------------------- 12-3 12-2
Grouting ------------------------------------ 12-4 12-2
iii
EM 1110-2-1901
30 Sep 86
Upstream Impervious Blanket----------------- 12-5 12-4

Downstream Berm----------------------------- 12-6 12-6
Slurry Trench Cutoff------------------------ 12-7 12-6
Relief Wells -------------------------------- 12-8 12-11
Drainage of Downstream Slope---------------- 12-9 12-12
CHAPTER 13. MONITORING PERFORMANCE OF SEEPAGE CONTROL

MEASURES

Piezometers for Seepage Pressures----------- 13-2 13-1
Flow Measurements--------------------------- 13-3 13-12
Seepage Water Analysis---------------------- 13-4 13-18
Remote Sensing Methods---------------------- 13-5 13-20
CHAPTER 14. INSPECTION, MAINTENANCE, AND REHABILITA-

TION OF SEEPAGE CONTROL MEASURES
Introduction -------------------------------- 14-1 14-1

Inspection---------------------------------- 14-2 14-1
Maintenance--------------------------------- 14-3 14-1
Rehabilitation------------------------------ 14-4 14-3
APPENDIX A. REFERENCES
Government Publications--------------------- A-1
Non-Government Publications----------------- A-11
APPENDIX B. APPROXIMATE METHODS FOR ANALYSIS OF FLOW

PROBLEMS
Introduction-------------------------------- B-1 B-1

Electrical Analogy-------------------------- B-2 B-1
Sand Tank Model----------------------------- B-3 B-2
Viscous Flow Models------------------------- B-4 B-5
Method of Fragments------------------------- B-5 B-5
Finite Difference Method-------------------- B-6 B-28
Finite Element Method----------------------- B-7 B-29
APPENDIX C. ANALYSIS OF PRESSURE INJECTION TESTS

(Ziegler 1976)
Water pressure Tests------------------------ C-1 C-1
Air pressure Tests-------------------------- C-2 C-11
pressure Holding Testy---------------------- C-3 C-17
iv
EM 1110-2-1901
30 Sep 86
APPENDIX D. FILTER AND DRAIN DESIGN AND CONSTRUCTION
General--------------------------------------- D-1 D-1

Stability------------------------------------- D-2 D-1
Permeability---------------------------------- D-3 D-2
Applicability--------------------------------- D-4 D-2
Perforated Pipe--------------------------- D-5 D-3
Gap-Graded Base--------------------------- D-6 D-3
Gap-Graded Filter------------------------- D-7 D-5
Broadly-Graded Base----------------------- D-8 D-5
Graded Filter Design---------------------- D-9 D-5
Geotextiles----------------------------------- D-10 D-6
Construction---------------------------------- D-11 D-6
Monitoring------------------------------------ D-12 D-8
v
EM 1110-2-1901
30 Sep 86
CHAPTER 1
INTRODUCTION
1-1. Purpose. This manual provides guidance and information concerning seep-
age analysis and control for dams.
1-2. Applicability. The provisions of this manual are applicable to all

HQUSACE/OCE elements and field operating activities (FOA) having responsibil-
ity for seepage analysis and control for dams.
1-3 References. Appendix A contains a list of Government and non-Government

references pertaining to this manual. Each reference is identified in the
text by either the designated publication number or by author and date.
Reference to cited material in tables and figures is identified throughout the
manual by superscripted numbers (item 1, 2, etc.) that correspond to similarly
numbered items in Appendix A.
1-4. Objective and Scope. The objective of this manual is to provide a guide
for seepage analysis and control for dams.
1-5. General Considerations. All earth and rock-fill dams are subject to
seepage through the embankment, foundation, and abutments. Concrete gravity
and arch dams are subject to seepage through the foundation and abutments.
Seepage control is necessary to prevent excessive uplift pressures, sloughing
of the downstream slope, piping through the embankment and foundation, and
erosion of material by loss into open joints in the foundation and abutments.
The purpose of the project, i.e., long-term storage, flood control, etc., may
impose limitations on the allowable quantity of seepage (Sowers 1977).
1-1
EM 1110-2-1901
30 Sep 86
CHAPTER 2
DETERMINATION OF PERMEABILITY OF SOIL AND
CHEMICAL COMPOSITION OF WATER
2-1. Darcy's Law.
a. Development of Darcy's Law. Henry Darcy, a French engineer, con-

ducted a laboratory experiment to study the flow of water in verticals and
filters which he published in his 1856 treatise. The results of his experiment
indicated that (Rouse and Ince 1957)
v = ki (2-1)
or since Q = VA
Q = kiAt (2-2)
or using q =
q = kiA (2-3)
\where
v = discharge velocity
k = Darcy's coefficient of permeability(1)
i = hydraulic gradient (head loss/length over which head loss occurs)
Q = quantity of discharge
A = cross-sectional area of flow
t = time of flow
q = rate of discharge
b. Extension to Inclined Soil Column. Darcy's law may be extended to

flow through an inclined soil column given in figure 2-1 (Harr 1962). As
indicated in equation 2-1, flow is a consequence of differences in total
(1) Commonly called the coefficient of permeability or the permeability.
2-1
EM 1110-2-1901
30 Sep 86
head(l) and not of pressure gradients (Harr 1962 and Bear 1972). As shown in
figure 2-1, flow is directed from point A to point B even though the pressure
at point B is greater than that at point A.
Figure 2-1. Darcy's law for flow through inclined soil column
(prepared by WES)
(1) The elevation head at any point is the distance from some arbitrary datum.
The pressure head is the water pressure divided by the unit weight of the
water. The total head is the sum of the elevation head and the pressure
head.
2-2
EM 1110-2-1901
30 Sep 86
c. Discharge Velocity and Seepage Velocity. The discharge velocity is

defined as the quantity of fluid that flows through a unit cross-sectional area
of the soil oriented at a right angle to the direction of flow in a unit time.
The discharge velocity is used in determining the quantity of flow or rate of
discharge through a soil. As flow can occur only through the interconnected
pores of the soil, as shown in figure 2-2, the actual rate of
Figure 2-2. Concepts of flow paths through a soil column

(prepared by WES)
movement of the water, as measured with dye tracers for instance, is the seep-
age velocity (Harr 1962 and Casagrande 1937) which exceeds the discharge
velocity.
(2-4)
2-3
EM 1110-2-1901
30 Sep 86
where
seepage velocity
n = porosity (ratio of volume of voids to the total volume of the soil

mass)
It follows that
Equation 2-5 is a useful expression in estimating field permeabilities using

dye tracers (Soil Conservation Service 1978).
2-2. Range of Validity of Darcy's Law.
a. Lower Bound. Darcy's law (equations 2-1 through 2-3) applies to

linear flow (adjacent flow lines are locally straight and parallel). For flows
through soils, there are two situations where the validity of this linear
relationship may not hold. For highly plastic clays of low permeability, there
may be a threshold hydraulic gradient below which flow does not take place.
Such conditions may occur in deeply buried clays and clay shales. For many
practical seepage problems the rate of flow through these soil layers is so
small that they can be considered to be impervious (Mitchell 1976, Chugaev
1971, Basak and Madhav 1979, and Muskat 1946).
b. Upper Bound. Of greater practical importance is the upper limit on

the range of validity of Darcy's law. It has been recognized that, at very
high flow rates, Darcy's law does not hold (Chugaev 1971). The upper limit is
usually identified using Reynolds number, a dimensionless number that expresses
the ratio of internal to viscous forces during flow. It is often used in fluid
mechanics to distinguish between laminar flow (fluid layer flows alongside of
another at approximately the same velocity with no macroscopic mixing of fluid
particles) at low velocities and turbulent flow (velocity fluctuations, both
parallel and transverse, are imposed upon the mean motion with mixing of the
fluid particles) at high velocities. The Reynolds number for flow through
soils is
(2-6)
where
Reynolds number
D = average diameter of soil particles
2-4
EM 1110-2-1901
30 Sep 86
= density of fluid
= coefficient of dynamic viscosity of fluid
The critical value-of Reynolds number at which the flow in soils changes from
laminar to turbulent has been determined experimentally by various investiga-
tors to range from 1 to 12 (Harr 1962 and Chugaev 1971). Assuming a water
-3
temperature of 20º C, substituting = 998.2 kg/m3 and µ = 1.002 x 10 kg/m
sec into equation 2-6, and assuming values of D and solving for v with
= 1 and = 12 gives the relationship shown in figure 2-3 which defines
the upper bound of the validity of Darcy's law. Depending on the discharge
velocity, Darcy's law is generally applicable for silts through medium sands.
c. Turbulent Flow.
(1) Estimating Permeability from Empirical Equation. For flow through

soils more pervious than medium sands, flow is likely to be turbulent. Under
turbulent conditions, the seepage velocity in a material with monosized soil
particles (coarse sands and/or gravels) can be estimated from the following
equation (Wilkins 1956, Leps 1973, and Stephenson 1979).
(2-7)
where
= seepage velocity in inches per second
w = an empirical constant, which depends on the shape and roughness of

the soil particles and viscosity of water and varies from 33 for
crushed gravel to 46 for polished marbles, in inch 1/2 per second
M = hydraulic mean radius of the rock voids (for a given volume of

particles equal to the volume of voids divided by the total surface
area of the particles, or the void ratio divided by the surface area
per unit volume of solids) in inches
i = hydraulic gradient
The coefficient of permeability is obtained from the seepage velocity using

equation 2-5. For well-graded soils, the D50 size (50 percent finer by weight)
can be used to calculate the hydraulic mean radius provided that the minus
1-in.-size material is less than 30 percent by weight. If there is more than
30 percent of minus 1-in .-size material, the permeability should be determined
experimentally (Leps 1973).
(2) Determining Permeability Experimentally. Alternatively, for flow

through soils more pervious than medium sands, the relationship between
hydraulic gradient and discharge velocity can be determined experimentally
(Cedergren 1977).
2-5
EM 1110-2-1901
30 Sep 86
Figure 2-3. Boundary between laminar and turbulent flow determined

using Reynolds number for temperature of 20° C (prepared by WES)
2-3. Coefficient of Permeability.
a. Darcy's (Engineer's) Coefficient of Permeability. The coefficient

of permeability used in seepage analysis for dams is called the Darcy's or
engineer's coefficient and is given by (Cedergren 1977)
2-6
EM 1110-2-1901
30 Sep 86
(2-8)
or since Q = VA
(2-9)
or using q =
(2-10)
The coefficient of permeability is defined as the rate of discharge-of water

at a temperature of 20º C under conditions of laminar flow through a unit
cross-sectional area of a saturated soil medium. The coefficient of perme-
ability has the dimensions of a velocity and is usually expressed in centi-
meters per second. Permeability computed on the basis of Darcy's law is
limited to the conditions of laminar flow and complete saturation of the soil.
Under conditions of partial saturation, the flow is in a transient state and
is time dependent. To analyze natural flow conditions which depart from the
Darcy flow condition, it is sometimes necessary to apply Darcy's law in condi-
tions where it is not strictly valid. When this is done, the effects of tur-
bulent flow and partial saturation on the permeability must be recognized and
taken into consideration (Cedergren 1975).
b. Intrinsic (Specific) Permeability. The coefficient of permeability

of a soil material varies for different pore fluids depending upon their
density and viscosity as follows:
(2-11)
where
= intrinsic permeability
= unit weight of pore fluid
µ = viscosity of pore fluid
The intrinsic permeability has the dimensions of length squared and is

expressed in square centimeters or Darcy's (equal to 1.01 x 10-8 cm2). Fig-
ure 2-4 is a chart for the conversion of permeability values from one set of
2-7
EM 1110-2-1901
30 Sep 86
2-8
10-5 w-6
~u~~~~u~~?~/_fo~,r-~~a-te_r~f·~t,
1ol 2 10
2
~,?u~~~~)~·~--ul~l!~•~·~·~·~~~~~~~u·~·~·~•u_u_
10 1
__~~~~~~_u~~-L-L--~~-LL-L-~ULLL~~--~~~~~--liLuuU-L-L-~
10-
1-- High -4)+1<<----Medium - - - - k - - - - - Low - - - - - * < - - - - V e r y low----k-----Practically impermeable ------'l-.1
• For water at 2o•c To convert from em/sec to: m/min JJ/sec ft/sec ft/min ft/yr Darcy*
Multiply by: 0.600 10 4 0.0328 1.968 1.034 )( 10 6 1.045 )( 103
201
Figure 2-4. Permeability conversion chart (courtesy of John Wiley & Sons, Inc. )
EM 1110-2-1901
30 Sep 86
units to another (Lohman et al. 1972). Substituting equation 2-11 into

equation 2-1
(2-12)
Equation 2-12 may be used when dealing with more than one fluid or with temper-
ature variations. This is widely used in the petroleum industry where the
presence of gas, oil, and water occur in multiphase flow systems (Freeze and
Cherry 1979, and Bureau of Reclamation 1977). In seepage analysis for earth
dams where we are primarily interested in the flow of water subject to small
changes in temperature, this refinement is seldom required.
c. Transmissivity Factor. In order to describe the flow characteris-

tics of an aquifer (saturated permeable geologic unit that can transmit
significant quantities of water under ordinary hydraulic gradients),
C. V. Thesis introduced the term transmissivity which is defined as (Bureau of
Reclamation 1977)
T = kt (2-13)
where
T = transmissivity factor
k = average permeability
t = aquifer thickness
Transmissivity represents the rate of discharge for a gradient of unity

through a vertical strip of aquifer one unit wide and has dimensions of length
squared per unit time and is usually expressed in square feet per day.
2-4. Factors Influencing Permeability.
a. Range of Values of Permeability. No other property of soil exhibits

a wider range of values (up to ten orders of magnitude) or shows greater
directional (anisotropy) and spatial variability in a given deposit as does
the coefficient of permeability. The approximate range in coefficients of
permeability for soils and rocks is shown in figure 2-5 (Milligan 1976).
Within the range, extreme variations of permeability in situ are possible due
to the degree of stratification or heterogeneity of the soil deposit.
b. Variation of In Situ Permeability. Natural soil deposits are

generally stratified in structure. Water-deposited soils are laid down in
'horizontal layers and are often more permeable in the horizontal than vertical
direction. Windblown sands and silts are generally more permeable vertically
than horizontally due to the presence of continuous vertical root holes. An
important example of stratification is openwork gravel which may occur in
2-9
EM 1110-2-1901
30 Sep 86
Figure 2-5. Approximate range in coefficient of

224
permeability of soils and rocks (from Milligan )
ordinary gravel or soil and have tremendous influence on the watertightness of

dam foundations and abutments as shown in figure 2-6 (Cedergren 1977). Fig-
ure 2-6a shows a soil profile surmised from several drill holes. The grain
size analysis of soil samples taken at frequent intervals erroneously indicated
that the deposit was composed of relatively uniform sandy gravels. Laboratory
permeability tests on disturbed samples produced coefficients of permeability
-6
of about 1 x 10 cm/sec. Using this value of permeability, the probable
seepage loss beneath the proposed dam was estimated to be 3 cu ft/day, which is
an insignificant quantity. However, the design engineer had observed many
openwork streaks in which the fines fraction of the material was almost
completely absent along the banks of the river and noted that the ground-water
table was level for several hundred feet away from the river and fluctuated
rapidly with changes in river stage. Field pumping tests were conducted which
indicated somewhat variable permeabilities but none approaching the magnitude
of openwork gravels. Based upon the available data, the dam was designed with
a cutoff trench to bedrock. During the excavation of the cutoff trench,
streaks of openwork gravel were found throughout the foundation. A revised
seepage computation based on a permeability of 30 cm/sec indicated that without
the cutoff trench, the theoretical underseepage would be about 1,000,000 cu
ft/day. If openwork gravel or other important discontinuities in earth dam
foundations remain undetected, serious problems from excessive seepage and
hydrostatic pressures will develop. This example illustrates the potential
serious effects of deviations between the design assumptions and the as-built
dam (Cedergren 1977). Also, thin continuous seams of cohesive soil can
drastically alter the vertical flow through what would otherwise be a highly
permeable site.
2-10
EM 1110-2-1901
30 Sep 86
a. Soil profile surmised from b. Soil profile revealed by

drill holes with estimated cutoff trench with esti-
quantity of seepage under mated quantity of seep-
dam equal to 3 cu ft/day age under dam equal to
1,000,000 cu ft/day
(without cutoff trench)
Figure 2-6. Influence of openwork gravel on underseepage

155)
(courtesy of John Wiley & Sons, Inc.
c. Properties of the Seepage Fluid. The properties of the seepage

fluid which influence the permeability of soils are the temperature, density,
viscosity, and chemical composition.
(1) As shown in table 2-1, for the range of temperatures ordinarily

encountered in seepage analysis of dams (0º C to 40º C) the density of water
is nearly constant (varies less than 1 percent).
Table 2-1. Properties of Watera
Temperature Density Viscosity

°C ºF kg/m3 kg/m sec
-3
0 32 999.8 1.787 x 10
5 41 999.9 1.519
10 50 999.7 1.307
15 59 999.1 1.139
20 68 998.2 1.002
25 77 997.0 0.890
30 86 995.6 0.798
35 95 994.0 0.719
40 104 992.2 0.653
a
Prepared by WES.
2-11
EM 1110-2-1901
30 Sep 86
(2) The viscosity varies up to 63.5 percent over the range of tempera-
tures ordinarily encountered in seepage analysis of dams. As indicated in
equation 2-11, the permeability is inversely proportional to the viscosity of
the water. As given in table 2-1, the viscosity of water decreases as tempera-
ture increases. Therefore, the coefficient of permeability of the soil
increases as the temperature of the water increases. Permeability tests are
run at the most convenient temperature and reported at 20º C.
(3) The total dissolved salts (TDS) present in the seepage water may
influence the permeability of the soil, particularly for cohesive soils (Quirk
and Schofield 1955 and Cary, Walter, and Harstad 1943). Available data indi-
cate that cohesive soils may be two to three orders of magnitude more permeable
to seepage water containing moderate amounts of dissolved salts (less than
300 parts per million by weight) than the distilled water (Carry, Walter, and
Harstad 1943).
d. Degree of Saturation. The degree of saturation of a soil
(2-14)
where
S = degree of saturation
= volume of water
Vv = volume of voids
has an important influence on permeability. A decrease in the degree of

saturation causes a decrease in the permeability as shown in figure 2-7 (Lambe
1951). When the degree of saturation is less than 85 percent, much of the air
would be continuous throughout the soil voids and Darcy's law would not hold.
When the degree of saturation is greater than 85 percent, most of air present
in the soil is in the form of small occluded bubbles and Darcy's law will be
approximately valid. The ratio of the permeability of the unsaturated sand to
the saturated sand at the same void ratio is given as (Scott 1963 and Parker
and Thornton 1976)
(2-15)
where
k = unsaturated permeability
us
m = constant with values between 2 (uniform grain size) and
4 (well-graded materials)
2-12
EM 1110-2-1901
30 Sep 86
Figure 2-7. Permeability versus degree of saturation for

200
various sands (courtesy of John Wiley & Sons, Inc. )
e. Hydraulic Gradient. The hydraulic gradient
(2-16)
where
H = head loss
L = length over which head loss occurs
2-13
EM 1110-2-1901
30 Sep 86
at which permeability is measured can have a significant influence on the

coefficient of permeability computed from Darcy's law under certain condi-
tions. The maximum hydraulic gradient for which laminar flow occurs for a
particular soil at a given density may be determined in the laboratory by
plotting the discharge velocity
(2-17)
versus the hydraulic gradient as shown in figure 2-8. A straight line

relationship indicates laminar flow
(2-18)
while deviations from the straight line at high gradients indicate turbulent
flow. Darcy's law for a fine sand, as shown in figure 2-8, is valid only for
the hydraulic gradient less than 2 for the loose state and 4.5 for the dense
state. For soils larger than a fine sand, Darcy's law is valid for progres-
sively smaller hydraulic gradients (Burmister 1948 and Burmister 1955).
f. Particle Size. For cohesive soils, the permeability increases with

increases in clay mineral size and increase in void ratio (ratio of the volume
of voids to the volume of solid particles in the soil mass) as shown in
figure 2-9 (Yong and Warkentin 1966).(l) For cohesionless soils, the size and
shape of the soil particles influence the permeability. Allan Hazan conducted
tests on filter sands for use in waterworks and found that for uniform loose
clean sands the permeability was given by (Taylor 1948)
(2-19)
where
k - coefficient of permeability in cm per second
D 1 0 = particle size in cm at which 10 percent of the material is finer

by weight (also known as Hazen's effective size)
(1) As shown in table 2-2, the exchangeable cation present influences the
permeability of clay minerals at constant void ratio (Scott 1963). The
permeabilities are much smaller when the exchangeable cation is sodium
which is one of the reasons why sodium montmorillonite is used to seal
reservoirs.
2-14
EM 1110-2-1901
30 Sep 86
Figure 2-8. Determination of maximum hydraulic gradient for

which laminar flow occurs for a fine sand (courtesy of
147
American Society for Testing and Materials )
Hazen's experiments were made on sands for which 0.1 mm < D10 < 0.3 mm and the
uniformity coefficient, C u < 5 , where
(2-20)
where
C u = uniformity coefficient
D 6 0 = particle size at which 60 percent of the material is finer by

weight
The coefficient 100 is an average of many values which ranged from 41 to 146,
but most of the values were from 81 to 117. Equation 2-19 makes no allowance
for variations in shape of the soil particles or void ratio.
2-15
EM 1110-2-1901
30 Sep 86
a. Edge view sketch to show relative

size and shape of clay particles
(dimension not shown in length)
b. Permeability versus void ratio for various

clay minerals
Figure 2-9. Influence of particle size and void ratio on

294
permeability of clay minerals (courtesy of Macmillan )
2-16
EM 1110-2-1901
30 Sep 86
Table 2-2. Coefficients of Permeability for Different Exchange

Cations and Void Ratios for Two Clay Mineralsa
Permeabilities are obtained by falling-head test on samples in consolidation

apparatus. Results indicate the following:
For montmorillonite at void ratio 8 the order of permeability in terms of

the exchangeable ion present is
for kaolinite at void ratio 1.5 the order is
For compacted soils it is also observed that the permeability is much lower
(x 10-l to 10-2) in soils compacted slightly wet of optimum than in soils
compacted dry of optimum; it is thought that this occurs because of the
parallel arrangement of clay platelets in the wetter material after
compaction.
a 251
Courtesy of Addison-Wesley Publishing Company, Inc. .
2-17
EM 1110-2-1901
30 Sep 86
g. Particle Shape and Surface Roughness. Cohesionless soil particles

have different particle shapes and surface roughness dependent on the distance
they have been transported by flowing water from the place of original ero-
sion. As shown in table 2-3, the measured permeability is several orders of
magnitude lower for angular sand particles with rough surfaces than for
rounded sand particles with smooth surfaces (Burmister 1948). For uniform
cohesionless soils, crushing of particles during compaction with resulting
decrease in permeability occurs to a higher degree in soils with angular
shapes and rough surfaces than in soils with rounded shapes and smooth sur-
faces. Crushing of particles during compaction leads to an increase in the
amount of silt-sized particles (smaller than No. 200 sieve or 0.074 mm) which
results in lower permeability. For this and other reasons (cementation in
limestones and arching due to particle angularity) crushed rock is generally
not used for filters in earth dams. Also, table 2-3 compares the measured
permeability with the permeability computed from equation 2-19 developed by
Hazen for uniform loose clean sands. The agreement between measured and
computed permeability is within one order of magnitude for uniform sands and
glass spheres. Therefore, Hazen's equation should be used only for uniform
sands (sphericity and roundness > 0.90). The sphericity and roundness may be
estimated for sands using figure 2-10 (Krumbein and Sloss 1951).
h. Void Ratio. The permeability increases as the void ratio increases.
(2-21)
where
e = void ratio
V s = volume of solids
There are considerable laboratory test data, shown in figure 2-11, which
indicate that a plot of void ratio versus log of coefficient of permeability
is frequently a straight line (Lambe and Whitman 1969).
i. Amount and Type of Fines Present. The permeability of sands and

gravels varies significantly with the amount and type of fines (material
smaller than the No. 200 sieve) (Barber and Sawyer 1952; Fenn 1966; Younger
and Lim 1972; Strohm, Nettles, and Calhoun 1967; Nettles and Calhoun 1967, and
Loudon 1952). As shown in figure 2-12a, the addition of 2.5 percent, by dry
weight, silt fines to concrete sand results in an order of magnitude decrease
in permeability (Barber and Sawyer 1952). The addition of 6.5 percent silt
fines to concrete sand decreases the permeability two orders of magnitude.
Similar results are obtained by the addition of somewhat larger amounts of clay
and limestone fines to concrete sand. As shown in figure 2-12b, the addition
of 2.0 percent silt fines to a sand-gravel mixture results in an order of
magnitude decrease in permeability (Barber and Sawyer 1952). The addition of
4.2 percent silt fines to sand-gravel mixture decreases the permeability
two orders of magnitude. Similar results are obtained by the addition of
2-18
Table 2-3. Influence of Particle Shape and Surface Roughness on Permeability of Sand(l)
Coefficient of Permeability
Relative
Void Density em/sec
Surface c<2) cO>
Tl~e of Material Rouahness S~hericity Roundness _ u_ _c_ Ratio ~ercent Measured Com~uted-Hazen( 4 )
Crushed Sudeten granite Very rough 0.68 0.10 3.09 I. 54 0.543 79 5.00 X 10- 5 1.21 X 10- 2
Nysd Klodzkd River sand Very rough 0.75 o. 15 5.47 2.20 0.451 79 5.19 X 10- 6 5.63 X 10- 3
Odra River sand Rough to smooth 0.84 0.55 2.58 0.87 0.595 58 9.41 X 10- 4 1.44 X 10- 2
Odra River sand Rough to smooth 0.87 0.65 2.42 0.83 0.608 63 5.44 X 10- 3 1.44 X 10- 2
Baltic Beach sand Smooth 0.87 0.85 I. 91 0.89 0.720 38 1.83 X 10- 3 1.32 X 10- 2
Glass spheres Very smooth 1.00 1.00 I. 38 0.92 0.691 67 6.45 X 10-J 6.40 X
10-3
2-19
146
(!)Courtesy of British Geotechnical Society
{2)
(3)
EM 1110-2-1901
cc 0 10 0 60
X
30 Sep 86
( 4) k • 100 D •
10
EN 1110-2-1901
30 Sep 86
Figure 2-10. Krumbein and Sloss standard chart for

visual estimation of sphericity and roundness of
cohesionless soils (courtesy of W. H. Freeman
198)
and Company
somewhat smaller amounts of clay and larger amounts of limestone, respec-

tively, to a sand-gravel mixture. As shown in figure 2-12c, the addition of
about 1 percent calcium montmorillonite fines to a uniform fine sand results
in an order of magnitude decrease in permeability, while over 10 percent
kaolinite fines would be required for a similar reduction in permeability
(Fenn 1966).
j. Summary of Factors Influencing Permeability. The significant

influence that various factors exert on the permeability emphasizes the
importance of duplicating field conditions when determining permeability in
the laboratory.
2-5. Indirect Methods for Determining Permeability.
a. Hazen's Equation. For uniform loose clean sands, classified SP in

the Unified Soil Classification System (U. S. Army Engineer Waterways Experi-
ment Station 1960), the permeability may be estimated from the previously
given Hazen's equation (Taylor 1948)
(2-22)
where k is in cm per second and D10 is in cm.
2-20
MEDIUM
0
i=
<(
~ 2.0 r-------+-------~~----~------~r-------+-------~-------+------~~------+-----~
0
0
>
2-21
7 5 1 3
10 10 10 10 6 10 10 10 10 I
COEFFICIENT OF PERMEABILITY, em/sec

SOIL IDENTIFICATION CODE
1. COMPACTED CALICHE 10. OTTAWA SAND 19. LEAN CLAY
2. COMPACTED CALICHE 11. SAND-GASPEE POINT 20. SAND-UNION FALLS
3. SILTY SAND 12. SAND-FRANKLIN FALLS 21. SILT -NORTH CAROLINA
4. SANDY CLAY 13. SAND-SCITUATE 22. SAND FROM DIKE
5. BEACH SAND 14. SAND-PLUM ISLAND 23. SODIUM-BOSTON BLUE CLAY
6. COMPACTED BOSTON BLUE CLAY 15. SAND-FORT PECK 24. CALCIUM KAOLINITE
EM 1110-2-1901
7. VICKSBURG BUCKSHOT CLAY 16. SILT -BOSTON 25. SODIUM MONTMORILLONITE
8. SANDY CLAY 17. SILT -BOSTON 26.-30. SAND DAM FIL TEA
9. SILT -BOSTON 18. LOESS
30 Sep 86
Figure 2-11. 201
Relationship between void ratio and log of permeability (courtesy of Wiley )
EM 1110-2-1901
30 Sep 86
a. Effect of fines on permea- b. Effect of fines on permea-

bility of concrete sand bility of sand-gravel mixture
(from Barber and Sawyer137) (from Barber and Sawyer137)
c. Effect of fines on permea-

bility of uniform fine sand
(from Fenn171)
Figure 2-12. Influence of type and amount of fines on permeability

of concrete sand, sand-gravel mixture, and uniform fine sand
(prepared by WES)
2-22
EM 1110-2-1901
30 Sep 86
b. Masch and Denny Method. For uniform or nonuniform dense clean sands,
classified SP or SW in the Unified Soil Classification System (U. S. Army
Engineer Waterways Experiment Station 1960), the permeability may be estimated
from an empirical method developed by Masch and Denny 1966, and Denny 1965.
The gradation curve is plotted in Krumbein 4 units (Krumbein and Pettijohn
1938) (using the chart in figure 2-13) as shown in figure 2-14 where
(2-23)
where
= phi scale units used to describe grain size distribution
d = grain size diameter in mm
The inclusive standard deviation is used as a measure of the spread of the

gradation curve where (Masch and Denny 1966)
(2-24)
where
= inclusive standard deviation
d 1 6 = grain size in units at which 16 percent is finer
d5 = grain size in units at which 5 percent is finer
2-23
10 I I I I I I I I I I I I I I I I I I I I I
30 Sep 86
EM 1110-2-1901
8
6 ~
4
~ ~
~ ~'<:,_,.,
zd
~
2-24
0 ~
-2 ~ .........._
~
-4 1 I I I I I I I I I I I I I I I I I I I I
0.01 0.1 10
GRAIN SIZE, d, mm
Figure 2-13. Conversion chart for ~ and grain size in mm, for the range 0.01 to 10 mm
(prepared by WES)
EM 1110-2-1901
30 Sep 86
a. Grain size distribution
b. Coefficient of permeability
versus median grain size
Figure 2-14. Masch and Denny relationship for permeability

as a function of median grain size and inclusive standard
deviation (courtesy of Prentice-Hall175)
2-25
EM 1110-2-1901
30 Sep 86
The median grain size, d50 in units, is determined from the gradation
curve as shown in figure 2-14a. Then knowing and d50 , the coefficient
of permeability in cm per minute can be obtained from figure 2-14b (Freeze and
Cherry 1979).
c. Kozeny-Carman Equation. For uniform loose to dense clean sands

classified SP in the Unified Soil Classification System (U. S. Army Engineer
Waterways Experiment Station 1960), the permeability may be estimated using
the Kozeny-Carman equation (Loudon 1952 and Perloff and Baron 1976)
(2-25)
where
k = coefficient of permeability
= unit weight of fluid
e = void ratio
C s = shape factor corresponding to a particular flow channel
T o = tortuosity factor related to the degree of sinuous flow
sS = specific surface (surface area of solids/volume of solids)
µ = coefficient of viscosity of fluid
For sands and silt-sized (finer than 0.074 mm and coarser than 0.005 mm)
particles CsTo2 = 5 is a good approximation (Perloff and Baron 1976). The
specific surface may be obtained from (Loudon 1952)
(2-26)
where
s s = specific surface
A = angularity factor
X 1 = percentage of total soil sample between adjacent sieves expressed

as a decimal
sl = specific surface of spheres uniformly distributed in size between

the mesh sizes of adjacent sieves
2-26
EM 1110-2-1901
30 Sep 86
The angularity factor, A , which varies from 1.0 for glass spheres to 1.8 for
crushed glass, may be determined by microscopic examination of the soil or
estimated from table 2-4 (Loudon 1952). The specific surface of spheres,
Si , between the mesh sizes dx and dy is (Loudon 1952)
(2-27)
Specific surfaces of spheres lying between selected U. S. standard sieves is

given in table 2-5.
d. Correlation of In Situ Horizontal Permeability and Hazen's Effective

Grain Size. For natural fine to medium, relatively uniform sands, classified
SP or SW in the Unified Soil Classification System (U. S. Army Engineer
Waterways Experiment Station 1960), in the middle and lower Mississippi River
Valley, the in situ horizontal permeability may be estimated from the Hazen's
effective size as shown in figure 2-15 (U. S. Army Engineer Waterways Experi-
ment Station 1956a). The relationship given in figure 2-15 should not be used
outside the geographic area for which it was developed. A similar relationship
between transmissivity and median grain size of sands is available for the
Arkansas River Valley (Bedinger 1961).
Table 2-4. Angularity Factor for Soil Grains (a)
Type of Material Description Angularity Factor
Glass sphere Well rounded 1.0
Natural sand Rounded 1.1
Subrounded 1.2
Subangular 1.3
Angular 1.4
Crushed rock Quartzite 1.5
Crushed rock Basalt 1.6
Crushed glass Pyrex 1.8
(a) Courtesy of the Institution of Civil Engineering 210 .
2-27
EM 1110-2-1901
30 Sep 86
Table 2-5. Specific Surface of Spheres Lying Between

Selected U. S. Standard Sieve Sizes (a)
U. S. Standard Sieve Specific Surface (b)

Numbers 1/cm
4 to 6 382
6 to 8 538
8 to 10 696
10 to 16 985
16 to 20 1524
20 to 30 2178
30 to 40 3080
40 to 50 4318
50 to 70 6089
70 to 100 8574
100 to 140 12199
140 to 200 17400
(a) Prepared by WES.

(b)
e. Computation of Permeability from Consolidation Test. The coeffi-

cient of permeability of normally consolidated clays and silts can be computed
from the consolidation test using the relationship (Lambe 1951 and Olson and
Daniel 1979)
(2-28)
where
c V = coefficient of consolidation
a v = coefficient of compressibility
eo = initial void ratio
2-6. Laboratory Methods for Determining Permeability.
a. General. Laboratory tests described in EM 1110-2-1906 can be used

to determine the coefficient of permeability of a soil, Unless otherwise
required, the coefficient of permeability shall be determined using deaired
distilled water and completely saturated soil specimens. The apparatus used
2-28
EM 1110-2-1901
30 Sep 86
Figure 2-15. Relationship between in situ horizontal

120
permeability and effective size (prepared by WES )
for permeability testing may vary depending upon whether the sample is fine-
grained or coarse-grained, undisturbed, remolded, or compacted, and saturated
or unsaturated. The permeability of remolded coarse-grained soils is deter-
mined in permeameter cylinders, while the permeability of undisturbed coarse-
grained soils in a vertical direction can be determined using the sampling
tube as a permeameter. Samples which have become segregated or contaminated
with drilling mud during sampling operations will not give reliable results.
The permeability of remolded coarse-grained soils is generally used to
approximate the permeability of undisturbed coarse-grained soils in a hori-
zontal direction. Usually the laboratory permeability of remolded coarse-
grained soils is considerably less than the horizontal permeability of the
coarse-grained soil in the field, so the approximation may not be conserva-
tive. Pressure cylinders and consolidometers are used for fine-grained soils
2-29
EM 1110-2-1901
30 Sep 86
in the remolded or undisturbed state. Fine-grained soils can be tested with

the specimen oriented to obtain the permeability in either the vertical or
horizontal direction.
b. Possible Errors. There are several possible errors in determining

permeability in the laboratory (FM 1110-2-1906; Olson and Daniel 1979;
and Mitchell, Guzikowski, and Villet 1978).
(1) Use of samples that are not representative of actual field condi-
tions. This can be minimized by thorough field investigation, attention to
details (take undisturbed samples from test fills for determination of perme-
ability of embankment materials, sampling along faults, fissures, clay seams,
and sand partings for determination of permeability of the dam foundation),
and by the use of large samples.
(2) Orientation of the in situ stratum to the direction of seepage flow

is seldom duplicated in the laboratory. This can be overcome by obtaining the
permeability of the soil (embankment material and/or foundation) in both the
vertical and horizontal direction.
(3) Incorrect hydraulic gradient used in the laboratory test. The

hydraulic gradient used in the laboratory should cover the range of expected
hydraulic gradient in situ. Where possible the hydraulic gradient should be
selected so that the flow is laminar (straight line relationship between dis-
charge versus hydraulic gradient) and Darcy's law will be applicable. It is
usually not practical to achieve laminar flow for coarser soils, and the
laboratory test should be run at the hydraulic gradient anticipated in the
field.
(4) Air dissolved in the water. As water enters the specimen, small
quantities of air dissolved in the water will tend to collect as fine bubbles
at the soil-water interface and reduce the permeability with increasing time.
Permeability tests on saturated specimens should show no significant decrease
in permeability with time if properly deaired distilled water is used. How-
ever, if such a decrease in permeability occurs, then a prefilter, consisting
of a layer of the same material as the test specimen, should be used between
the deaired distilled water reservoir and the test specimen to remove the air
remaining in solution.
(5) Leakage along the sides of the permeameter can result in an

increased permeability. One major advantage of the triaxial compression
chamber for permeability tests is that the specimen is confined by a flexible
membrane which is pressed tightly against the specimen by the chamber pressure
thus reducing the possibility for leakage along the sides.
2-7. Origin, Occurrence, and Movement of Ground Water.
a. Hydrologic Cycle. Precipitation, runoff, storage, and evaporation

of the earth's water follow an unending sequence called the hydraulic cycle,
as shown in figure 2-16. Radiation from the sun evaporates water from the
oceans into the atmosphere. The moisture is condensed and rises to form cloud
formations. From these clouds, the earth receives precipitation of rain,
2-30
EM 1110-2-1901
30 Sep 86
Figure 2-16. Schematic diagram of the hydrologic cycle

189
(courtesy of Johnson Division, Universal Oil Products )
snow, sleet, or hail which runs into lakes and streams or seeps into the soil
and thence into the underlying rock formations. The percolating water moves
through the saturated subsurface materials and may reappear at the surface, at
a lower elevation than the level where it entered the ground, in the form of
springs and seeps which maintain the flow of streams in dry periods (TM 5-545;
Bureau of Reclamation 1977; and Johnson Division, Universal Oil Products
1972).
b. Water Table. The surface below which the soil or rock is saturated
is the water table, as shown in figure 2-17. The water table is not a level
surface but varies in shape and slope depending upon the variations in perme-
ability and areas of recharge and discharge. In general, the water table
reflects the surface topography but with less relief. Ground water is said to
be perched if it is separated from the main water table by unsaturated mate-
rials, as shown in figure 2-17. An aquifer is a saturated permeable geologic
unit that can transmit significant quantities of water under ordinary hydrau-
lic gradients. An unconfined aquifer is one that does not have a confining
layer overlying it as shown in figure 2-17. The water table, or upper surface
of the saturated ground water is in direct contact with the atmosphere through
the open pores of the overlying material and movement of the ground water is in
direct response to gravity. The aquifer may be a layer of gravel or sand,
permeable sedimentary rocks such as sandstones or limestones, a rubbly zone
between lava flows, or even a large body of massive rock, such as fractured
2-31
EM 1110-2-1901
30 Sep 86
..• o;
c
0
•
!:'
;;;
~
,,_ ___ _
- - - _ _ _cPiezometric sur foce
--- II
II II
II II
II I I
II II
2-32
1'1 I I
II II
II II
Wate~ II II
level II II
II
Unconfined oquifer
II
I I
I I
II II
Perforated
II Well cosin9 II
II
See text for explanations

and definitions.
Aquiclude
Figure 2-17. .
Groundwater relationsh1ps (courtesy of Soil Conservation Service 68 )
EM 1110-2-1901
30 Sep 86
granite, which has sizable openings. An aquiclude is a saturated geologic

unit that is incapable of transmitting significant quantities of water under
ordinary hydraulic gradients as shown in figure 2-17. A confined or artesian
aquifer has an overlying confining layer of lower permeability than the aqui-
fer and has only an indirect or distant connection with the atmosphere as.
shown in figure 2-17. The water in the artesian aquifer is under pressure and
if the aquifer is penetrated by a tightly cased well or a piezometer, the
water will rise above the bottom of the confining layer to an elevation at
which it is in balance with atmospheric pressure. If this elevation is
greater than that of the land surface at the well, artesian water will flow
from the well as shown in figure 2-17. The imaginary surface, conforming to
the elevations to which water will rise in wells penetrating an artesian
aquifer, is the piezometric surface as shown in figure 2-17 (Soil Conservation
Service 1978, TM 5-545, Bureau of Reclamation 1977, Freeze and Cherry 1979,
and Anonymous 1980).
2-8. Field Methods for Determining Permeability.
a. General. In sands it is difficult to obtain undisturbed soil

samples for laboratory testing and the structure (void ratio, stratification,
etc.) has an important influence on-permeability. Therefore, field tests for
determining permeability are necessary. Because sampling operations do not
necessarily indicate the relative perviousness of foundations containing large
amounts of gravelly materials, field pumping tests are required to determine
the foundation permeability for dams where positive measures are not proposed
to completely cut off underseepage in the gravelly formations.
b. Test Pits and Bore Hole Tests.. In sands and gravels above the
ground-water level, field tests are normally carried out by measuring the down-
ward seepage from test pits or shallow boreholes (Cedergren 1977). Below the
ground-water table information about the order of magnitude and variability of
the coefficient of permeability may be obtained by conducting falling head
permeability tests in the exploratory boring as drilling proceeds. The hole
is cased from the ground surface to the top of the zone to be tested and
extends without support for a suitable depth below the casing. If the per-
vious stratum is not too thick, the uncased hole is extended throughout the
full thickness, otherwise the uncased hole penetrates only a part of the
pervious stratum. Water is added to raise the water level in the casing and
then the water level descends toward its equilibrium position. The elevation
of the water level is measured as a function of time and the coefficient of
permeability is calculated (Terzaghi and Peck 1967).
(2-29)
where
2-33
EM 1110-2-1901
30 Sep 86
A = inside cross-sectional area of casing
drop in water level in casing during time interval At
time interval
C = dimensionless quantity depending on shape of cylindrical hole and

depth of penetration into pervious zone (see figure 2-18)
= mean radius of hole below casing
= mean distance during time interval At from water level in casing

to equilibrium water level in pervious zone
The falling head field permeability test often gives an observed permeability
that is too low because silt particles which are suspended in the water may
form a filter skin over the walls and bottom of the hole in the pervious mate-
rial. The results of such tests are little more than an indication of the
order of magnitude of the in situ permeability. More reliable data are
obtained from field pumping tests.
c. Field Pumping Tests. The most reliable method for determining in

situ permeability is a field pumping test on a test well which fully pene-
trates the aquifer. The test procedures for equilibrium (steady-state flow)
and nonequilibrium (transient flow) are given in Appendix III to TM 5-818-5
and Civil Works Engineer Letter 63-16 (U. S. Army Corps of Engineers 1963).
The ratio of the horizontal to vertical permeability can be determined from
specially conducted field pumping tests (Mansur and Dietrich 1965).
2-9. Chemical Composition of Ground Water and River (or Reservoir) Water.
a. Ground Water. The chemical composition of the ground water is impor-

tant because some ground waters are highly corrosive to metal screens, pipes,
and pumps, or may contain dissolved minerals or carbonates which form incrus-
tations in wells or filters and, with time, cause clogging and reduced effi-
ciency of the dewatering or drainage system. Indications of corrosive and
incrusting waters are given in table 2-6 (TM 5-818-5; and Johnson Division,
Universal Oil Products 1972). General information concerning ground-water
properties is available in an Atlas (Pettyjohn et al. 1979). Sampling, sample
preservation, and chemical analysis of ground water is covered in handbooks
(Moser and Huibregtse 1976, and Environmental Protection Agency 1976).
b. River (or Reservoir) Water. The total amount of cations (calcium,

magnesium, potassium, and sodium) in the river water (for dams not yet con-
structed) and in the reservoir water (for existing dams) significantly influ-
ences the erosion through a possible crack in the core of the dam (Perry 1975).
Usually, as the total amount of cations in the eroding water decreases, the
erodibility of the soil increases. For dams constructed of dispersive clay,
the susceptibility of the dam to piping depends, in part, upon the total amount
of cations in the seepage water (Perry 1979).
2-34
EM 1110-2-1901
30 Sep 86
(b)
L ..
For if; <20"-, vse L••L
5 m ~ M ~ ~ ~ ~ ~
Ratio~
ro
(a)
4000 I IT II I I I II -r If T1
2000
~
-- --
/
/
/
1000
700
=
~
/ = -
- / -
1..> ~00
'
-~
-"
,;::
2~0
-
v
v -
-...: I/
!00
"'
(3
70
~ / =
-
t-
1-- I/ -
40
1--
l/
v -
20
v
JO
7 v
v I I I II I I I l11 I I I II ~
I 2 ~ 7 /0 20 40 70 /00 200 ~ 700 1000 2000 4000
Rt~tio ~
ro
(c)
Figure 2-18. Field permeability test in bore hole

(courtesy of John Wiley and Sons175)
2-35
EM 1110-2-1901
30 Sep 86
Table 2-6. Indicators of Corrosive and Incrusting Waters (a)
Indicators of Corrosive Water Indicators of Incrusting Water

1. A pH less than 7 1. a pH greater than 7
2. Dissolved oxygen in excess of 2. Total iron (Fe) in. excess of

2 ppm( )b 2 ppm
3. Hydrogen sulfide (H2S) in excess of 3. Total manganese (Mn) in excess of

1 ppm, detected by a rotten egg 1 ppm in conjunction with a high
odor pH and the presence of oxygen
4. Total dissolved solids in excess of 4. Total carbonate hardness in

1000 ppm indicates an ability to excess of 300 ppm
conduct electric current great
enough to cause serious electro-
lytic corrosion
5. Carbon dioxide (C02) in excess of

50 pm
6. Chlorides (CL) in excess of 500 ppm
(a) From TM 5-818-51.

(b) ppm = parts per million.
2-36
EM 1110-2-1901
30 Sep 86
CHAPTER 3
DETERMINATION OF PERMEABILITY OF ROCK
3-1. Permeabilities of Rock Masses. Permeability of rock, as with soil, is a

measure of the ease-with which fluids may travel through a medium under the
influence of a driving force. The term "permeability," however, has several
definitions for describing the flow of water in rock masses.
a. Coefficient of Permeability. The engineer's coefficient, or Darcy's

coefficient, is normally referred to as simply the "coefficient of permeabil-
ity." It is defined as the discharge velocity through a unit area under a
unit hydraulic gradient and is dependent upon the properties of the medium, as
well as the viscosity and density of the fluid (Section 2-3.a.).
b. Intrinsic Permeability. The physicist's coefficient, or the intrin-

sic permeability, is occasionally used in the determination of the hydraulic
conductivity of a rock mass. It is defined as the volume of a fluid of unit
viscosity passing through a unit cross section of a medium in unit time under
the action of a unit pressure gradient (Section 2-3.b.). The intrinsic perme-
ability thus varies with the porosity of the medium and is independent of both
the viscosity and the density of the fluid.
c. Equivalent Permeability. The complex system of interconnected void

space in a rock mass may be described in terms of an equivalent porous con-
tinuum, and the flow assumed to occur uniformly throughout the mass rather
than within individual passageways. Under these assumptions the term equiva-
lent permeability is used to describe the permeability of a rock mass.
\
d. Parallel Plate Permeability. The permeability of a fissure or a
fissure set is occasionally determined by modeling the rock mass conditions
with parallel plates. The parallel plate permeability can be computed from
the value of the aperture between the plates (Ziegler 1976) and this perme-
ability provides inferences for use in the modeled rock mass. The accuracy of
computed parallel plate permeabilities has been verified consistently with
laboratory tests (Snow 1965).
e. Fissure Permeability. Fissure permability is the permeability of

an individual fissure or a set of fissures and, whether measured in the
laboratory or in the field, is determined using the parallel plate permeability
theory. Fissure permeability is determined using an equivalent parallel plate
aperture, rather than the actual fissure aperture. This in effect incor-
porates roughness into the parallel plate law.
3-2. Flow Characteristics in Rock Masses. The determination of the perme-

ability of a rock mass, whether it be a rock slope, dam foundation, or dam
abutment, can only be accomplished after certain criteria are defined or
specified.
a. Continuum Approximations. From an overall regional point of view,

most rock masses may be treated as continua, for all practical purposes. One
of the primary considerations in selecting the continuum approach,-for evalu-
ating the permeability of a rock mass, is the relative size, frequency, and
3-1
EM 1110-2-1901
30 Sep 86
orientation of the inherent discontinuities in the rock mass, compared with

the size of the area of interest or area under study. If the flow charac-
teristics of a rock mass are to be treated as those of continua, the net vari-
ation in flow over the study area should be relatively small, and the frequency
and orientation of the discontinuities should be such that they provide an
overall averaging effect on the flow with respect to the area of interest.
b. Discontinuum Approximations. As areas of investigation become

smaller and smaller, i.e. more specific, the inherent discontinuities in rock
masses start to play larger and larger roles in the interpretation of ground-
water flow and its path of movement. The permeabilities or relative perme-
abilities of individual fissures and fissure systems are important for esti-
mating the amount of seepage into various sections of underground excavations
such as powerhouse tunnels or diversions , or through specific strata in a dam
foundation or abutment. Discontinuum approximations of permeability are nor-
mally used when the flow in the area of investigation is governed either by a
single fissure or by a fissure system.
(1) Single Fissure Flow. The permeability of single fissures can be

very important in karstic terrains, basalt, or rhyolite flows, or in areas
where tunnels are driven through faulted zones. The flow through a single
fissure, under certain geologic conditions, can be the key to estimating the
seepage through a dam foundation or abutment, or for determining the best
route for a proposed tunnel.
(2) Flow Through Fissure Systems. Fissure systems such as joints,

fractures, and bedding planes can yield, or contribute to, unpredictable flow
paths, seepage patterns, and uplift pressures. The permeability of fissure
systems should be evaluated by a discontinuum method of analysis when the
size, frequency, and orientation of such systems make a continuum approach to
the area under investigation unrealistic.
c. Ground-water Velocity. Flow of water in rock masses is generally

considered to be governed by one of two laws. Under conditions of laminar
flow, Darcy's law, previously presented as equation 2-1 in Chapter 2 is assumed
to govern the flow. For turbulent flow conditions in rock there is a nonlinear
flow velocity versus hydraulic gradient relationship and equations presented by
Forchheimer (1914) and Missbach (1937) are assumed to govern the flow. The
Missbach law is the most convenient to analyze and it takes either of
two forms,
(3-1)
or
(3-2)
3-2
EM 1110-2-1901
30 Sep 86
where
m and = degrees of nonlinearity
k' and B' = turbulent coefficients of permeability
Since the two equations are identical, resolving them indicates that m =
and The use of the above laws, however, is usually restricted
to a homogeneous, isotropic, porous continuum. Since in soil and rock masses
there are complex systems of interconnected void spaces, an equivalent rather
than an absolute permeability should be determined. The coefficient of equiv-
alent permeability from the continuum approach assumes that flow occurs uni-
formly throughout the mass rather than within individual passageways. There-
fore, for equivalent permeability, Darcy's law and Missbach's law are written,
respectively,
(3-3)
and
(3-4)
where
ke = laminar equivalent permeability
turbulent equivalent permeability
The continuum approach, in some cases, is not applicable and therefore, the
discontinuum method of analysis for evaluating an equivalent permeability
should be used. Formulae for the discontinuum analysis for equivalent
permeability are presented later with various other methods of analysis.
3-3. Methods for Determining Rock Mass Permeability. Numerous methods have
been developed for determining or estimating rock mass permeabilities. All of
the available testing, as well as analytical techniques should be considered
and evaluated for each individual study, in order to optimize the advantages
and minimize the disadvantages inherent within each method for determining the
permeability of a rock mass.
a. Laboratory Permeability Tests. Laboratory permeability tests are

used for evaluating the permeability of rock cores or samples, determining the
flow characteristics of rock fissures, and performing parametric studies of
the factors affecting the permeability of rock masses. Laboratory test
methods for permeability provide a convenient research and evaluation tool
3-3
EM 1110-2-1901
30 Sep 86
because a variety of parameters may be controlled and varied to yield a broad

spectrum of conditions which may be encountered in a rock mass.
(1) Model Tests. Model tests are conducted by constructing parallel

plate models to simulate given geologic information. The tests generally tend
to be a parametric evaluation, but have been used extensively in evaluating
both theoretical and empirical rock mass permeability formulae. To model fis-
sures as equivalent parallel plate conductors, the flow between parallel
plates must be defined (Snow 1965 and Wilson and Witherspoon 1970). Laminar
flow of an incompressible viscous fluid between smooth, parallel plates can be
expressed as
(3-5)
where
= unit weight of water
= dynamic viscosity of water
d = aperture between smooth parallel plates
The volume flow rate per unit width, q , becomes
(3-6)
Comparison of the flow velocity equation with Darcy's law indicates that the
parallel plate permeability, kp , can be expressed by
(3-7)
where
(2) Individual Fissure Tests. The majority of the laboratory tests

which have been conducted, to date, have been on individual fissures. Tests
on individual fissures are perhaps the most flexible of the laboratory tests
for rock mass permeability. The tests are generally conducted at various flow
rates within the individual fissure and at various normal loads to simulate
the in situ effective stress flow conditions. The data are used to develop
3-4
EM 1110-2-1901
30 Sep 86
correlations for predicting the permeability under the anticipated in situ

stress and hydraulic gradient conditions. The permeability measured from
individual fissures provides direct input into discontinuum analyses and may
be applied to continuum analyses under certain conditions. Before fissure
permeability, kj , can be determined, however, the fissure roughness must be
incorporated into the equation for the parallel plate permeability, kp . To
accommodate the roughness, the aperture between smooth, parallel plates, d ,
is replaced with an equivalent parallel plate aperture, e . Thus, the fissure
permeability is expressed by
(3-8)
where Thus, it follows that the flow rate per unit width
becomes
(3-9)
The value of e can be determined from flow experiments by rearranging the

above equation to the form
(3-10)
Values for e are determined from laboratory tests on individual fissures,

and equations for the laminar equivalent permeability have been developed
yielding
(3-11)
where
e = average of individual values of e for fissures in the set under

avg consideration
b = average of the individual spacing between fissures

avg
3-5
EM 1110-2-1901
30 Sep 86
Tests of individual fissures can also be analyzed for turbulent flow according
to the Missbach law
(3-12)
where the volume flow rate per unit width can be expressed as
(3-13)
The turbulent coefficient of permeability thus given by
(3-14)
The degree of nonlinearity, m , is determined as the arithmetic slope of

log i versus log q (Ziegler 1976). For turbulent flow analysis the
equivalent parallel plate aperture, e , is estimated from analysis of the
linear portion of the q versus i curve given by
(3-15)
(3) Representative Sample Tests. Another laboratory approach to

measuring permeability is to test a representative sample from a rock mass.
The obvious difficulty with such tests, however, is the problem of obtaining a
representative specimen of reasonable dimensions. The tests may be conducted
as standard laboratory permeability tests, but on a larger scale. In addition
to the standard permeability tests, small-scale pressure injection tests may be
conducted on such specimens, but only a limited amount of success should be
expected.
(4) Evaluation of Methods of Analysis. Model tests, as well as the

testing of representative samples, have only limited application and the
results are frequently subject to much speculation. Measured fissure perme-
ability and computed equivalent permeability can be directly applied to dis-
continuum and continuum analyses, respectively. In general, laboratory tests
are normally representative of only a very small portion of the rock mass under
consideration. In addition, the results of such tests can be altered
significantly by either the control or the interpretation of the parameters
involved in the test. Laboratory tests, therefore, should be used as a sup-
plement to, rather than in lieu of, field tests.
3-6
EM 1110-2-1901
30 Sep 86
b. Interpretation from Geologic Properties. Theoretical and empirical

formulae have been developed which relate permeability to geologic properties.
The parameters generally required for interpretation, or computation, are the
average fissure aperture, the average surface roughness of the fissures, and
the average fissure spacing or number of fissures per given length. The
governing assumption for any interpretation by this method is that the rock
mass permeability is controlled by fissures and that the fissures may be
modeled as equivalent parallel plate conductors. Extensive borehole logging
or observation and mapping of exposed surfaces is required for determining the
parameters to be used in the analysis.
(1) Analytical Procedures. If no laboratory tests are performed to

determine an equivalent parallel plate aperture, an equivalent permeability can
be estimated solely from field data. The requirements are core samples for
determining surface roughness of fissures and borehole logging to determine the
fissure apertures and spacing. To evaluate the surface roughness of fissures,
Louis (1969) defined a surface roughness index, S , as
(3-16)
where
y = mean height of the asperities on the fissure walls
a = mean fissure aperture
For S < 0.033 equations for laminar equivalent permeability have been devel-
oped which yield
(3-17)
where a = average of the individual values of a . For S > 0.033 equa-

avg
tions for the laminar equivalent permeability have been developed which yield
(3-18)
where S = average of the individual values of S . The above equations

avg
generated by Louis (1969) are empirical and are the result of numerous pipe
flow experiments and separate tests of fissures with different roughness,
3-7
EM 1110-2-1901
30 Sep 86
modeled as openings between parallel slabs of concrete. In addition to laminar

flow formulae, equations may be developed for turbulent flow for both the
hydraulically smooth and the completely rough flow regimes.
(2) Evaluation of Method of Analysis. Comparisons of permeability

interpreted from geologic properties with permeability measured by other
methods have indicated the potential for large differences between the com-
puted and the actual permeability. Interpretations of permeability from fis-
sure properties are made difficult by obvious problems in measuring fissure
apertures and roughness.' Natural fissures can have complex surficial geom-
etries and can only be observed in boreholes or exposed surfaces, which may be
disturbed during excavation. Potentially large discrepancies are possible, due
to the possibility of many fissures near the borehole being either exaggerated,
constricted, or discontinuous, due to the disturbance during drilling. While
interpretation of permeability from geologic properties is theoretically pos-
sible, the results should be used with caution.
c. Field Measurements. By far the most accurate and most reliable

technique for determining the permeability of a rock mass is that of field
testing. The use of field tests results in larger volumes of the rock mass
being tested and the tests are performed under in situ conditions. Field
tests have generally been limited to ground-water velocity measurements, pump-
ing tests, and injection tests.
(1) Ground-water Velocity Measurements. The equivalent permeability can

be computed for a rock mass by measuring the ground-water velocity and the
hydraulic gradient when certain criteria are met. It must be assumed that
steady-state horizontal flow intersects the well and flow is governed by
Darcy's law. There are several techniques available for measuring ground-water
velocity downhole as discussed below.
(a) Temperature Probes. The velocity of ground water moving through a

borehole may be determined with the use of temperature probes or sensors. Such
devices consist of a small heater strip or coil mounted beneath a thermistor.
The amount of heat dissipated is a function of the ground-water velocity, and
properly calibrated, the devices can sense very low velocities (below
1 ft/min). The directional components of flow may be obtained by rotation of
the device or by the construction of orthogonal flow channels within the device
itself.
(b) Flowmeters. When conditions of high ground-water flow rates are

encountered, small horizontally mounted commercial flowmeters may be placed
downhole for measuring the velocity. The direction of flow may be determined
by varying the orientation of the flowmeter.
(c) Tracer Tests. Tracer tests involve the injection of an inert solu-
tion, or tracer, into an existing flow field via a borehole or well. Tracer
tests are often desirable because they are passive-type tests and do not place
unnatural stress conditions on the flow system. The dilution rate of the
tracer at the injection well or its time of travel to another well can be used
to calculate the ground-water velocity and ultimately the permeability. Detec-
tion of the tracer, or concentration measurements, can be made by either
3-8
EM 1110-2-1901
30 Sep 86
manual or probe sampling. Generally, the probe method of sampling is desirable

to avoid any disturbance to the flow system due to sample extraction. The
"travel time" tracer tests normally involve large portions of the rock mass,
and thus have the advantage of averaging the effects of exceptionally high- or
low-permeability zones within the mass. The dilution method of testing is
particularly applicable to determining permeability profiles within a single
borehole by injection of the tracer into borehole sections isolated by packers.
Commonly used tracers are radioisotopes, salt solution, and fluorescent dyes.
(d) Methods of Analysis. Downhole ground-water velocity measurement

devices such as a temperature probe or flowmeters measure the Darcy velocity or
discharge according to his original equation, expressed as
(3-19)
or
(3-20)
With the hydraulic gradient determined from observation wells in the area, the
permeability can be computed directly. For tracer tests the seepage velocity
is determined according to the equation
(3-21)
where
d W = distance between injection well and observation well
t r = tracer travel time between wells
Using the relationship
(3-22)
(3-23)
(3-24)
3-9
EM 1110-2-1901
30 Sep 86
In analyzing a tracer dilution test, the flow velocity, v , is related to the

rate at which the tracer concentration diminishes within the test section of
the injection well. For an assumed homogeneous isotropic porous medium, the
velocity is determined from the following equation given by Lewis, Kriz, and
Burgy (1966).
(3-25)
where
w d = well diameter
C r = ratio of the final to the initial tracer concentration
td = dilution time period
Analysis of dilution tests in fractured and fissured rock masses is made by

applying the parallel plate analogy.
(2) Pumping Tests. Pumping tests have become an established means of

determining the permeability of hydraulic characteristics of water bearing
materials. In a pumping test, water is pumped from a well normally at a
constant rate over a certain time period, and the drawdown of the water table
or piezometric head is measured in the well and in piezometers or observation
wells in the vicinity. Since pumping tests, as with tracer tests, involve
large volumes of the rock mass, they have the advantage of averaging the
effects of the inherent discontinuities, such as joints, fissures, fractures,
etc. Most classical solutions for pump test data are based on the assumptions
that the aquifers are homogeneous and isotropic, and that the flow is governed
by Darcy's law. Applications of such solutions to interpretation of pumping
tests in rock masses have resulted in varying degrees of success. For cases
where the normal solutions have proven to be unsuccessful or inadequate,
mathematical models have been developed which are capable of modeling the flow
regime in various types of rock masses. With pumping tests, the major dis-
advantage is the period of time required to perform a test. Test durations of
one week or longer are not unusual when attempting to approach steady-state
flow conditions. Additionally, large diameter boreholes or wells are required
since the majority of the conditions encountered require the use of a downhole
pump. The analysis of pumping test results obtained from rock masses is
generally completely analogous to the analyses used in classical soil
mechanics. Since such analyses are well-documented, they will not be pre-
sented here.
(3) Injection Tests. Injection tests, which are the reciprocal of

pumping tests, commonly involve the steady-state transmission of a fluid from a
borehole into the surrounding medium. The permeability of a rock mass can be
related to the relationship between the injection pressure and the flow rate.
Equations and techniques have been developed for both the steady-state and the
3-10
EM 1110-2-1901
30 Sep 86
unsteady-state conditions and for using either air or water for injection.
Methods of analysis of pressure injection tests are presented in Appendix C.
(a) Water Pressure Tests. Water pressure tests, also known as packer
tests (in Europe they are called Lugeon tests) are normally conducted by
pumping water at a constant pressure into a test section of a borehole and
measuring the flow rate. Borehole test sections are commonly sealed off by
one to four packers, with the use of one or two packers being the most widely
used technique. In comparison with a pumping test, a water pressure test
affects a relatively small volume of the surrounding medium, because frictional
losses in the immediate vicinity of the test section are normally extremely
large. The test, however, is rapid and simple to conduct, and by performing
tests within intervals along the entire length of a borehole, a permeability
profile can be obtained. Additionally, the water pressure test is normally
conducted in NX boreholes, and has the advantage of being conducted above or
below the ground-water table.
(b) Air Pressure Tests. Air pressure tests are similar to water pres-
sure tests except that air rather than water is used for the testing fluid.
The air pressure test was developed for testing above the ground-water table
and has predominantly been used for testing areas of high permeability such as
those characteristic of rubblelike, fallback material adjacent to explosively
excavated craters in rock. In such areas, water pressure tests have been
inadequate due to an inability to provide water at a flow rate high enough to
pressurize the surrounding media. Air pressure tests have an unlimited supply
of testing fluid, as well as the advantage of a wide variety of high capacity
air compressors. The disadvantage of such tests is that permeability equations
must be modified for application to a compressible fluid and a conversion from
the air permeability to a water permeability must be made to obtain usable
results.
(c) Pressure Holding Tests. Pressure holding or pressure drop tests are
usually conducted in conjunction with water pressure tests. The test is
analogous to the falling head test used in soil mechanics; however, in rock
the test section is normally pressurized to a value above that of the static
head of water between the test zone and the ground surface. The pressures are
normally measured in the test section with a transducer, and pressures versus
time are recorded. The pressure holding test offers the advantage of being
quick and simple to perform, as well as requiring significantly less water
than that used in conventional constant pressure tests.
3-4. Applications of Rock Mass Permeability.
a. Assessment of Ground-water Movements. The permeability of a dam's

foundation or abutments is one of the controlling factors in the movement of
ground water; therefore,. a valid assessment is imperative. The rock mass
permeability of the dam foundation or abutment has numerous applications, each
having potentially significant impact on the design or safety of the
structure.
(1) Seepage Patterns. Seepage patterns which are expected to develop

after impoundment of a reservoir should be evaluated during the design phase of
3-11
EM 1110-2-1901
30 Sep 86
a dam. Seepage patterns which do actually develop after a reservoir is in

operation provide valuable data for an evaluation of the as-built performance
of any cutoff or relief measures, sources and exits of seepage, and input for
planning required remedial measures, such as additional cutoffs, drains, etc.
In rock masses, generally, the overall trend of the seepage patterns can be
determined by the continuum approach; however, in some cases, a detailed
analysis of a specific seepage problem requires the use of a discontinuum
approach.
(2) Flow Rates. Given the permeability and gradient, the flow rate
across a given area can be computed. With the effective porosity, the rate of
advance of a seepage front can be determined. Ground-water flow rates are
required for determining construction dewatering requirements for both surface
and subsurface excavations, as well as the seepage losses through dam founda-
tions and abutments.
b. Foundation and Abutment Drainage Requirements. The permeability of

a dam's foundation and abutments is the major factor involved in evaluating
drainage requirements. Accurate and reliable rock mass permeability measure-
ments are required in the design phase for determining the necessity for, and
extent of, cutoff, vertical drains, drainage blankets, relief wells, etc.
Postconstruction problems with leakage, excessive uplift pressures, etc., also
require an in-depth evaluation of the flow of the water in a foundation or
abutment. Evaluations of drainage or relief requirements are generally
directed toward specific areas and, in most cases, while a continuum approach
can give satisfactory approximations, the discontinuum analysis is required
for describing the flow characteristics in details.
c. Grouting Requirements and Effectiveness. Consistent with determining

the drainage requirements for a structure, the permeability is also useful for
estimating the grouting requirements. In the design phase of a dam the
requirements for cutoff or grouting, and drainage or relief, compliment each
other and are balanced to obtain a desired seepage pattern and uplift pressure
distribution beneath or within the structure. In foundations, and particularly
in abutments, permeability measurements can indicate possible solution chan-
nels, faults, fissures, or other highly permeable zones which require grout-
ing. If permeability tests are conducted at a given location, both before and
after a grouting operation, an evaluation of the grouting effectiveness can be
made. Such tests should be conducted as a matter of routine since they will
either establish a confidence level or indicate the need for additional
grouting.
d. General Considerations. The application of rock mass permeability

to seepage control methods and evaluations has become increasingly useful as
more complex methods of analysis evolve.
(1) Index Tests. The determination of rock mass permeability for dam
foundations and abutments has historically been used more as an index test
than as an absolute test. As the state of the art advances this trend is
gradually changing. While index tests are a valuable tool to the experienced
foundation engineer, more complex methods of analysis are becoming available
3-12
EM 1110-2-1901
30 Sep 86
which allow a complete evaluation of the mechanisms involved as water flows

through a fractured or fissured rock mass.
(2) Continuum Analyses. The accuracy of continuum approximations of

permeability is dependent upon the geology and the size of the area under
consideration. In general, continuum analyses provide average permeabilities
or regional permeabilities in which the assumptions of isotropy and homogene-
ity can be used in a gross sense. From a practical point of view the continuum
approach has been and continues to be a useful tool for evaluating rock mass
permeability.
(3) Discontinuum Analyses. Recently the discontinuum approach to

determining permeability has become more and more promising. As techniques for
determining individual fissure permeabilities advance, the availability of
better modeling techniques increases, and understanding of the influences of
orientation, spacing, apertures, and surficial geometries of fissures and fis-
sure sets increase, the discontinuum approach to determining rock mass perme-
ability becomes more reliable.
3-13
EM 1110-2-1901
30 Sep 86
CHAPTER 4
SEEPAGE PRINCIPLES
4-1. General Considerations. Seepage as used in this manual is defined as

the flow of water through homogeneous saturated soil under steady-state condi-
tions. Additionally, the soil particles, soil structure, and water are
assumed incompressible and flow obeys Darcy's law. Thus transient conditions
such as a wetting front or other movement of water in unsaturated soil, con-
solidation, and subsidence are not considered for analysis. Principles which
characterize movement of energy through conducting media also apply to the
movement of water through soils. Seepage has been modeled for study by using
flow of electricity and heat. Both conditions are governed by Laplace's equa-
tion in homogeneous media. As explained in Chapter 2, water moves from a
higher energy state to a lower energy state, and in seepage the difference in
energy states is the amount of energy required to move the water through the
soil, i.e., to overcome the soil's resistance to the flow of water. Chapter 4
will consider factors controlling seepage, equations describing seepage,
methods of determining pressure distribution and pressures at particular points
in the soil, and seepage quantities. For example, in figure 4-1(a): What is
the uplift pressure at point 5? How much water will exit at point 8? How
fast? Will the sand at point 8 be eroded? If the sheet pile at point 6 is
removed, how will it affect pressure distribution beneath the dam?
4-2. Boundary Conditions.
a. Basis. The saturated soil which is considered for analysis must be

defined by boundaries, permeability of the soil, and heads imposed upon the
water. This section considers the types of boundaries which may define a
particular porous soil mass considered for analysis. The nature and location
of these boundaries are determined by a soils exploration program, assumptions
based on engineering judgment and conditions imposed by the proposed design.
Normally, simplifying assumptions are required in order to establish boundaries
which will make analysis feasible. Generally, seepage analysis problems
associated with dams will involve four possible types of boundaries (Harr
1962). Examples of the four general types of boundary conditions are shown in
figure 4-1.
b. Impervious Boundaries. The interface between the saturated, per-

vious soil mass and adjacent materials such as a very low permeability soil or
concrete is approximated as an impervious boundary. It is assumed that no flow
takes place across this interface, thus flow in the pervious soil next to the
impervious boundary is parallel to that boundary. In figure 4-1, lines AB and
1-8 are impervious boundaries.
c. Entrances and Exits. The lines defining the area where water enters
or leaves the pervious soil mass are known as entrances or exits, respec-
tively. Along these lines (O-l and 8-G in figure 4-1(a) and AD and BE in
figure 4-1(b)) are lines of equal potential; that is, the piezometric level is
the same all along the line regardless of its orientation or shape. Flow is
perpendicular to an entrance or exit. Entrances and exits are also called
reservoir boundaries (Harr 1962).
4-1
EM 1110-2-1901
30 Sep 86
Figure 4-1. Examples of boundary conditions (courtesy of

180
McGraw-Hill Book Company )
d. Surface of Seepage. The saturated pervious soil mass may have a

boundary exposed to the atmosphere and allow water to escape along this
boundary, line GE, figure 4-1(b). Pressure along this surface is atmospheric.
The surface of seepage may also be called a seepage face.
e. Line of Seepage. Known also as the free surface, this boundary is

located within the pervious soil where water is at atmospheric pressure,
line DG, figure 4-1(b). Because of capillary forces, the saturated zone of
pervious soil extends slightly above the line of seepage, but this capillary
zone rarely has significant influence on seepage analysis. Whereas the first
two boundaries are normally defined by the geometric boundaries of the satu-
rated porous soil mass, the line of seepage is not known until the flow dis-
tribution within the pervious soil is known. Again, as for an impervious
boundary, the assumption is made that no flow takes place across the line of
seepage, thus flow in the pervious soil next to this boundary is parallel to
the boundary.
4-3. Confined and Unconfined Flow Problems. Two general cases of seepage are
considered in this manual: confined and unconfined flow. Confined flow
exists in a saturated pervious soil mass which does not have a line of seepage
boundary. Figure 4-1(a) is an example of confined flow. Unconfined flow,
figure 4-1(b), exists when the pervious soil mass has a line of seepage. Thus
confined flow has all boundaries defined while for unconfined flow the surface
of seepage and line of seepage must be defined in the analysis.
4-2
EM 1110-2-1901
30 Sep 86
4-4. Laplace's Equation.
a. Seepage Analysis. In order to do a seepage analysis, a general

model describing the phenomena of seepage must be available. Supplied with
specific boundary conditions and soil properties, this model can be used to
determine head and flow distribution and seepage quantities. The Laplace
equation is the mathematical basis for several models or methods used in
seepage analysis.
b. Basis of Laplace's Equation. Figure 4-2 shows a general seepage

condition from which an element is taken. Development of Laplace's equation
depends on six assumptions:
(1) Heads h and h are constant and thus flow is steady state.
1 2
(2) Water is incompressible.
(3) Volume of voids does not change--soil is incompressible.
(4) Flow is laminar--Darcy's law applies.
(5) The element has a dimension, dy , into the plane of the figure
which gives an element volume but no flow takes place perpendicular to the
plane of the figure, i.e., the flow is two-dimensional.
(6) The saturated pervious soil stratum is homogeneous. From

figure 4-2(b) let:
v ,v = components of discharge velocity in x and z directions,

x z
respectively
hydraulic gradient in the x direction
hydraulic gradient in the z direction
(The minus sign indicates that gradient is in a direction opposite increasing

piezometric head.) Assumptions 1, 2, and 3 assure continuity of flow which
means that water entering the element per unit of time, q e (where qe = vx
dz dy + vz dx dy) equals water leaving the element per unit of time,
(where Setting
equal to gives:
4-3
EM 1110-2-1901
30 Sep 86
(4-1)
Using Darcy's law, v = ki and assuming the same permeability in the x and
z directions:
kh is called a potential or velocity potential and is normally given the

symbol Thus
and
Substituting into equation 4-1 gives
(4-2)
which is a form of the Laplace equation for laminar, two-dimensional flow in

homogeneous, isotropic, porous media. The development here follows Terzaghi
1943. Rigorous developments can be found in Bear 1972, Cedergren 1977, and
Harr 1962.
4-4
EM 1110-2-1901
30 Sep 86
Figure 4-2. Flow of water through saturated pervious soil

beneath a hydraulic structure (courtesy of John Wiley and
Sons274)
4-5. Methods for Solution of Laplace's Equation. Solutions to steady-state,

laminar flow, seepage problems must solve Laplace's equation. Several methods
have been developed to solve exactly or approximately Laplace's equation for
various cases of seepage, figure 4-3 (Radhakrishnan 1978). One of the most
widely used methods, the flow net, can be adapted to many of the underseepage
and through-seepage problems found in dams and other projects involving
hydraulic structures. This method will be covered in detail in Section 4.6.
a. Models. Models which scale or simulate the flow of water in porous

media can provide a good feel for what is occurring during seepage and allow a
physical feel for the reaction of the flow system to changes in head, design
geometry, and other assumptions. Appendix B contains examples of the various
model types.
(1) As previously mentioned, processes which involve movement of energy

due to differences in energy potential operate by the same principles as
4-5
EM 1110-2-1901
30 Sep 86
I SEEPAGE ANALYSIS I
I J J 1
FLOW NETS MODELS ANALYTICAL METHODS NUMERICAL AND
COMPUTER METHODS
a. ELECTRICAL ANALOGY a. MAPPING OR a. FINITE DIFFERENCES

(1) WET--AQUEOUS TRANSFORMATION (RELAXATION)
SOLUTION OR 1,2,3-DIMENSIONAL
GEL b. SEMI EMPIRICAL
2-D OR 3-D -METHOD OF b. FINITE ELEMENT
4-6
FRAGMENTS 1,2,3-DIMENSIONAL
(2) DRY 2-D
b. SAND c. CLOSED FORM
2-D OR 3-D
c. VISCOUS FLOW
2-0
66
Figure 4-3. Seepage analysis methods (from Radhakrishnan )
EM 1110-2-1901
30 Sep 86
movement of confined ground water. These processes include electricity and

heat flow which have been used as seepage analogies. Electrical analogies have
proven particularly useful in the study of three-dimensional problems and in
problems where geometric complexities do not allow adequate simplifying
assumptions for analytical methods. "Wet" electrical analogies normally use a
conducting aqueous solution or gel to model the volume of the confined, satu-
rated, porous soil. Wet models are well suited to projects where an irregular
structure penetrates a confined aquifer. By probing the gel or solution when
a set potential or voltage is applied across it, electrical potential can be
determined at various points of interest in the model aquifer (McAnear and
Trahan 1972, Banks 1963, 1965). When field conditions can be characterized by
a two-dimensional plan or section, conducting paper models may be used to
inexpensively determine the effect of various configurations on the flow and
pressures in the aquifer, figure 4-4 (Todd 1980).
(2) Sand models which may use prototype materials can provide informa-
tion about flow paths and head at particular points in the aquifer. The sand
or porous material may be placed underwater to provide a homogeneous condi-
tion, or layers of different sand sizes may be used to study effects of inter-
nal boundaries or layers. If the flow is unconfined and the same material is
used for model and prototype, the capillary rise will not be scaled and must
be compensated for in the model. Flow can be traced by dye injection and
heads determined by small piezometers. Disadvantages include effects of
layering when the porous material is placed, difficulty in modeling prototype
permeability and boundary effects. Prickett (1975) provides examples of sand
tank models and discusses applications, advantages, and disadvantages.
(3) Viscous flow models have been used to study transient flow (e.g.,
sudden drawdown) and effects of drains. This method depends on the flow of a
viscous fluid such as oil or glycerin between two parallel plates and is nor-
mally used to study two-dimensional flow. As with sand models, dye can be
used to trace flow lines.' Construction is normally complicated and operation
requires care since temperature and capillary forces affect the flow. Flow
must be laminar, which can be difficult to achieve at the boundaries or at
sharp changes in boundary geometry.
b. Analytical Methods.
(1) Harr (1962) explains the use of transformations and mapping to

transfer the geometry of a seepage problem from one complex plane to another.
In this manner, the geometry of a problem may be taken from a plane where the
solution is unknown to a plane where the solution is known. While this method
has been used to obtain solutions to general problems it is not frequently
used for solutions to site-specific seepage problems since it requires the use
of complex variable theory and proper choice of transformation functions.
(2) Pavlovsky (1936, 1956) developed an approximate method which allows

the piecing together of flow net fragments to develop a flow net for the total
seepage problem. This method, termed the Method of Fragments, allows rather
complicated seepage problems to be resolved by breaking them into parts,
analyzing flow patterns for each, and reassembling the parts to provide an
4-7
EM 1110-2-1901
30 Sep 86
Figure 4-4. Use of two-dimensional conducting paper to find flow

lines and equipotential lines (courtesy of John Wiley
279)
and Sons
overall solution. Appendix B contains details of the Methods of Fragments

based on Harr's (1962) explanation of Pavlovsky's work.
(3) Closed form solutions exist for simpler seepage conditions such as
flow to a fully penetrating well with a radial source (Muskat 1946). Seepage
problems associated with dams typically require approximate solutions because
of complicated flow conditions.
4-8
EM 1110-2-1901
30 Sep 86
c. Numerical and Computer Methods. Computer models are used to make

acceptable approximations for the Laplace equation in complex flow conditions.
The two primary methods of numerical solution are finite difference and finite
element. Both can be used in one-, two-, or three-dimensional modeling.
Several computer programs for these methods are available within the Corps of
Engineers (Edris and Vanadit-Ellis 1982).
(1) The finite difference method solves the Laplace equations by

approximating them with a set of linear algebraic equations. The flow region
is divided into a discrete rectangular grid with nodal points which are
assigned values of head (known head values along fixed head boundaries or
points, estimated heads for nodal points that do not have initially known head
values). Using Darcy's law and the assumption that the head at a given node
is the average of the surrounding nodes, a set of N linear algebraic equa-
tions with N unknown values of head are developed (N equals number of
nodes). Simple grids with few nodes can be solved by hand. Normally, N is
large and relaxation methods involving iterations and the use of a computer
must be applied. Appendix B provides details of this method.
(2) The finite element method is a second way of numerical solution.

This method is also based on grid pattern (not necessarily rectangular) which
divides the flow region into discrete elements and provides N equations with
N unknowns. Material properties, such as permeability, are specified for
each element and boundary conditions (heads and flow rates) are set. A system
of equations is solved to compute heads at nodes and flows in the elements.
The finite element has several advantages over the finite difference method
for more complex seepage problems. These include (Radhakrishnan 1978):
(a) Complex geometry including sloping layers of material can be easily

accommodated.
(b) By varying the size of elements, zones where seepage gradients or

velocity are high can be accurately modeled.
(c) Pockets of material in a layer can be modeled.
4-6. Graphical Method for Flow Net Construction. Flow nets are one of the
most useful and accepted methods for solution of Laplace's equation (Casagrande
1937). If boundary conditions and geometry of a flow region are known and can
be displayed two dimensionally, a flow net can provide a strong visual sense
of what is happening (pressures and flow quantities) in the flow region.
Equation 4-2, paragraph 4.4, is an elliptical partial differential equation
whose solution can be represented by sets of orthogonal (intersecting at right
angles) curves. One set of curves represents flow paths of water through the
porous media while curves at right angles to the flow paths show the location
of points within the porous media that have the same piezometric head. The
former are called flow lines, the latter equipotential lines. The flow net is
a singular solution to a specific seepage condition, i.e., there is only one
family of curves that will solve the given geometry and boundary conditions.
This does not mean that a given problem will have only one flow net--we may
choose from the family of curves different sets of curves to define the
problem, figure 4-5. The relationship between the number of equipotential
4-9
EM 1110-2-1901
30 Sep 86
drops, Nd , and flow channels, Nf , does not change. A brief study of

figure 4-5 will provide a feel for where quantity of flow is greatest,
velocity highest, and gradient highest, i.e., in the area of the porous soil
nearest the sheet pile (flow channel 4, figure 4-5(a); flow channel 5,
figure 4-5(b)). This section draws upon several publications which give
a. Net drawn for four flow b. Net drawn for five flow
channels. channels.
Figure 4-5. Flow net for a sheetpile wall in a permeable foundation

(from U. S. Army Engineer District, Little Rock92)
detailed explanation of flow net derivation and drawing instructions

(Casagrande 1937; Cooley, Harsh, and Lewis 1972; Soil Conservation Service
1973; and Cedergren 1977). One of the best ways to develop an understanding of
seepage and flow nets is to study well-drawn flow nets found in these and other
references and to practice drawing them.
a. Assumptions for Flow Net Construction. In order to draw a flow net,

several basic properties of the seepage problem must be known or assumed:
(1) The geometry of the porous media must be known.
(2) The boundary conditions must be determined (see paragraph 4.2).
(3) The assumptions required to develop Laplace's equation must hold

(see paragraph 4.4b).
(4) The porous media must be homogeneous and isotropic (anisotropic

conditions are dealt with in paragraph 4.7).
4-10
EM 1110-2-1901
30 Sep 86
b. Guidelines for Flow Net Drawing. Once the section of porous media
and boundary conditions are determined, the flow net can be drawn following
general guidelines:
(1) Determine flow conditions at the boundaries:
(a) Flow will be along and parallel to impermeable boundaries lines BCD
and FG, figure 4-5.
(b) Entrances and exits are equipotential lines, lines AB and DE,
figure 4-5, with flow perpendicular to them.
(c) Flow will be along and parallel to a line of seepage--line AB,

figure 4-6.
Figure 4-6. Seepage through an embankment

underlain by an impermeable foundation
155)
(courtesy of John Wiley and Sons
(d) Entrance and exit conditions for a line of seepage are shown in
figure 4-7 under "Conditions for Point of Discharge."
This will provide a feel for the flow net.
(2) Equipotential and flow lines must meet at right angles and make
curvilinear squares. Usually, it is best to make either the number of flow
channels a whole number (if the number of flow channels is a whole number, the
number of equipotential drops will likely be fractional).
(3) Generally, a crude flow net should first be completed and adjust-
ments applied throughout the net rather than defining one portion since
refinement of a small portion tends to shift the whole net.
(4) The initial emphasis should be on getting intersections of flow

lines and equipotential lines at 90º, then shifting lines to form squares.
(5) If, in the finished flow net, either equipotential drops or flow
channels end up as a whole number plus a fractional line of squares (equi-
potential drop or flow channel), this should not be a problem but must be used
in any calculations based on the flow net. It is convenient to locate a par-
tial equipotential drop in an area of uniform squares since this will make
accurate estimation of the fraction easier.
4-11
EM 1110-7-1901
30 Sep 86
Figure 4-7. Entrance and discharge (exit) conditions for a

line of seepage (courtesy of New England Waterworks
151
Association )
(6) Use only enough flow lines and equipotential lines to bring out
flow net definition. If more information is needed in particular areas, the
squares may be subdivided into smaller squares for more detail of flow and
pressure distribution.
(7) As shown in figure 4-6, equipotential line intersections at a line

of seepage, line AB, and a surface of seepage or discharge face, line BC, are
controlled by elevation since pressure is atmospheric along these lines.
Along the discharge face BC, the equipotential lines and flow lines do not
form squares since the discharge face is not a flow line or an equipotential
line but a line at atmospheric pressure and changing elevation potential.
(8) Figures 4-7 and 4-8 provide some guidelines for entrances and exits
and particular areas within the flow region.
4-12
EM 1110-2-1901
30 Sep 86
For foundations, furthermost upstream and down-

stream flow lines and equipotential lines should
intersect at or near the center of the pervious
foundation.
The flow line and equipotantial line nearest an

angle shouId intersect on the bisector of the
angle.
Same as (b) except for an upstream toe on an

impervious foundation.
2:1 length ratios to establish shape of the

“square” in a pervious foundation at the toe of
an impervious fill.
2:1 length ratios used with angle bisectors to

shape flow around an imbedded 90-degree angle.
2:1 length ratios to establish flow directions

beneath a thin cutoff wall taken to the midpoint
of the pervious stratum.
Subdivide to check odd-shaped "squares”. Re-

sulting smaller odd-shaped "squares” should
have the general shape of the one subdivided.
Figure 4-8. Guides for flow net construction (from U. S. Department

123
of Agriculture )
4-13
EM 1110-2-1901
30 Sep 86
(9) It is helpful to lay out the boundaries which will contain the flow
net in ink and use a soft pencil and eraser to develop the flow net to final
form.
(10) Accuracy of squares may be checked by drawing diagonals for a

square or subdividing the square by sketching an additional flow line and
equipotential line orthogonal to it (ad infinitum). The diagonals should be
smooth curves intersecting at right angles. Also, if the intersection of
two flow lines and two equipotential lines is a square, a circle, tangent to
each of the sides, may be inscribed within the square.
(11) For calculation of seepage quantity only a crude flow net is

required. Accurate flow nets are required to determine pressure distribution.
4-7. Flow Net for Anisotropic Soil. Most naturally occurring soils and many
man-placed soils have greater horizontal permeabilities than vertical. This
affects the shape of a flow net since the flow net provides a solution to
Laplace's equation which is based on the assumption of an isotropic porous
media (paragraph 4.4b). To compensate for anisotropy, the dimensions of the
porous media are changed by the square root of the ratio of the two perme-
abilities. If kh is the horizontal permeability and kv is the vertical
permeability, then the horizontal dimensions of the porous media cross section
are changed by a ratio of , e.g., if the base of a dam is 300 feet,
then it would be changed by a factor of , or would be 300 feet times
The same ratio would be applied to all other horizontal dimensions
to produce a transformed section. Next the flow net is drawn on the trans-
formed section, as described in paragraph 4-6. Then the section, including the
flow net, is returned to the original (true section) which produces a nonsquare
flow net. Computations are made using the nonsquare flow net just as a square
flow net is used for isotropic conditions. This procedure is illustrated in
figures 4-9 and 4-10. In the same manner, dimensions in the vertical direction
could be changed by the factor , square or normal flow net drawn
on the transformed section, then returned to true section. Pore pressure
distribution and hydrostatic uplift may be taken from either section while
gradient and magnitude of seepage forces must be determined from the true
section.
4-8. Flow Net for Composite Sections. Commonly, projects requiring seepage
analysis involve different soils with different permeabilities, e.g., strati-
fied foundation materials and zoned dams. Certain rules apply to flow lines,
equipotential lines, and lines of seepage crossing internal boundaries between
soils of different permeabilities. Figure 4-11 illustrates the deflection of
flow lines and equipotential lines at interfaces. The essential principle is
that the more permeable soil allows the same amount of water to flow with less
restriction, thus drops in potential within the higher permeability soil will
be farther apart (i.e., less energy loss in the higher permeability soil for
the same length of flow as in the low permeability soil). It should be noted
in figure 4-11 that when flow goes from lower permeability soil to higher
permeability soil, the distance between flow lines decreases (flow channel gets
smaller) and the distance between equipotential drops increases, Figures 4-12
through 4-14 are examples of flow net construction for seepage through soils of
4-14
EM 1110-2-1901
30 Sep 86
a. True section for kh = kv and transformed section for anisotropic

conditions.
b. True section for flow net in a for kh = 4kv
c. True section for flow net in a for kh = 9kv
Figure 4-9. Flow nets constructed on transformed section and redrawn on true
155
section (courtesy of John Wiley and Sons )
4-15
--------r---------1--------------
30 Sep 86
EM 1110-2-1901
40
,., ~
io u Stole (fl!ct
H ~
.s=S
----- --- ----~1

Anisotropic foundulion
lch = 100 ky
luopcr vious hotnulnr y

fJAM
flow t honnel
Notes• I. llorl,onlol Ira dormolion
futtur -v'li~ik "VIiiu() '0.1
1
2. k = v'(ky)(lch
3. Nrtu,, · 6.6!j 10 =o.GGS
4-16
4. q ,tt'h(t~r!N~,J)=k'(36X0665)
=7j 9k
__ ______ -----
:}/ Noles• I. Toe or~o mu I be enlor!Jed
to ol.lhJin vul e or l\ L.
2.
Figure 4-10. Example of technique for developing a flow net for an anisotropic foundation (from
123
U. S. Department of Agriculture )
[~ = b I
Interface Inter foce

4-17
k2
d k.J tan
c = k2 tan
----Flow lines
---Equipotential lines
Figure 4-11. Deflections of flow lines at interfaces of soils having different

15 1
permeabilities (courtesy of New England Waterworks Association . )
EM 1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
a. Identify boundaries
b. Determine deflection of flow net at interface between soils with

different permeabilities.
c. Trial flow net
d. Final flow net
Figure 4-12. Flow net construction for a composite section

4-18
EM 1110-2-1901
30 Sep 86
a. Construct flow net assuming an impermeable foundation.
b. External equipotential lines into foundation without adjusting lines of net

in dam.
c. Adjust flow net until balanced.
Figure 4-13. Flow net construction for embankment on a foundation of lower

permeability (courtesy of John Wiley and Sons155)
4-19
EM 1110-2-1901
30 Sep 86
a. Construct flow net assuming impermeable embankment
b. Extend equipotentials up into dam locating initial position of

line of seepage
c. Adjust flow net to meet basic flow net requirements.
Figure 4-14. Flow net construction for an embankment on a foundation

155
of higher permeability (courtesy of John Wiley and Sons )
4-20
EM 1110-2-1901
30 Sep 86
differing permeabilities. In all cases, flow lines and equipotential lines

maintain continuity across the interface between the soils though direction
will change abruptly. Additionally, the number of flow channels must remain
constant throughout the flow net. For the two examples of embankments with
foundations of differing permeabilities, figures 4-13 and 4-14, the flow is
more or less parallel to the interface, and the more permeable zone will
dominate flow location and quantity. Because of this, the flow net can be
started by assuming all flow goes through the most permeable zone. Once this
flow net is drawn, it is extended into the lower permeability zone and refined
to meet the general flow net criteria of paragraph 4.6. Transferring a line of
seepage across the interface of soils of differing permeability, such as in a
zoned dam, is more involved than transferring of flow lines and equipotential
lines and will be described in Chapter 6.
4-9. Determination of Seepage Quantities, Escape Gradients, Seepage Forces,

and Uplift Pressures. A flow net is a picture of seepage conditions under
given geometry and boundary conditions. It explains how pressures are dis-
tributed and where flow is being directed. Coupled with the knowledge of head
imposed on and the permeability of the porous media, the flow net can supply
important information about stability and flow quantity in two-dimensional
idealization of the real situation.
a. Seepage Quantities. Each of the complete flow channels passes an

equal volume of water per unit of time, while partial channels carry a propor-
tional flow. Each of the complete potential drops between equipotential lines
is an equal portion of the total head, h , applied across the flow net with
partial drops having a proportionally smaller part. The number of flow chan-
nels, including any partial channel, is given the symbol Nf while the number
of equipotential drops, including any partial drops, is given the symbol Nd .
The ratio of Nf/Nd is called the shape factor, , which is a characteristic
of the given geometry and boundary conditions and permeability ratios (k1/k2,
kv/kh). Quantity of flow per unit length through the porous media can be
determined by using Darcy's law, q = kiA and the shape factor. Total flow is
the sum of the flows through each flow channel, i.e., q = where q
is the total flow and is the flow through each complete flow channel. In
figure 4-10, q = Since is the head loss between each
equipotential line (h = is the dimension of a flow net square:
and from the Darcy equation:
4-21
EM 1110-2-1901
30 Sep 86
where a is the area of the rectangle perpendicular to the flow direction.

If one side of the rectangle is one unit of length perpendicular of the plane
of the flow net, and the other dimension is , thus a = This
leads to:
Then:
which gives the quantity of seepage flow for each unit of thickness of porous
media perpendicular to the plane of the flow net. Figure 4-10(b) gives an
example of this calculation for anisotropic seepage conditions in a dam
foundation. The permeability, k' , used for anisotropic conditions,
k' = is derived by Casagrande (1937).
b. Escape and Critical Gradients. The escape or exit gradient, ie ,

is the rate of dissipation of head per unit of length in the area where
seepage is exiting the porous media. For confined flow, the area of concern is
usually along the uppermost flow line near the flow exit, e.g., at the
downstream edge of a concrete or other impermeable structure, figure 4-15.
Escape gradients for flow through embankments may also be studied by choosing
squares from the area of interest in the flow net (usually at or near the exit
face and downstream toe) and calculating gradients. If the gradient is too
great where seepage is exiting, soil particles may be removed from this area.
This phenomenon, called flotation, can cause piping (the removal of soil
particles by moving water) which can lead to undermining and loss of the
structure. The gradient at which flotation of particles begins is termed the
critical gradient, icr . Critical gradient is determined by the in-place
4-22
ttcad Water Pnuure. 1lonu 111•1•P.r liniN IJ,.unriJrv
lbollout ol w"h I
Ground Su~·-~..__ _ _ _ _ _ , _ __
'N.
CP . Ql
GJ (\)
-.
Ol
w
-- It
I ul ar
4-23
!TTTTr
·------·.-'
• Fu!l'l
-- ------
l~u
ol dropa
----
fhwd lou • Upllll preasur t
I desivnu!lon to point l~h • flo. drops) ( IB.B'- hd.lon) Y. E1cap1 gredlent 11 tocwoll end lram~·· ''" sill!
\!lj ___ (p_ s rI "
5,5 5,1 85 5
-----
5.6
---67 ---6.5 824
768
- ---
EM 1110-2-1901
7,6 7.2 72~L ___
---·--·-- --- -
__ ---
-F---e
E 0.0
.. fl. 2 661
30 Sep 86
------·~---
9.5 0.8 62 4
Figure 4-15. Base uplift and escape gradient for concrete drop spillway
. 1 ture 123)
(f rom U• S . Department o f A gr1cu
EM 1110-2-1901
30 Sep 86
unit weight of the soil and is the gradient at which upward drag forces on the
soil particles equal the submerged weight of the soil particles, figure 4-16.
The critical gradient is dependent on the specific gravity and density of the
soil particles and can be defined in terms of specific gravity of solids, Gs ,
void ratio, e , and porosity, n :
or, since e = n and

e
1-n n = 1 + e
If typical values of Gs , e , and n for sand are used in the above

equations, i c r will be approximately 1. Investigators have recommended
ranges for factor of safety for escape gradient, FSG from 1.5 and 15,
depending on knowledge of soil and possible seepage conditions. Generally,

factors of safety in the range of 4-5 (Harr 1962, 1977) or 2.5-3 (Cedergren
1977) have been proposed.
c. Heave. In some cases, movement of soil at the downstream seepage

exit may not occur as flotation followed by particle-by-particle movement. A
mass of soil may be lifted initially, followed by piping. This phenomenon is
called heave and occurs when the upward seepage force due to differential head
equals the overlying buoyant weight of soil. Heave occurs under conditions of
critical hydraulic gradient. For field conditions, the point at which minimum
differential head offsets the overlying buoyant weight must be determined by
judgment and calculations. Terzaghi and Peck (1967) have evaluated the factor
of safety with respect to heave for a row of sheet piles. Resistance to heave
4-24
EM 1110-2-1901
30 Sep 86
Figure 4-16. Definition of critical gradient (prepared by WES)
may be developed by placing very pervious material on the exit face, which will
allow free passage of water but add weight to the exit face and thus add
downward force. This very pervious material must meet filter criteria to
prevent loss of the underlying soil through the weighting material.
d. Seepage Forces. Forces imposed on soil particles by the drag of

water flowing between them must be considered when analyzing the stability of
slopes, embankments, and structures subject to pressures from earth masses.
These forces are called seepage forces. The magnitude of this force on a mass
of soil is determined by the difference in piezometric head on each side of
the soil mass, the weight of water, and the area perpendicular to flow. The
4-25
EM 1110-2-1901
30 Sep 86
seepage force acts in the same direction as flow, i.e. along flow lines. Con-
sider the seepage force on plane A-A in figure 4-16. Since flow is vertically
upward, the direction of seepage force is up, the difference in piezometric
head is h , and the area perpendicular to flow is A . The seepage force,
Fs , is the part of the upward forces due to differential head, h , or:
In terms of gradient and unit volume;
and using f as seepage force per unit volume:

S
Two methods of applying this force to use in stability analysis are described
by Cedergren (1977) and termed the gradient method and boundary pressure
method. EM 1110-2-1902 gives examples of embankment stability analyses
considering seepage forces. Additionally, the effect of buoyant forces on soil
mass stability must also be considered. The upward or buoyant force, Fb ,
causing reduction in effective stress on plane A-A, figure 4-15, is the
remainder of the upward forces on plane A-A:
e. Uplift Pressures. When seepage occurs beneath concrete or other

impermeable structures or strata, the underside of this impermeable barrier is
subject to a force which tends to lift the structure upward. The determina-
tion of this pressure or force is important in analyzing the stability of the
structure. An example of the analysis is given in figure 4-15. Summing of
the uplift pressures over the bottom area of the spillway will give the total
uplift force on the structure for a stability analysis. Harr's text (1962)
provides methods other than flow net construction to determine uplift
pressures.
4-26
EM 1110-2-1901
30 Sep 86
CHAPTER 5
CONFINED FLOW PROBLEMS
5-1. General Considerations. As explained in Chapter 4, confined flow exists

when the saturated pervious soil mass does not have a line of seepage boundary.
Impervious weirs or gravity dams on pervious soil or rock are typical projects
which have confined flow conditions. This chapter will consider two of these
cases. While examples of this chapter use flow nets, other methods for deter-
mining uplift and gradient such as the Method of Fragments may be used.
5-2. Gravity Dam on Pervious Foundation of Finite Depth. Figure 5-1, a copy
of figure 4-15, provides an example of an impervious structure on a pervious
foundation of finite depth with calculation of uplift and escape gradient.
Figure 4-12 illustrates a gravity dam on a composite pervious foundation of
finite depth and figure 4-10 shows an example of an impervious dam (though not
a gravity dam) on a finite anisotropic foundation. A cross section of a
classical gravity dam/weir, figure 5-2 indicates the effect of a partially
penetrating cutoff placed beneath the upstream portion of the dam. Comparing
figure 4-10(b) with figures 5-1 and 5-2 will show the reduction in gradient at
the downstream toe caused by embedment of the structure in the pervious
foundation. Embedment provides a longer upper flow line for a given structure
width and reduces the gradient at the downstream toe of the structure. Other
measures to reduce uplift and/or high gradients at the downstream toe include
cutoffs beneath the downstream portion of the dam and placement of drains
beneath the downstream portion of the dam.
5-3. Gravity Dam on Infinitely Deep Pervious Foundations. This case, illus-
trated by figure 5-3, is symmetrical (if there are no asymmetrical cutoffs or
drains beneath the dam) and at large distances becomes a series of half circle
arcs for flow lines and radial lines for equipotential lines. Since the up-
stream (entrance) and downstream (exit) boundaries extend to infinity and the
foundation depth is infinite , the flow quantity is infinite. Obviously the
analyst must decide the limits of the problem and calculate flow quantity
based on those limits. The same manipulations and effects of embedment, cut-
off, and drain described for the previous case will apply to this case.
5-1
lleod Woler
EM 1110-2-1901
30 Sep 86
_ _ _V..___ _ _ _ _ _ _ _~--- - -
Ground
Toil Wuler
lrurrervious
Uountlor y \
77.11111 I rrr I 11 TTirrTT~
5-2
13.8;
----
Porn! No. of drops Ueod loss : Uplift pressure
desiqnotio11 to point 11\h a No drops) ( 18. a'- hd loss) Y. Escape yr adient at torwall and tramverse sill:
0
(fl.) (ps.U .~
A "5,5 5.1 855
--·----
B 6 5.6 824
1---·-- ---·----
c 7 6.5 768
f----··
0
·------
l.8
---·-
l.2 724
-- --
E 8.8 8.2 G61
-·---- -------- ---- -- --
F 9.5 8.8 62 'l
··-
Figure 5-l. Base uplift and escape gradient for concrete drop spillway (from
123
U. S. Department of Agriculture )
EM 1110-2-1901
30 Sep 86
Figure 5-2. Gravity dam on pervious foundation of finite depth

181
(courtesy of McGraw-Hill Book Company )
Figure 5-3. Gravity dam on an infinitely deep pervious foundation

(prepared by WES)
5-3
EM 1110-2-1901
30 Sep 86
CHAPTER 6
UNCONFINED FLOW PROBLEMS
6-1. Introduction. This chapter will consider unconfined flow problems for
cases involving earth dams. Because of their ability to give a strong visual
sense of flow and pressure distribution, flow nets will be used to define seep-
age. Other methods, such as transformations (Harr 1962), electrical analogy
models, and numerical methods, can provide pressures and flows and be used to
develop flow nets. Unconfined flow problems require the solution of flow and
pressure distribution within the porous media and definition of the line of
seepage boundary (phreatic surface within the dam).
6-2. Homogeneous Earth Dam on Impervious Foundation. The simplest earth dam
configuration consists of a homogeneous, pervious embankment on an impervious
foundation. Though rarely encountered in engineered embankments, this case
will introduce general methods of defining flow in embankments.
a. Definition of Unknown Seepage Boundaries and Calculation of Flow per

Unit Length of Embankment, q . It is desired to define the flow and pressure
distribution within the embankment and total flow through the embankment. The
first step is determination of the upper flow line (which is the line of seep-
age boundary) and the length of the seepage exit face on the downstream slope
of the earth dam. This provides all necessary boundary conditions for flow
net construction and complete seepage definition. The two unknown boundaries,
BC and CD, figure 6-1, are a combination of an entrance condition, fig-
ure 4-7(c), BBl; part of a parabola, B1B2; a smooth transition between points
of tangency, B2C, and a straight line discharge face along the downstream
slope, CD. A parabola, shown by the dashed line, is the basic geometric member
used to define the location and extent of the two boundaries. Casagrande
(1937) provided the standard reference for flow through embankments while
others (Harr 1962, Cedergren 1977, and others) added to and refined the basic
methods. Figure 6-2 provides the nomenclature and formulas for drawing the
line of seepage and exit face and determining the quantity of seepage per unit
length of embankment, q . In a given problem, embankment geometry and head
water elevation provide values for h , m and which allow location of
points A and B and determination of distance, d , as shown in figure 6-2.
Figure 6-1. Line of seepage, BC, and seepage exit face, CD, for a
homogeneous earth dam on an impermeable foundation (prepared by
WES)
6-1
EM 1110-2-1901
30 Sep 86
I
/I
h
II
I I
r I
- ........
........ ('
.
........................
I I .......
tl
METHOD . EQUATIONS
I SCHAFFERNAK -
! a = _d- - j d2 - h2
I cos 0: cos2 0: ~
VAN ITERSON I
1 q • k .
a sm o: tan 0:
; L. CASAGRANDE
a= S -
o
J S 2 - -h2 -
o sin2 0:
FOR a< 60°, USE S = ..j d 2 + h2 . FOR 0

! 60° <a< 90°, USE MEASURED S = Ac +CD 0
i q = k a sin2 0:
~ KOZENY i
I a
Yo 1
=- =- [ VI d 2 + h
2
- d
·
l
I 0 2 2
I
I
I q = 2ka 0 = ky 0
A. CASAGRANDE I DETERMINE (a+ L\a) AS THE INTERSECTION OF

THE BASIC PARABOLA AND DAM SLOPE. THEN
DETERMINE L\a FROM C VALUE ON FIG. 6-5.
q = k a sin2 0:
OR
q = ky 0 = k ( J d 2 + h2 - d )
Figure 6-2. Determination of line of seepage and seepage exit

face for embankments on impervious foundations (adapted from
151
New England Waterworks Association )
6-2
EM 1110-2-1901
30 Sep 86
b. After this is done one of the four methods shown in figure 6-2 and
explained below can be used to determine the location of the exit face CD and
the line of seepage BC.
(1) < 30º Schaffernak-Van Iterson. The two formulas for this method
given in figure 6-2 assume gradient equals dy/dx and allow direct determi-
nation of a and q . Construction of basic parabola shown in figure 6-3 is
the first step in determining the upper line of seepage (Casagrande 1937).
From embankment geometry and headwater height, point A is located. d and y
are determined by scribing an arc, with radius DA through point E. Then the
point of vertical tangency of the basic parabola, F, is determined. Line AG,
parallel to the embankment base and horizontal axis of the parabola, is drawn
and divided into an equal number of segments (6 in the case in figure 6-3).
Line GF, the vertical tangent to the parabola, located at yo/2 from the
downstream toe of the embankment is divided into the same number of equal
segments as line AG. The points dividing line AG into segments are connected
with point F. The intersection of these lines with their counterpart lines
drawn from the points on line GF define the parabola. Thus the basic parabola,
dashed line A-F, is defined. The upstream portion of the line of seepage,
dotted line BH, is drawn by starting at point B perpendicular to the upstream
slope (since the upstream slope is an equipotential line and the line of
seepage is a flow line) and continuing downstream to make line BH tangent to
the basic parabola at point H which is selected based on judgment. This is an
entrance condition as shown in figure 4-7. The central portion of the line of
seepage is along the basic parabola while the downstream portion is a smooth
transition from the basic parabola to tangency with the downstream slope at
point C. Point C is located a distance a. from the downstream toe as
determined by the equation for Schaffernak-Van Iterson shown in figure 6-2.
With all seepage boundaries known and using the rules of Chapter 4, a flow net
may be constructed within the boundaries as shown in figure 6-4. This figure
points out the important basic flow net requirement that all equipotential
lines intercept the line of seepage and exit face at points with equal verti-
cal separation (in this case H/10 apart).
Figure 6-3. Construction of basic parabola (prepared by WES)
6-3
EM 1110-2-1901
30 Sep 86
Figure 6-4. Seepage through a permeable

embankment underlain by an impermeable
foundation (prepared by WES)
(2) < 90° L. Casagrande. The gradient assumption for this method is
i = dy/ds where s is the distance along the line of seepage, and allows
greater accuracy than Schaffernak-Van Iterson method for steeper downstream
slopes. Use of the equations in figure 6-2 and the same general procedures
used for the Schaffernak-Van Iterson method apply for up to 60º. For
60° < < 90°, since a and s are interdependent, the location of point C
(or distance a) must be estimated to determine the value of so
then distance a calculated. This procedure is repeated until there is
satisfactory agreement between the portion of the distance so as measured
and a as calculated. Thus the seepage boundaries are established allowing
flow net construction.
(3) = 180° Kozeny. For this special case Kozeny described a solution
adapted by Casagrande (1937). Figure 6-5 illustrates the nomenclature and
construction method for this case. Embankment geometry, h , and drain loca-
tion control construction of the basic parabola. For this case the seepage
face is the distance a0 and the correction is not used. Again with
boundary definition, the flow net can be drawn.
(4) 30° < < 18.0° A. Casagrande. After study of model experiments
and construction of flow nets for various , A. Casagrande (1937)
developed a curve, figure 6-6, which relates a to the ratio,
Construction of the basic parabola is the first step in this procedure. The
point, Co , as shown in figure 6-2, where the basic parabola intercepts the
downstream slope is determined and distance a + is measured. Knowing
C can be found in figure 6-6 and calculated. Information is then suf-
ficient to draw the line of seepage and discharge face, determine q , and con-
struct the flow net. Casagrande (1937) provides a procedure for the condition
of tailwater on the downstream slope:
For the comparatively rare case in which the pres-

ence of tailwater must be considered in the design, the
determination of the line of seepage and of the quantity
can be performed by dividing the dam horizontally at tail-
water level into an upper and lower section. The line of
seepage is determined for the upper section in the same
6-4
EM 1110-2-1901
30 Sep 86
Figure 6-5. Construction of basic parabola and seepage line for

= 180° Kozeny (courtesy of New England Waterways
Association151)
Figure 6-6. c vs (courtesy of New England

151
Waterworks Association )
6-5
EM 1110-2-1901
30 Sep 86
manner as if the dividing line were an impervious bound-

ary. The seepage through the lower section is determined
by means of Darcy's law, using the ratio of the difference
in head over the average length of path of percolation as
the hydraulic gradient. The total quantity of seepage is
the sum of the quantities flowing through the upper sec-
tion and the lower section. The results obtained by this
rather crude approximation agree remarkably well with the
values obtained from an accurate graphical solution.
Harr (1962) explains an additional method, known as Pavlovsky's solution, for

determining a0 and q for the case of a homogeneous, pervious embankment on
an impervious foundation. Pavlovsky analyzed the embankment by dividing it
into three zones, writing an equation for q in each of the zones and, by
assuming continuity of flow, equating the three equations for q . Fig-
ure 6-7 provides the nomenclature for Pavlovsky's solution. The embankment is
divided as shown with Zone I between the upstream slope and a vertical line at
Figure 6-7. Nomenclature for Pavlovsky's solution

180
the intersection of the crest and upstream slope (y axis), Zone II between the
y axis and a vertical line at the intersection of the line of seepage with the
downstream slope, and Zone III which is composed of the remainder of the
downstream toe. Pavlovsky assumed horizontal flow in each zone and wrote the
basic equation q for each zone using the nomenclature of figure 6-7.
The equations for each zone are:
Zone I
(6-1)
6-6
EM 1110-2-1901
30 Sep 86
Zone II
(6-2)
Zone III
for ho > 0
(6-3)
for ho = 0
(6-4)
It is assumed that a , , b , hd , hw , hO , and k are known for a

given problem, thus since qI = qII = qIII = q (continuity of flow, steady
state conditions) only a0 , h1 , and q are unknown. This analysis pro-
vides three equations, (6-1), (6-2), and (6-3), or (6-4), and three unknowns.
The equations may be solved in a number of ways. One method for ho= 0 is to
equate (6-1) and (6-4) and solve for a then equate (6-2) and (6-4) and solve
for a0 :
(6-5)
(6-6)
then a plot of ao versus h1 may be made of equations (6-5) and (6-6). The
intersection of the two curves representing (6-5) and (6-6) is the value of a0
and h 1 for solution. Equation (6-4) will then provide q . An example from
Harr (1962) is provided in figure 6-8.
6-7
EM 1110-2-1901
30 Sep 86
Example: Determine the quantity of seepage, q , ao , and h1 , using

Pavlovsky's solution for the dam shown above. Compare with
Schaffernak-Van Iterson and L. Casagrande Method, figure 6-2.
A. Pavlovsky's solution
For assumed values of h1 in equations (6-5) and (6-6) the resulting

values for ao are given in the following tabulation and plotted in
the accompanying graph.
Equation (6-5) Equation (6-6)

h1 ao hl ao
50 19.6 50 15.9
52.5 18.7 52.5 17.7
55 17.4 55 19.7
60 13.9 60 24.1
65 8.4 65 29.3
Figure 6-8. Example problem of homogeneous embankment on

an impervious foundation comparing solutions (Continued)
(prepared by WES)
6-8
EM 1110-2-1901
30 Sep 86
From equation (6-4)
0.002 ft/min 18.3 ft

q= = 0.0122 ft3/min per ft of embankment length
3
a = 3ao = 54.9 ft
B. Schaffernak-Van Iterson solution:
since m = 3(70 ft) = 210 ft
d = 0.3(210 ft) + 3(10 ft) + 20 ft + 3(80 ft) = 353 ft
h W = 70 ft
a = 73 ft
q = (0.002 ft/min)(73 ft)(sin 18°26')(tan 18°26')

3
q = 0.015 ft /min per ft of embankment length
C. L. Casagrande solution:
a = 76 ft
2
q = (0.002 ft/min)(76 ft sin 18º26')
3
q = 0.015 ft /min per ft of embankment length
Figure 6-8 (Concluded)
6-9
EM 1110-2-1901
30 Sep 86
6-3. Earth Dam with Horizontal Drain on Impervious Foundation. Figure 6-5
presents this case for a homogeneous embankment. However, since most earth
dams are built in horizontal layers, they very likely have a stratified
structure which may allow considerable flow to bypass the horizontal drain,
figure 6-9. The difference in vertical and horizontal structure also causes
differences in horizontal and vertical permeabilities. This can strongly
affect the location of the upper line of seepage, figure 6-10, which affects
stability considerations and methods of controlling seepage.
Figure 6-9. Schematic of effect of earth dam stratification on flow of

water through embankment with horizontal drain (prepared by WES)
6-4. Earth Dam with Toe Drain on Impervious Foundation. Toe drains are
another method of controlling the line of seepage, figure 6-11. Again the
effects of anisotropy must be considered, figure 6-12. The geometry of the
embankment and the toe drain, height of reservoir, and the degree of anisotropy
will control the location of the line of seepage.
6-5. Earth Dam with Vertical or near Vertical Horizontal Drains on Impervious
Foundation. One very effective method of intercepting horizontal flows due to
stratification of the embankment, figure 6-9, is the incorporation of an
inclined or vertical drain into the central portion of the embankment, fig-
ure 6-13. This seepage analysis of a zoned, anisotropic embankment assumed
the rockfill to have infinite permeability with respect to the core materials
and used the method recommended by A. Casagrande for drawing the parabola to
determine the upper line of seepage. The interface of the core and inclined
drain is used as the downstream slope for the seepage face since the drain has
a much higher permeability than the core material. Provision must be made in
sizing the drain to pass all the water coming out of the core without building
up a tailwater on the downstream slope of the core. It can be noted that the
designers of this example used m/3 instead of 0.3m to determine the intercep-
tion point of the parabola and headwater elevation and the formula
which can be derived trigonometrically, to determine

The geometry of the embankment and toe drain, height of reservoir,
and the degree of anisotropy will control the location of the line of seepage.
6-6. Flow Net for a Composite Zoned Dam. If differences in permeabilities

between zones in a zoned dam are great enough (e.g. 100 to 1000 times or more)
the more permeable zone may be considered to have infinite permeability rela-
tive to the less permeable zone for purposes of seepage analysis. In the
seepage analysis example of figure 6-13 the upstream rockfill and the inclined
6-10
EM 1110-2-1901
30 Sep 86
Figure 6-10. Effect of difference in horizontal and vertical permeability

(anisotropy) on location of line of seepage within an embankment with a
horizontal drain (after U. S. Department of Agriculture123)
Figure 6-11. Homogeneous embankment on impervious foundation with a

toe drain (from EM 1110-2-191311)
Figure 6-12. Effect of anisotropy for homogeneous embankment

on impervious foundation with a toe drain (courtesy of New
151
England Waterworks Association )
6-11
30 Sep 86
EM 1110-2-1901
£DAM
I
6-12
I
~CURTAIN GROUTING
I
TRANSFORMATION
EMBANKMENT SEEPAGE
KH ~ 4Kv, TRANSFORMATION r ACTOR--~ ='f'[ = _!.""' 0.5
v~ v; 2
TRUE SECTION
FOR NON-ISOTROPIC SOIL THE EFFECTIVE COEFFICIENT OF PERMEABILITY
TRANFORMEO SECTION HORIZONTAL DIMENSION= 0.5 TRUE SECTION HORIZONTAL DIMENSION
K' = KvKH, SINCE THE EMBANKMENT IMPERVIOUS Fill MATERIAL IS
DETERMINATION OF PHREATIC' LINE NON ISOTROPIC, K' = ~ ~ ~(04x10 6) (1.6)(10-61 CM S!:C

POOL OF SPILLWAY CREST E L 937', TAILWATER EL 806 1, MAXIMUM NET HEAD" 1311
K"- '(0"64 )( 10· 6 CM•SEC = 0.8)( 10 6 CM'SEC OR 26.2~)( lo-9rT SfC
ASSUMING TH!:. I::MEJANKMENT RESTS ON IMPERVIOUS FO\INOATIQN AS IN THE CASE
OF THE LEFT ABUTMENT, THE POSITION OF THE PHREATIC LINt: IS DETERMINED AFTER SEFPAGE OUANTITY, q FOOT,=~ K H ~~)( 26 25)(1Q-9H: 11 4A)( 1Q-9H FTJ.SEC
A CASAGRANDE THE ABOVE ASSUMPliON PERTAINS TO THE DETERMINATION OF
THE PHREATIC LINE ONLY WHICH LEADS TO MORE CONSERVATIVE DESIGN. WHENH- 131FT, q/FOOT-1148)(10-9)( 131"' 1504 )( 10 7 FT3·SEC
Y0 ~ fH~+dJ_-d"' f( 156.96)2+!13l.OUI] -156 96 ~ 47.48'
a•.la- ~~~DS<L. TANa-2,tF2.a=63°-26",C05a-0.447
<~+..'ia = ~ 5~~- 85.86 a~:a

7
-= 0 321AFTEA A. CASAGRANDE!
..'IJ- 0 32 )( 85.86= 27. 48"
Figure 6-13. Seepage analysis for Kettle Creek earth embankment, Kettle Creek, Pa. (from
78
U. S. Army Engineer District, Baltimore )
£DAM
F/3•36.~ I'
MAX WATER SURFACE EL 962.7 "
SPILLWAY CREST EL 937.0
I
:I:
:TRANSFORMED SECTION
6-13
FLOW NET
LEFT ABUTMENT-STATION 17+00
TRANSFORMATION EMBANKMENT SEEPAGE
fK; f1 1 FOR NON-ISOTROPIC SOIL THE EFFECTIVE COEFFICIENT OF PERMEABILITY

KH • 4Kv. TRANSFORMATION FACTOR "i""i(,:; "j"4 • "2 • 0.5
K' • KvKH, SINCE THE EMBANKMENT IMPERVIOUS FILL MATERIAL IS
TRANFORMED SECTION HORIZONTAL DIMENSION • 0.5 TRUE SECTION HORIZONTAL DIMENSION
NON ISOTROPIC, K' • oJKv KH • V10.4x1o-s) (1.6x1o-6) CM/SEC
DETERMINATION OF PHREATIC LINE
K' •I/Q.64 x 10- 6 CM/SEC • O.Bx1o-6 CM/SEC OR 26.25x1o-9FT/SEC
POOL OF SPILLWAY CREST EL 937 1, TAl LWATER EL 806 1, MAXIMUM NET HEAD • 131 1
'N, 7
ASSUMING THE EMBANKMENT RESTS ON IMPERVIOUS FOUNDATION, AS IN THE CASE SEEPAGE OUANTITY,q/FOOT,•N K'H•l6 x 26.25x1o-9H•11.48x10-9H FT3/SEC
OF THE LEFT ABUTMENT, THE POSITION OF THE PHREATIC LINE IS DETERMINED AFTER p
WHEN H • 131FT. q/FOOT • 11.4Bx1o-9x 131 • 15.04 x 10-7 FT3/SEC
A. CASAGRANDE. THE ABOVE ASSUMPTION PERTAINS TO THE DETERMINATION OF
THE PHREATIC LINE ONLY WHICH LEADS TO MORE CONSERVATIVE DESIGN.
y0 • iH 2 +d 2 -d. il156.96)2+( 131.00)2 -156.96. 47.48'
EM 1110-2-1901
yo
a+Jl.a = -COS a , TAN a • 2, a=2, a=63' -26', COS a • 0.447
1
47.48 Jl.a
30 Sep 86
a+Jl.a • _ = 85.86; a+Jl.a • 0.32 (AFTER A. CASAGRANDE)
0 553
l>a = 0.32 x 85.86 • 21. 48'
Figure 6-13. Concluded

EM 1110-2-1901
30 Sep 86
drain are considered infinitely permeable with respect to the central earth
portion of the embankment. In some cases dams may contain adjoining zones
which, have relatively small but marked differences in permeability and it may
be desired to accurately analyze the flow through these zones. Equipotential
lines and flow lines, with the exception of the upper line of seepage, will
cross the interface of the zones in the same manner as given for confined
composite sections in Chapter 4, but the location of the upper line of seepage
(phreatic line) must first be determined. Once its location is determined the
upper line of seepage transfers between regions of different permeabilities in
the manner shown in figure 6-14. In determining the location of the upper
seepage line, the essential principle to remember is that the flow rate must
be the same through each zone. That is, for a unit depth of embankment per-
pendicular to the plane of the flow net, Q must be the same for each zone.
More permeable zones require less gradient and/or cross-sectional area to pass
the flow transmitted to them from less permeable zones. This idea can be seen
in the example and instructions taken from Cedergren 1977.
Figure 6-14. Transfer of line of seepage between regions of

different permeability (courtesy of New England Waterworks
151
Association )
a. Locate the reservoir level and the tailwater level, noting the
difference in head as h and dividing h into a convenient number of equal
parts of increments Draw a series of light horizontal guide lines (head
lines) at intervals of across the downstream part of the section
(figure 6-15a).
6-14
EM 1110-2-1901
30 Sep 86
b. Guess a trial position for the phreatic line in both zones and draw a
preliminary flow net as shown in figure 6-15a, making squares in zone 1 and
rectangles in zone 2. Make the length-to-width ratios of all of the rectangles
in zone 2 approximately equal by adjusting the shape of the saturation line,
using an engineer's scale to measure the lengths and widths of the figures.
When this step is completed, the trial flow net should be reasonably well
drawn. It should satisfy the basic shape requirements of a flow net, but the
length-to-width ratio of the shapes in zone 2 probably will not satisfy
c/d = k2/k1 . Although the flow net has been drawn for a composite section,
the ratio of k2/k1 probably does not equal the k2/k1 ratio originally
assumed for the section.
a. First trial-flow net (not correct)
b. Completed flow net (correct)
Figure 6-15. Method for constructing flow nets for com-

posite sections (courtesy of John Wiley and Sons 155 )
c. Calculate the actual ratio of k2/k1 for the trial flow net just
constructed. To make this important check proceed as follows:
(1) Count the number of full flow channels between any two adjacent
equipotential lines in zone 1 and call this number nf-1. In the trial flow
net in figure 6-15a , n f - 1 = 4.0 .
(2) Count the number of full flow channels between any two adjacent
equipotential lines in zone 2 and call this number nf-2 . In figure 6-15a,
nf-2 is equal to the width-to-length ratio of the figures in zone 2, d/c ,
and equal to 0.5.
(3) The actual value of k2/k1 for the trial flow net in figure 6-15a
can now be determined from the equation
6-15
EM 1110-2-1901
30 Sep 86
or
(4) If the calculated k2/k1 ratio is too high, the saturation line in
zone 2 is too low and must be raised. If the calculated k2/k1 ratio is too
low, the saturation line in zone 2 is too high and must be lowered. Raise or
lower the general level of the saturation line in zone 2 as indicated and
construct another trial flow net.
(5) Repeat steps (1) through (4) until a flow net of the desired
accuracy is obtained. (Usually a few trials will be sufficient.) By applying
the above equation to the first trial flow net in this example (figure 6-15a)
k2/k1 = 4.0/0.5 = 8.0. Because the ratio of k2/k1 for this example was
assumed to be 5, the k2k1 ratio of the trial flow net is too high; hence the
general level of the saturation line in zone 2 is too low and must be raised.
For the second trial flow net (figure 6-15b) nf-1 = 3.5 and nf-2 = 0.7 .
The calculated ratio of k2/k1 = 3.5/0.7 = 5.0 , the value originally assumed.
The above equation may be derived by recalling that the quantity of seepage in
zones 1 and 2 (figure 6-15) must be equal. Using
composite flow net being examined for accuracy. It is an essential criterion

to be used in constructing accurate flow nets for composite sections. Fig-
ure 6-16 illustrates the additional detail which can be obtained by further
subdivision of the flow net, This may be necessary for a stability analysis or
other reasons. Another example provided in figures 6-17 and 6-18 illustrates a
flow net for a zoned dam on an impervious foundation; figure 6-17 is the trans-
formed section for the true section of figure 6-18 (Cedergren 1975 and U. S.
Army Engineer District, Sacramento 1977).
6-16
EM 1110-2-1901
30 Sep 86
Figure 6-16. Three forms of one flow net (courtesy of

155)
John Wiley and Sons
6-7. Zoned Earth Dam on Pervious Foundation.
a. Zoned embankments on pervious foundations may require rather involved

seepage analyses unless simplifying assumptions are made. This is particularly
true when relatively small permeability differences exist between adjacent
zones and foundation and it is desired to consider these differences in perme-
ability. Figure 6-19 provides an example of such an analysis (U. S. Army Engi-
neer Waterways Experiment Station 1956b). Development of the flow net required
use of a number of principles and methods explained in Chapter 4, and previous
portions of Chapter 6, e.g., dimensional changes due to anisotropy, upper seep-
age lines and other flow lines crossing the interface of materials having dif-
ferent permeabilities and composite sections. It should be noted that several
assumptions were made.
(1) The flow net for the foundation was drawn independently of the
embankment, i.e. considering the embankment to be impermeable, and assuming the
foundation to be isotropic.
(2) The embankment flow net reflects the influence of foundation flow
net in the location of equipotential and flow lines and is drawn for aniso-
tropic conditions.
6-17
EM 1110-2-1901
30 Sep 86
Figure 6-17. Buchanan Dam flow net studies full reservoir-

Sta 9+50 D/S transition fill - kh = 16 kv transformed
section (from Cedergren25 and U. S. Army Engineer District,
Sacramento 1977)
6-18
EL 602
6-19
k-->oo
BUCHANAN DAM
H. R. CEDERGREN
FLOW NET STUOIES
3-18-75 FULL RESERVOIR - STA 9+50
DIS TRANSITION FILL- kh = 16kv
EM 1110-2-1901
Figure 6-18. Buchanan Dam flow net studies full reservoir - Sta 9+50 D/S
25
30 Sep 86
transition fill - ~ = 16 kv true section (from Cedergren and U. S.
Army Engineer District, Sacramento 1977)·
30 Sep 86
EM 1110-2-1901
SEEPAGE PER FOOT OF EMBANKMENT LENGTH
0,. 1< 1 h S,. 7.5 X 10-8 CM/SEC X 192FT X 15/8 X 60 SEC/MIN X 1/30.48 FT/CM X 7.48 GAL/FT 3
0 a 0.00040 GAL/MIN PER FOOT OF EMBANKMENT
FOR 600FT EMBANKMENT
QTOTAL,. 600FT X 0.00040 GAL/MIN/FT E 0.24 GAL/MIN
DISTANCE. FT FROM CENTER LINE OF DAM
900 BOO 700 600 500 400 300 200 100 100 200 300 400 500 600 700
650 650
t<v 2.5 K 10- 8 CM/SEC (

= _8 NOTE. DIMENSION NUMBERS WIT IN THE
1<, ( Kh- 22.5 x 10-8 CM/SEC K 1 : 7.5 K 10 CM/SEC
~600 FLOW NET INDICATE THE MBER 600~
" K (K" ~SO){ 10-B CMISEC

OF FLOW PATHS WITH REF RENeE:.
~
( K : 150 x 10-B CM/SEC TO 'THE CORE FLOW NET,
"
,..
6-20
2 Kh .. 450 ._ t0- 8 CM/SEC 2

~
z' K 1 x 1o- 4 CM/SEC z·
'2 3
0
~ 500 K4 50 x 10- 4 CM/SEC soo;:
«
~
~
•::~~~~==~~;;~~~~~;:::::::::~--~><0°________~_-_-j'"oc_-_-_-_i_~_~-~- - -'~.~ t~!.~ 1~ ~~~~~~~-~~~~~~~::~~~~;:~~-~-:-~~~O:C~K~T~O~Eil~~~
00000
00 0
ROCK
~ GROUT CURTAIN~A~O AND GRAV~L DRAIN ROCK TAILWATER
::
a. IMPERVIOUS FOUNDATIOI~
Figure 6-19. Seepage analysis, Blakely Mountain Dam, Arkansas (from U. S. Army Engineer
121
Waterways Experiment Station )
SEEPAGE PER FOOT OF EMBANKMENT LENGTH
CleMBANKMENT "''R'1 h , ... 7.5 X 10-B CM/SEC X 192FT X 13.5/8.6 X 60 SEC/MIN X 1/30.48 FT/CM
X 7,48 GAL/FT3"' 0.00033 GAL/MIN
OFOUNDATION • K 5 h #•
5 X 10- 4 CM/SEC X 192FT X 418.6 X 60 SEC/MIN X 1/30.48 FT/CM X 7.48
GAL/FT3 "' 0.66 GAL/MIN PER FOOT OF EMBANKMENT
FOR 600 FT EMBANKMENT
OTOTAL .. 600FT FT X (0.66 GAL/MIN/FT +0.00033 GAL/MIN/FTI "'400 GAL/MIN
660 660
600 600
COEFFICIENTS OF PERMEABILITY
K 1 , K 2, K 3 , AND K 4 SAME AS ABOVE
K 5 = Kv • Kh= 5 x 1o-4 CM/SEC
500 500
400 400
... ...
; "'>-
6-21
t:aoo 300 "

~
z- £
o 0
;::
~
~200 200
~"w-'
100 100
TAILWATER ASSUMED TO BE AT EL 385 FOR

CONSTRucnON OF FLOW NETS.
FOR CONSTRUCTION OF FLOW NET IN THE
FOUNDATION, IT WAS ASSUMED THAT THE
NOTE: DIMENSION NUMBERS WITHIN THE UPSTREAM COFFERDAM AND SHELL HAVE A
FLOW NET INDICATE THE NUMBER HIGH PERMEABILITY AND THAT THE FLOW THROUGH
TO THE C07
OF FLOW PATHS WITH REFERENCE
FLOW NET.
THE CORE AND RANDOM SECTION WOULD NOT
AFFECT THE FOUNDATION FLOW NET.
-100
b. PERVIOUS FOUNDATION
Figure 6-19. Concluded
EM 1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
(3) The flow nets were constructed assuming a tailwater elevation of

388 ft (approximate elevation of the base of the embankment) while seepage
quantities calculations assumed a tailwater elevation of 400 ft as shown on
the cross sections (i.e., h for calculation of Q is 192 ft).
(4) For calculation of Q in figure 6-19(b) the flow nets were con-
sidered to be separate but to have the same number of equipotential drops.
b. This example provides evidence that a large permeable member of the

embankment-foundation material controls the quantity of flow. In this case the
relatively impermeable embankment materials needed to be considered in order to
determine the upper seepage line, the extent of saturated materials, and the
expected pore pressures within the embankment. Seepage for this problem might
also be evaluated using a finite element computer program.
6-22
EM 1110-2-1901
30 Sep 86
CHAPTER 7
SEEPAGE TOWARD WELLS
7-1. Use of Wells. Wells are used in a variety of ways to control seepage.
They may be placed-on the landward side of water retention structures to
reduce pressure at the lower boundary of impervious strata. Wells are also
used to maintain dry conditions in excavations during construction. In
addition to seepage control, well pumping tests serve as an accurate means of
field determination of permeability (see Chapter 2).
7-2. Analysis of Well Problems. The graphical flow net technique described
in Chapter 4 or the approximate methods described in Appendix B can be used in
the analysis of well problems. However, formulas obtained from analytical
solutions to well problems are the most common methods of analysis.
a. Flow Nets. An example of a flow net for a simple flow problem is

shown in figure 7-1. The flow between flow lines is given by (Taylor 1948)
(7-1)
where
k = permeability (L/T)
= total head loss between equipotential lines (L)
r = distance from well (L)
b = dimension of element in Z direction (L)
= dimension of element in r direction (L)
As for a plan flow net, must be the same for all elements within
the net. Thus is a constant. When drawn in plan view (figure 7-1b)
the flow net consists of square elements as in the plane case described in
Chapter 4. When drawn in profile (figure 7-1c) the elements' aspect ratios
(b/R) are proportional to the radial distance r and are therefore not
squares. Thus, graphical construction of flow nets for radial flow problems
is generally not practical except for cases where the water bearing has a
constant thickness and only the plan view of the net is required.
b. Approximate Solutions. The numerical and analog methods described

in Appendix C can be used for problems involving complicated boundary condi-
tions. Electrical analog methods are especially advantageous as most compli-
cated well problems cannot be idealized in two dimensions.
7-1
EM 1110-2-1901
30 Sep 86
a. HORIZONTAL FLOW TO WELL
b. PLAN VIEW OF FLOWNET
c. PROFILE VIEW OF FLOWNET

Figure 7-1. Flowout of simple radial
flow problem (courtesy of John Wiley
and Sons268)
c. Analytical Formulas. The analysis of flow to a single well can

often be solved by analytical methods. Also, the analysis of flow to multiple
wells and many problems involving complicated boundary conditions can be
solved by superposition of solutions for single well problems. Analytical
solutions can be obtained for nonsteady flow problems.
7-3. Basic Well Equations for Steady State Flow. Steady flow conditions
exist when the well flow rate and piezometric surface do not change with time.
If the regional piezometric surface does not fluctuate, steady state condi-
tions are achieved by pumping from a well at a constant rate for a long time
period. Design of wells for seepage control are often based on computations
assuming steady state conditions.
7-2
EM 1110-2-1901
30 Sep 86
a. Artesian Conditions. When significant flow to a well is confined to

a single saturated stratum, the problem can be idealized as shown in fig-
ure 7-2a. An artesian condition exists when the height h of the piezometric
surface lies above the top of the water bearing unit b . If the properties of
the soil are constant in all directions from the well, the discharge Q from
the well must be equal to the flow through a cylinder defined by the radius
r , height b , and differential thickness dr . Thus from Darcy's law
(equation 3-3), the flow can be written as
(7-2)
where
Q = constant discharge from well (L3/T)
k = coefficient of permeability -(L/T)
= hydraulic gradient along radius (L/L)

r = radius of cylinder (L)
B = thickness of aquifer (L)
Upon integrating equation 7-2, the relationship between r and h is found.
(7-3)
The constant can be determined by specifying that at the radius re , the

total head h is equal to a known head H , the total head that existed
before starting discharge from the well. That is,
(7-4)
Inserting equation 7-4 into equation 7-3, the constant term is found to be:
7-3
EM 1110-2-1901
30 Sep 86
a. HORIZONTAL AQUIFER CONFINED BETWEEN IMPERVIOUS STRATA

(ARTESIAN FLOW)
b. HORIZONTAL UNCONFINED AQUIFER (GRAVITY FLOW)
c. COMBINED AND CONFINED AND GRAVITY FLOW
Figure 7-2. Radial flow to horizontal aquifers (courtesy of

164
John Wiley and Sons )
7-4
EM 1110-2-1901
30 Sep 86
By substituting the constant term into equation 7-3 and combining logarithmic
terms, the well equation for confined flow is obtained.
(7-5)
The distance re is often defined as the radius beyond which the well has no
influence or radius of influence.
b. Gravity Flow Conditions. Flow to a well under gravity (figure 7-2b)

differs from the confined flow problem in the important aspect that the height
B of the differential cylinder is equal to the variable h . Thus,
equation 7-2 must be written as:
(7-6)
which upon integration and substitution of boundary condition gives
(7-7)
The constant term can be evaluated from the condition at the radius of influ-
ence r e as was done in equations 7-4 and 7-5. The constant term is given
by:
which when substituted back into equation 7-7 gives the well equation for
gravity flow
Development of equation 7-6 is based on the Dupuit assumption (Chapter 4)

which limits the applicability of equation 7-7 to those cases where the slope
of the piezometric surface is small (less than 5 percent). The error is
greatest in the vicinity of the well.
7-5
EM 1110-2-1901
30 Sep 86
c. Combined Artesian and Gravity Flow. When drawdown of the potentio-

metric surface becomes large near the well, combined gravity and confined con-
ditions can occur (figure 7-2c).
d. Flow to Well Groups (Method of Superposition). The piezometric sur-

face h caused by discharge from a group of wells can be determined by super-
imposing the solution for the individual wells given by either equation 7-3 or
7-6. For multiple wells, flow cannot be idealized by concentric cylinders and
the problem must be stated in terms of the plan coordinates x and y
(figure 7-3b). By noting that for a well located at
a general well equation can be written as
(7-8)
where
= potential required at point (x,y) to sustain a discharge Qi from

a well at
= hi (confined flow)
= (unconfined or gravity flow)
qi = intensity factor
( confined flow)
(unconfined or gravity flow)
Ci = constant
The head distribution can be determined by summing the individual

As the sum of the constants Ci is a constant, the multiple well
equation can be written as
(7-9)
where n is the number of wells. The constant is determined from a known

value of at a specified location. For example, the superposition formula
7-6
EM 1110-2-1901
30 Sep 86
a. FLOW BETWEEN TWO WELLS
b. COORDINATE SYSTEM FOR FLOW TO MULTIPLE WELLS
Figure 7-3. Flow to multiple wells (adapted from John Wiley

164
and Sons )
7-7
EM 1110-7-1901
30 Sep 86
for the wells shown in figure 7-3 would be
At a distance re from both wells h(x,y) = H . The constant term is found

to be
Substituting the above equation into equation 7-9, the well formula for two
wells becomes
e. Hydrologic Boundaries (Image Well Method). When there are hydrologic

boundaries within the radius of influence of the well, equations 7-3 and 7-7
are no longer valid. Examples of boundaries are:
(1) A stream or river which can be idealized as a line source of equal

potential.
(2) A rock bluff line at the edge of an alluvial fill valley which can
be idealized as an impervious boundary.
The superposition of solutions (equation 7-9) can be used to analyze the flow
near a boundary by introducing an artificial device called an image well. An
image well is identical to the actual well and located symmetrically on the
opposite side of the boundary. The superimposed effect of the real and image
well for an infinite well is identical to the influence of the real well and
boundary. If the real well is a pumping well then a recharging image well is
used to represent boundaries such as rivers and a pumping image well is used
to represent an impervious barrier. For either case, the absolute value of
the flow Q for the image well is equal to that of the real well. For
7-8
EM 1110-2-1901
30 Sep 86
example, the head distribution created by a discharging well in the vicinity of

a river is identical to that created by the combined influence of a recharge
and discharge well (see figure 7-4). The head distribution created by the
discharge well in an infinite confined aquifer is given by
(7-10a)
and by the image recharge well in the infinite aquifer
(7-10b)
By superposition, the head distribution for the true actual problem is
(7-11)
Note that at the river rI = rR and Thus, constant = H , the
head at the river. Substituting the constant term into equation 7-11, the
formula for a single well near a recharge boundary is
To describe the head distribution for confined flow near an impervious boundary
an image discharge well is used (figure 7-4b), By the procedure used above, h
would be obtained as
(7-12)
7-9
EM 1110-2-1901
30 Sep 86
a. IMAGE WELL ANALYSIS OF DISCHARGE WELL NEAR

RECHARGE BOUNDARY (RIVER)
b. IMAGE WELL ANALYSIS OF DISCHARGE WELL NEAR

IMPERMEABLE BOUNDARY (ROCK BLUFF)
Figure 7-4. Application of image well method for analysis of flow

287
near boundaries (courtesy of Illinois State Water Survey )
7-10
EM 1110-2-1901
30 Sep 86
The head at the impervious boundary is unknown, thus additional information is

needed to determine the constant. Note that when rR and rI are both equal
to the radius of influence that h = H . Thus
(7-13)
The image well method can also be applied to problems involving multiple
boundaries. For example, a common geologic situation involving multiple
boundaries would be a discharge well pumping from an alluvial terrace located
between a river and rock bluff (figure 7-5). In this case, the image well for
the river would have a second image well with respect to the rock bluff, which
in turn would have an image with respect to the river and so on. A similar
progression of image wells would be needed for the impermeable barrier.
Eventually, the location of each added-image well extends beyond its radius of
influence re from the pumping well and has no practical influence in the
solution.
7-4. Special Conditions. Although the simple well formula (equation 7-8) is
often used to analyze flow problems, it describes a relatively idealized con-
dition that is found rarely in practice. It is generally desirable to con-
sider the effects of partial penetration of wells, sloping aquifer, and
stratification of water bearing units in the analysis.
a. Partially Penetrating Wells. In deriving equations 7-3 and 7-7 it

is assumed that the flow lines are horizontal at the entrance of the well.
This assumption is valid only if the well completely penetrates the water
bearing strata. An approximate solution for flow to a well partially pene-
trating a confined aquifer was developed by Muskat (1946). The head can be
computed from
(7-14)
where C and C are constants to be determined from boundary conditions

1 2
is a function of the radius from the well (Warriner and Banks 1977).
The expression for given by Muskat (1946) was based on simplifying assump-
tions. Duncan (1963) and Banks (1965) assessed its validity from electrical
analogy model studies and developed a more accurate expression for The
alternative empirically determined relationship for developed by Duncan
(1963) is given in figure 7-6. The constants C1 and C2 are determined
7-11
EM 1110-2-1901
30 Sep 86
b. SECTION
a. PLAN
c. PRIMARY IMAGE
WELLS TO ACCOUNT FOR
INFLUENCE OF BOUNDARIES
ON REAL WELL
d. SECONDARY IMAGE WELLS TO

ACCOUNT FOR INFLUENCE OF
BOUNDARIES ON PRIMARY IMAGE
WELLS
Figure 7-5. Multiple image wells for a two-boundary problem
7-12
EM 1110-2-1901
30 Sep 86
124
Figure 7-6. Beta function curve (from Warriner and Banks )
7-13
EM 1110-2-1901
30 Sep 86
from the boundary conditions at the well and at the radius of influence as
(7-15)
where
hw = total head at well (L)
= value of at well radius r w (dimensionless)
H = total head at radius of influence re (L)
The well discharge can be determined by using an empirically determined shape
(7-16)
with
where
b. Flows to Groups of Partially Penetrating Wells. An empirical method

developed by Warriner and Banks (1977) provides a means to modify the rela-
tionship obtained by superimposing solutions for individual fully penetrating
wells for the effects of partial penetration. First, the head at each well is
computed from the assumption that they fully penetrate the aquifer:
(7-17)
with
7-14
EM 1110-2-1901
30 Sep 86
where
hj = head at well j (L)
c = constant of integration (L)

3
Q i = discharge from well i (L /T)
k = coefficient of permeability (L/T)
B = aquifer thickness (L)
a = constant (L)
r ij = distance between well i and well j (L)
r w j = radius of well j (L)
N = number of wells in group
In addition, the head at a point on the source boundary is given by:
(7-18)
where H is the head at the source and ris is the distance between well i
and the source. The drawdown at each well is computed from combining
equations 7-17 and 7-18
(7-19)
with rjj = rwj .
Equation 7-19 gives the drawdown for each well within a group of fully
penetrating wells. The values of Qi required to cause the drawdown
can be determined by solving the system of N equations (7-19) for the N
unknowns Qi . As for the single well, a shape factor can be defined as:
(7-20)
7-15
EM 1110-2-1901
30 Sep 86
where is the shape factor for each well within a group. This shape
factor can be corrected to account for partial penetration by
(7-21)
By replacing in equation 7-16 with the flow from the well group is
given as
(7-22)
The computations required to evaluate equations 7-19 through 7-22 are straight-
forward though they are time consuming for large well groups. Warriner and
Banks (1977) provide a FORTRAN code to compute discharge and drawdowns for
partially penetrating well groups within an arbitrarily shaped source boundary.
c. Wells in Sloping Aquifer. If the regional potentiometric surface

has a significant slope, the effect of superimposing the initial regional
gradient on the well drawdowns must be considered. For example, when pumping
from floodplain locations, the existing piezometric gradient from upland areas
to the river may be as great as those caused by pumping from the well. The
significant parameters for confined flow to a single well are shown in
figure 7-7. At a large distance from the well, the regional flow net would
not be affected. All flow into the well would be contained within the stream
lines separated by the dimension f . Thus by Darcy's law for one-dimensional
flow
(7-23)
where
Q = discharge from well (L3/T)
h 1 = total head (L) for regional flow alone
7-16
EM 1110-2-1901
30 Sep 86
x = coordinate selected to be parallel to initial regional flow (L)
f = width of flow lines enclosing all flow to well (L)
Figure 7-7. Superposition of well drawdown on regional gradient

(courtesy of International Institute for Land Reclamation and
Improvement199)
7-17
EM 1110-2-1901
30 Sep 86
The corresponding differential equation for the well would be
(7-24)
where
= total head due to flow to well
At a distance Xe downgradient from the well, a groundwater divide develops

(culmination point) at which
(7-25)
In view of equations 7-23 and 7-24
(7-26)
or
By substitution of equation 7-26 into equation 7-23
(7-27)
By integrating equations 7-24 and 7-27
(7-28)
7-18
EM 1110-2-1901
30 Sep 86
and superimposing the effects
(7-29)
The distance Xe can be removed from the expression by substitution of

equation 7-27
(7-30)
where i = the regional slope of the aquifer.
For conditions of unconfined flow, the regional gradient would be defined by a

parabola
which when combined with the well equations for unconfined flow gives
(7-31)
d. Layered Aquifers. Natural soils often occur in layers and a well

may penetrate units having different permeabilities. If flow to the well is
horizontal, the simple well equations can be used by assigning an average
value of permeability given by
(7-32)
where
k m = horizontal permeability of layer m
d m = thickness of layer m
d = total thickness of layers
7-19
EM 1110-2-1901
30 Sep 86
Note that the permeability determined from a field pumping test is an average
of all units penetrated by the pumping well. A case where vertical flow can be
important is shown in figure 7-8. The discharging well is pumping from a
Figure 7-8. Flow to well with significant vertical flow through confining
164
layer (courtesy of John Wiley and Sons )
permeable unit overlain by a less permeable unit through which significant

vertical flow can occur. The flow to the well is given by
(7-33)
where
H = original total head (L)
h = total head at distance r from well at steady state condition

(L)
3
Q = discharge rate (L /T)
L = (leakage factor) (L)
B = thickness of aquifer
C = B'/k' (L)
B' = thickness of overlying low permeability unit (L)
7-20
EM 1110-2-1901
30 Sep 86
k ' = permeability of overlying low permeability unit (L/T)
= Hankel function (tabulated in table 7-1) (dimensionless)
7-5. Nonsteady State Flow. Nonsteady state flow may arise in several ways.
When pumping is started, time is required to establish a virtually steady state
condition. Flow during this period must be assumed to be nonsteady state. If
pumping occurs intermittently, a steady state condition may not be established.
Also, if large fluctuations occur at the source, potential steady state flow
conditions are not maintained. The steady state condition can be viewed as the
end condition that is reached after pumping for a long time period. In the
design of a well system for seepage control, it is generally adequate to
consider only the steady state condition. However, the determination of
coefficient of permeability from test data often requires analysis based on
nonsteady state condition. The duration of many well tests is too short to
Table 7-1. Values of Ko r/L for Selected Values of r/L to

Evaluation Equation 7-33 (a)
f = 10-2 f = 10-l f = 1.0
1.0 4.721 2.427 0.421
2.0 4.028 1.753 0.114
3.0 3.623 1.372 0.035
4.0 3.336 1.114 0.011
5.0 3.114 0.924 0.004
6.0 2.933 0.777 0.000
7.0 2.780 0.660 0.000
8.0 2.647 0.565 0.000
9.0 2.531 0.487 0.000
Example: = 0.5 , f = 10-1 , K(0.5) = 0.924
(a) Prepared from more extensive tables presented by Kruseman and De Ridder
(1970).
7-21
EM 1110-2-1901
30 Sep 86
reliably establish the steady state condition. Also, in practice, hydrologic

boundaries may be present within the steady state radius of influence. In
either case the use of the steady state flow equations could lead to substan-
tial error in determining the permeability.
a. Nonsteady State Confined Flow. Theis (1935) developed the following

relationship for nonsteady state flow in a confined aquifer (Davis and DeWeist
1966):
(7-34)
where
Q = constant discharge rate (L3/T)
W(u) = function given in table 7-2
The parameter u is given by
(7-35)
where
r = radius from well (L)
S = storage coefficient (dimensionless)
t = time from start of pumping (T)
The storage coefficient S represents the amount of water removed from stor-
age as a result of consolidation of the aquifer and expansion of water in
response to the decline in head. Physically S is given by
(7-36)
7-22
EM 1110-2-1901
30 Sep 86
Table 7-2. Values of W(u) for Selected Values of 1/u

to Evaluate Equation 7-34 (a)
1 1 W(u)
f u f = 1.0 f = 2.0 f = 8.0
10-1 0.000 0.001 0.146
1 0.219 0.600 1.623
10 1.823 2.468 3.817
102 4.034 4.726 6.109
103 6.332 7.024 8.410
104 8.633 9.326 10.71
105 10.94 11.63 13.02
106 13.24 13.93 15.32
107 15.54 16.23 17.62
108 17.84 18.54 19.92
10g 20.15 20.84 22.22
1010 22.45 23.14 24.53
1011 24.75 25.44 26.83
1012 27.05 27.75 29.13
1013 29.36 30.05 31.44
1014 31.66 32.35 33.74
Example : u = 0.005 , = 200 , f = 2 , W(u) = 4.726
(a) Prepared by WES.
7-29
EM 1110-2-1901
30 Sep 86
where
= mass density fluid (m/L3)
g = acceleration of gravity (L/T2)
B = thickness of aquifer
= bulk compressibility of aquifer (LT2/M)
n = porosity (dimensionless)
= bulk compressibility of fluid (LT2/M)
The determination of the aquifer properties kb and S from equation 7-34

requires a complete drawdown versus time history for each observation
piezometer. The Theis method for data analysis is based on the logarithmic
representation of equations 7-34 and 7-35
From the equations above it is seen that if Q is constant that log(H - h)

varies with log(r2/t) in the same way as log [W(u)] varies with log (u)
regardless of the units used. Therefore, it should be possible to superimpose
the data curve on the theoretical curve because the two curves are offset from
each other only by the constant terms log and log S/4kB . By deter-
mining the value of the offsets from the superimposed curves, kb and S can
be determined. The computation consists of the following steps:
(1) A plot is made of W(u) (log scale) versus u (log scale). This
plot is referred to as the type curve.
(2) For each observation well, a plot is made of drawdown H - h (log

scale) versus (log scale).
(3) Superimpose the test data over the type curve in such a way that
the drawdown data best fit the type curve (figure 7-9). The coordinate axes
of the two curves should be kept parallel.
(4) Determine the values W(u) , u , H - h , and r2/t from an

arbitrarily chosen matching point on the two curves.
7-24
EM 1110-2-1901
30 Sep 86
Figure 7-9. Use of type curve for analysis of nonsteady state flow
164
(courtesy of John Wiley and Sons )
(5) Compute the value of kB from equation 7-34 using the matching
point value of H - h and W(u) . Compute the value of S from equa-
2
tion 7-35 using the matching point values of u and r /t combined with the
previously computed value of kB . The above procedure is carried out for
each observation well. Ideally, the computed values of kB and S should be
the same for all observation wells. Differences in the computed values may be
caused by geologic variations in the aquifer and hydrologic boundaries not
accounted for in the analysis.
b. Simple Method for Coefficient Determination (Jacob's Method). Jacob

(1950) introduced a simplification to the determination of kB and S by
noting that for small values of u (small r and/or large t) equation 7-34
reduces to (Davis and DeWeist 1966)
7-25
EM 1110-2-1901
30 Sep 86
(7-37)
Equation 7-37 can be written in a form convenient for graphical solution by

substituting equation 7-35 and writing in terms of base 10 logarithms:
(7-38)
From equation 7-38 it is seen that the relationship between drawdown H - h

and time t for a particular observation piezometer (r = constant) can be
represented as a straight line on a plot of H - h versus log t
(figure 7-10). The slope of the line is equal to Also, the time,
to , corresponding to H - h = o gives
which can be used to determine S . An alternative analysis consists of plot-

ting H - h versus log r . The following relationship can be obtained by
rearranging the term in equation 7-38.
(7-39)
Equation 7-39 defines a straight line on a plot of H - h versus r

(figure 7-10b). The slope of the line is - and can be used to
determine kb . The line intersects the H - h = o axis at ro . This
intercept can be used to determine S from
Note that ro represents the radius of influence for the well at time equals
t . Thus the radius of influence for the steady state condition re is equal
to ro as t tends to infinity. This implies that the radius of influence
7-26
EM 1110-2-1901
30 Sep 86
a. One observation well
b. Simultaneous observations
Figure 7-10. Use of Jacob approximation for nonsteady state flow

(courtesy of John Wiley and Sons 164 )
7-27
EM 1110-2-1901
30 Sep 86
expands indefinitely and cannot be defined. However, the value of r

e
selected has a relatively small influence on computed drawdowns near the well
and equation 7-39 can be used to determine reasonable values for re .
c. Nonsteady Unconfined Flow with Vertical Gravity Drainage (Delayed

Yield). Initial response (generally after first few minutes of pumping) is
given by (Kruseman and DeRidder 1970)
(7-40)
where
S A = storage coefficient for instantaneous release of water from

storage
W(uA , r/B) = Boulton well function (figure 7-11a)
r/B = formation constant to be determined from piping test data
Later time response is given by
(7-41)
where
Sy = specific yield
W(u y , r/B) = delayed yield well function
The application of equations 7-40 and 7-41 through use of a type-curve is

similar to that of equation 7-34. The following should also be noted:
(1) Type curves for several values of r/B should be plotted. The
curve giving the best fit to the initial time-drawdown data is used to
estimate r/B .
7-28
EM 1110-2-1901
30 Sep 86
a FAMILY OF BOULTON TYPE CURVES: W (UA, r/B) VERSUS 1/uA AND W (Uy, r/b) VERSUS
1/Uy FOR DIFFERENT VALUES OF r/B.
a. Family of Boulton type curves: W(UA , r/B) versus 1/uA and

W(Uy/r/b) versus 1/Uy for different values of r/B
b. Boulton's delay index

curve
Figure 7-11. Type curves for Boulton's analysis of nonsteady unconfined

flow with delayed yield (courtesy of International Institute for Land
Reclamation and Improvement199)
7-29
EM 1110-2-1901
30 Sep 86
(2) The time-drawdown data overlay may be moved to obtain the best fit
for the latter time-drawdown data. Both initial time and latter time fits
should give the same value of r/B and kB .
(3) Eventually, the effects of vertical gravity drainage become negli-

gible and the latter time curve merges with the Theis curve. The time-
coordinate where the two curves merge is determined from Boulton's delay-index
curve (figure 7-11b).
(4) A number of type-curve solutions to the problem of nonsteady uncon-

fined flow to wells have been developed (Fetter 1980). For example, Neuman
(1975) presented a type-curve method similar to Boulton's that accounts for
anisotropy of the aquifer.
d. Nonsteady Confined Flow with Vertical Drainage Through Confining

Layer (Leaky Aquifer). The leaky aquifer equation for nonsteady flow is based
on the assumptions that flow to the well is horizontal and vertical flow is
restricted to seepage through the confining layer. These assumptions are
identical to those made for the steady state case described by equation 7-33.
The drawdown is given by
(7-42)
where
r 2S
u=4kBt
r = radius from well (L)
L = leakage factor (L) =
B ' = thickness of confining unit (L)
k' = permeability of confining unit (L/T)
W(u , r/L) = well function given in figure 7-12
7-30
NONEQUILIBRIUM
TYPE CURVE
0.0
.1 0.075 0 .05
0 3 0.2 0.15
·0 5°.4 .
o.s-·
1.0 0.8
1 14.6 OW!u, r/Bl
~-------- 1.5 5"'
T
----------2.0 u=
2100 ,-2s
Tt
7-31
~---------r/8=2.5
0.1~----~~~----------------~----------~~~~r~~-------1------~
r /8= /IK'/b') vT
0.01~~~~~-~~~~-~~~~~~~WUUL~~~~~~~~~ti-~-L~~~-L-L~~
10- 1 1.0 10 lol 103 10 4 10 5 106 107
1/u
Figure 7-12. Type curve for nonsteady flow in a semiconfined aquifer (courtesy of
EM 1110-2-1901
164
John Wiley and Sons )
30 Sep 86
EM 1110-2-1901
30 Sep 86
The application of the type curve method for the leaky aquifer problem is simi-
lar to the application to the delayed yield problem. The time-drawdown data
are matched to the standard type curve with the curve giving the best fit being
used to estimate r/B .
e. Nonsteady Unconfined Flow with Little Vertical Drainage. If the

delayed response component of the drawdown is small, the Theis equation
(equation 7-34) can be used to analyze the flow by inserting a "corrected"
drawdown into the flow equation. The corrected drawdown is given by
(7-43)
f. Nonsteady Flow with Hydrologic Boundaries. The method of super-

position presented for steady-state flow problems (equation 7-9) is applicable
to nonsteady flow problems. Therefore, the image well method can be used to
investigate the effects of hydrologic boundaries. For example, the image well
analysis for a discharging well near a river (recharge boundary) is (Davis and
Dewiest 1966)
(7-44)
where
r R = radius from real well (L)
rI = radius from image well (L)
7-32
EM 1110-2-1901
30 Sep 86
Note, then, when the function W(u) can be replaced with a logarithmic
approximation, as in the Jacob's method (equation 7-37), equation 7-44 can be
approximated as
(7-45)
From equation 7-45 it is seen that as u becomes small, flow becomes virtually
steady state (compare equation 7-45 with the steady state case, equation 7-11).
Thus the presence of a recharge boundary in an aquifer tends to shorten the
time needed to reach steady state (Davis and Dewiest 1966).
7-33
EM 1110-2-1901
30 Sep 86
CHAPTER 8
SEEPAGE CONTROL IN EMBANKMENTS
8-1. General. All earth and rock-fill dams are subject to seepage through
the embankment, foundation, and abutments. Seepage control is necessary to
prevent excessive uplift pressures, instability of the downstream slope, pip-
ing through the embankment and/or foundation, and erosion of material by
migration into open joints in the foundation and abutments. The purpose of
the project, i.e., long-term storage, flood control, etc., may impose limita-
tions on the allowable quantity of seepage.
8-2. Methods for Seepage Control. The three methods for seepage control in
embankments are flat slopes without drains, embankment zonation, and vertical
(or inclined) and horizontal drains.
8-3. Flat Slopes Without Drains. For some dams constructed with impervious
soils having flat embankment slopes and infrequent, short duration, high
reservoir levels, the phreatic surface may be contained well within the down-
stream slope and escape gradients may be sufficiently low to prevent piping
failure. For these dams, when it can be assured that variability in the
characteristics of borrow materials will not result in adverse stratification
in the embankment; no vertical or horizontal drains are required to control
seepage through the embankment. A horizontal drain may still be required for
control of underseepage (see Chapter 9). Examples of dams constructed with
flat slopes without vertical or horizontal drains are Aquilla Dam, Aubrey Dam,
and Lakeview Dam (U. S. Army Engineer District, Fort Worth 1976a, 1976b, 1980).
Figure 8-1 shows the analysis of through embankment seepage for Aubrey Dam,
Texas (now called Ray Roberts Dam). As shown in figure 8-1a, this is a zoned
embankment with relatively flat slopes due to a weak stratum in the founda-
tion. The slopes could be steepened from IV:10.6H to lV:8H if the weak foun-
dation gains shear strength due to consolidation during construction (the dam
is scheduled for completion in 1988). A 3-ft-thick horizontal drainage blanket
and collector system will be provided under the downstream embankment from
sta 136+00 to sta 142+60 to control any seepage through the foundation. For
the analysis of through embankment seepage, shown in figure 8-1b, the steady
state phreatic surface was developed graphically for the conservation pool by
considering a homogeneous nonisotropic embankment and an impervious foundation.
Since the escape seepage gradients were computed to be less than 0.3 to 0.4
(see paragraph 4.9b), it was concluded that no vertical or horizontal drains
were required.
8-4. Zoning Embankments.
a. General. Embankments are zoned to use as much material as possible

from required excavation and from borrow areas with the shortest haul dis-
tances and the least wastage and at the same time maintain stability and con-
trol seepage. The different zones of an embankment are shown in figure 8-2.
For most effective control of through seepage and seepage during reservoir
drawdown, the permeability should progressively increase from the core out
toward each slope as shown in figure 8-2 (EM 1110-2-2300).
8-1
30 Sep 86
EM 1110-2-1901
ct_ EMBANKMENT
SELECT
IMPERVIOUS I
44'
FILL
UPSTREAM 18" RIPRAP ON 6 DOWNSTREAM
.--
EL 511.0
3' HORIZONTAL DRAINAGE BLANKET ST.
FLOODPLAIN EMBANKM'NTSECTION
STATION 100+00 TO STATION 142+25
...
SCALES
HORIZONTAL
30 0 30 60FT
MW ~ERT!CAL
a. FLOODPLAIN EMBANKMENT SECTION
8-2
£EMBANKMENT
UPSTREAM DOWNSTREAM
FLOODPLAIN EMBANKMENT
LEGEND
...
b. EMBANKMENT SEEPAGE ANALYSIS
50 100FT
KH =HORIZONTAL PERMEABILITY
Kv ~VERTICAL PERMEABILITY
Figure 8-1. Analysis of through embankment seepage for Aubrey Dam, Texas
81
(from U. S. Army Engineer District, Fort. Worth )
EM 1110-2-1901
30 Sep 86
Figure 8-2. Different zones of an embankment (prepared by WES)
b. Impervious Zone or Core. The purpose of the core is to minimize

seepage losses through the embankment. As a general rule, sufficient imper-
vious material is available to result in small seepage losses through the
embankment. Therefore, the quantity of seepage passing through the foundation
and abutments may be more significant than the quantity passing through the
core. Important material properties of the core are permeability, erosion
resistance, and cracking resistance. A core material of very low permeability
may be required when the reservoir is used for long-term storage. A core
material of medium permeability may be utilized when the reservoir is used for
flood control. The erosion resistance of core material is important in evalu-
ating piping potential (Arulanandan and Perry 1983). The tensile strength of
the core material is important in evaluating the cracking resistance
(Al-Hussaini and Townsend 1974). In general, the base of the core or the cut-
off trench should be equal to or greater than a quarter of the maximum differ-
ence between reservoir and tailwater elevations (U. S. Army Engineer District,
Mobile 1976, EM 1110-2-2300). A core top width of 10 ft is considered to be
the minimum width on which earth-moving and compaction equipment can operate.
The maximum core width is controlled by stability and by availability of mate-
rial. The top of the core should be above the maximum reservoir elevation
8-3
EM 1110-2-1901
30 Sep 86
but below the bottom of the frost zone, A vertical core located near the
center of the dam is performed over an inclined upstream core because the
former provides higher contact pressure between the core and foundation to
prevent leakage, greater stability under earthquake loading (Sherad 1966,
1967), and better access for remedial seepage control. An inclined upstream
core allows the downstream portion of the embankment to be placed first and the
core later and reduces the possibility of hydraulic fracturing (Nobari, Lee,
and Duncan 1973).
c. Filter Zones. Filters may be required in various locations in earth

dams such as vertical (or inclined) and horizontal drains within the down-
stream section of the embankment as shown in figure 8-2, around outlet con-
duits passing under the downstream portion of the embankment, under concrete
structures such as stilling basins, around relief wells, beneath riprap where
drawdown may occur, and between the embankment and abutment. Important prop-
erties of the filter material are gradation, compacted density, and permeabil-
ity. Filters are designed to permit free passage of water and prevent migra-
tion of fines through the filter as discussed in Appendix D. The average
in-place relative density of the filter should be at least 85 percent and no
portion of the filter should have a relative density less than 80 percent
(EM 1110-2-2300). This requirement applies to vertical (or inclined) and
horizontal drains and filters under concrete structures but not to bedding
layers under riprap. Special care must be taken to assure that compaction does
not degrade the filter material (by grain breakage and/or segregation) and
reduce its permeability. When the filter material is sand or contains signifi-
cant portions of sand sizes, the material should be maintained in as saturated
a condition as possible during compaction to prevent bulking. The discharge
capacity of the filter zones should be determined in dimensioning the filters
(Cedergren 1977). The filter material should pass the 3-in. screen for mini-
mizing particle segregation and bridging during placement. As discussed pre-
viously in Chapter 2 (see figure 2-12), the permeability of sands and gravels
varies significantly with the amount and type of fines (material smaller than
the No. 200 sieve) present. Also, the amount and type of fines present
influence the capacity of a filter to self-heal by collapsing any cracks within
the filter (see figure 8-3). Therefore, the maximum percent fines and type
(silt, clay, etc.) to be allowed in the filter of an earth dam must be shown
to be sufficiently pervious by laboratory filter tests (1) and self healing by
collapse tests (Vaughn 1978). If vibration is present, such as in the vicinity
of a stilling basin or powerhouse, the. laboratory filter tests should be con-
ducted with vibration effects. If the base material to be protected is
dispersive, large-scale box filter tests will be required (McDaniel and Decker
1979, Bordeaux and Imaizumi 1977, and Logani and Lhez 1979). The procedure to
use in identifying dispersive clays is given in EM 1110-2-1906. Generally, two
or more filter zones, each with a uniform or narrow gradation (sand, pea
gravel, etc.) are preferable to a single well-graded filter zone which often
becomes segregated during processing, stockpiling, and placement. Care must
(1) Laboratory filter tests are not a routine laboratory test. Standard test-
ing procedures have not been developed. The conduct of laboratory filter
tests should be under the direction of a specialist and should be carried
out in a research laboratory.
8-4
EM 1110-2-1901
30 Sep 86
a. Self-healing (by collapse) of filter
b. Laboratory test for ability of filter material

to self-heal (by collapse)
Figure 8-3. Self-healing (by collapse) of crack
within a filter downstream of a core (courtesy
282
of American Society of Civil Engineers )
be taken during construction to prevent reduction in permeability of the filter

by intrusion of fines carried by surface runoff, spillage by compaction equip-
ment, or degradation during compaction. Also, care must be taken to prevent
coarse material from rolling down the surface of the filter and collecting
between the core and filter (or between filter zones if two or more filters are
used) forming a "tube" (in cross section) of more permeable material through
which core (or filter) material could be lost by piping.
8-5
EM 1110-2-1901
30 Sep 86
d. Transition Zones. The purpose of transition zones is to separate

zones of different permeability and compressibility within the embankment, to
prevent core material from being drawn into the upstream shell during rapid
drawdown of the reservoir, to provide a source to feed material into a crack
in the core and preventing piping (see figure 8-4). Important material
properties of the transition material are angularity of particles (upstream of
core, rounded particles and better for feeding material into cracks and pre-
venting piping), gradation, permeability, and compressibility. Transition
zones may be located both upstream and downstream of the core and generally
have a width >10 ft. Wide transition zones between the filter and the down-
stream shell will control the rate of flow through a crack in the core and
extending through the filter in the event that self-healing (by collapse) of
the filter does not occur. Transition zones (and filter zones) should be
widened near abutments where tension zones may induce cracking.
Figure 8-4. Crack stopping transition upstream of a core

(prepared by WES)
e. Random Zones. The purpose of random zones is to utilize required

excavation. Random zones are assumed to have the properties of the least
desirable material in the excavation. Random zones may be located either
upstream or downstream of the core. For most effective control of through
seepage and seepage during reservoir drawdown, the more pervious material
should be routed to the outer portions of the embankment.
f. Cuter Zones or Shells. The purpose of the outer zones or shells is

to permit steeper embankment slopes. Important material properties of the
shell material are durability (soundness) of rock, gradation (well-graded),
and permeability (free-draining). The upstream shell affords stability
8-6
EM 1110-2-1901
30 Sep 86
against end of construction, rapid drawdown, and earthquake loading. The

downstream shell acts as a drain and controls the line of seepage and provides
stability under high reservoir heads. When suitable materials are not avail-
able for pervious downstream shells, control of seepage through the embankment
is provided by vertical (or inclined) and horizontal drains.
g. Upstream Drawdown Blanket. The purpose of the upstream drawdown

blanket is to provide stability of the upstream slope during rapid drawdown of
the reservoir. Important material properties of the upstream drawdown blanket
are durability (soundness) of rock and permeability (free-draining). Fig-
ure 8-5 shows the improvement in the factor of safety resulting from the
Figure 8-5. Improvement in factor of safety

resulting from the upstream drawdown blanket
8-7
EM 1110-2-1901
30 Sep 86
upstream drawdown blanket (Cedergren 1977). The required minimum permeability

of the upstream drawdown blanket can be calculated from (Cedergren 1977):
(8-1)
where
k m i n = minimum permeability of the upstream drawdown blanket
v d d = velocity of drawdown of the water surface in the reservoir
L = defined in figure 8-6
n e = effective porosity of the upstream drawdown blanket
h = defined in figure 8-6
= angle between the median flow line in the upstream drawdown

blanket and the horizontal
The ratio h/L should not exceed 10.
Figure 8-6. Computation of required minimum permeability of the

155
upstream drawdown blanket (courtesy of John Wiley and Sons )
8-5. Vertical (or Inclined) and Horizontal Drains.
a. Need. As stated previously, vertical (or inclined) and horizontal

drains may be required to control seepage through the embankment by preventing
material eroded through a crack in the core from washing into the downstream
shell by seepage water under reservoir head. Also, because of the often
variable characteristics of borrow materials, it is frequently advisable to
provide vertical (or inclined) and horizontal drains within the downstream
section of the embankment, as shown in figure 8-7, to ensure satisfactory
seepage control. For a stratified soil, the vertical permeability is
8-8
EM 1110-2-1901
30 Sep 86
a. HORIZONTAL DRAIN ONLY
b. INCLINED AND HORIZONTAL DRAINS
Figure 8-7. Use of inclined and horizontal drains to ensure seepage

control against variable characteristics of borrow materials
155
(courtesy of John Wiley and Sons )
controlled by the least permeable layer. Therefore, the horizontal permeabil-

ity is always greater than the vertical permeability. Compacted soils in earth
dams are stratified due to variability in the characteristics of borrow mate-
rials and the tendency for soil particles to align horizontally during compac-
tion. The ratio of vertical to horizontal permeability may range from 2 to 10
or greater. For stratified soils, as shown in figure 8-8, a horizontal drain-
age blanket is not sufficient to prevent the downstream slope from becoming
saturated and susceptible to piping and/or slope failure. However, when a
properly designed and constructed inclined drain and horizontal drain is used,
as shown in figure 8-8, complete control is provided over seepage through the
embankment.
b. Filter Requirements. Vertical (or inclined) and horizontal drains

should be designed as filters (see Appendix D). If crushed rock is used for
the drain material (see paragraph 2-2.g), material to be protected is disper-
sive, or material to be protected contains cracks, filter tests will be
8-9
EM 1110-2-1901
30 Sep 86
KH = KV
KH = 9KV
KH = 9KV
Figure 8-8. Effect of anisotropy of permeability on seepage through

earth dam with and without an inclined drain (courtesy of John Wiley
and Sons155)
8-10
EM 1110-2-1901
30 Sep 86
required. Well-graded materials are internally unstable and should not be used
as filters when Cu > 20.(l)
c. Discharge Requirements. Vertical (or inclined) and horizontal

drains must have sufficient discharge capacities to remove seepage quickly
without inducing high seepage forces or hydrostatic pressures (Cedergren
1977). When drains are designed and constructed with ample discharge capac-
ity, the line of seepage will not rise above the drain zone. Since drains are
small compared to the overall dimensions of the earth dam, it is difficult to
construct accurate flow nets within the drains themselves. The total quantity
of seepage from all sources that must discharge through the drain should be
evaluated from a flow net analysis in which it is assumed that the drains have
an infinite permeability. Figure 8-9 shows an example of the design procedure
for inclined and horizontal drains to assure adequate drain capacity (Cedergren
1977). The probable rate of discharge through the dam and foundation is esti-
mated from composite flow nets (see figures 4-13 and 4-14). For the example
shown in figure 8-9 , the seepage through the dam Q1 = 2 cu ft/day and the
seepage through the foundation Q 2 = 10 cu ft/day. Therefore, the inclined
drain must be capable of discharging Q1 = 2 cu ft/day and the horizontal drain
must be capable to discharging Q1 + Q2 = 12 cu ft/day. These are discharge
rates per running foot of dam and drain. Assuming the inclined drain was
designed with a width of 12 ft to permit its placement with normal earth-moving
equipment, the cross-sectional area normal to the direction of flow within the
inclined drain Ac = 11 sq ft (see figure 8-9b) and its required minimum
permeability may be found from Darcy's law:
(8-2)
(1)
where C u = coefficient of uniformity
= size of filter material at 60 percent passing
= size of filter material at 10 percent passing
8-11
EM 1110-2-1901
30 Sep 86
where
Q = quantity of discharge
i = hydraulic gradient
A = cross-sectional area of flow
a. Cross section and flow net
b. Dimensions of inclined drain
c. Dimensions of horizontal drain
Figure 8-9. Example of design procedure for inclined and horizontal drains
155
to assure adequate drain capacity (courtesy of John Wiley and Sons )
8-12
EM 1110-2-1901
30 Sep 86
Substituting for Q , 1 , and A
(8-3)
Every filter must be permeable enough to have a reasonable reserve for higher
than expected flows. The filter should have a minimum permeability after
placement and compaction of at least 20 times that calculated theoretically.
Therefore, the required permeability for the inclined drain is
(8-4)
Clean, washed concrete sand is usually about this permeable. As previously

stated, the horizontal drain must be capable of discharging Q1 + Q2 = 12 cu
ft/day. Since the drain is to be designed so that the line of seepage does
not rise above the drain zone, the allowable maximum head in the horizontal
drain can be no greater than its height. The required minimum permeability
from Darcy's law is
(8-5)
Substituting Ab = hb (width is one running foot of dam and drain)
8-13
EM 1110-2-1901
30 Sep 86
(8-6)
To design the horizontal drain, select a drain height and calculate the
required minimum permeability. Apply a factor of safety of 20 to the calcu-
lated permeability and select a drain material from available aggregates.
Select a drain height of 4 ft.
The required permeability for design is
This permeability could be obtained by screened fine gravel (3/8-in. to 1/2-in.

size) which has a permeability of about 30,000 ft/day or 10.6 cm/sec
8-14
EM 1110-2-1901
30 Sep 86
(Cedergren 1977). Seepage in coarse aggregate is likely to be turbulent and a

reduction factor should be applied to the permeability. The hydraulic gradient
in the horizontal drain is
(8-7)
From figure 8-10, for screened fine gravel (3/8-in. to 1/2-in. size) with
1 = 0.007 the reduction factor for permeability is 0.9.
(8-8)
The permeability of the screened fine gravel (3/8-in. to 1/2-in. size) reduced
for turbulence is greater than the required permeability for design:
Therefore, it should be adequate when properly placed and compacted to conduct

seepage water through the horizontal drain. The screened fine gravel (3/8-in.
to 1/2-in. size) will be protected top and bottom with a 1-ft-thick clean
washed concrete sand filter. Since the seepage from the foundation must flow
across the fine filter to enter the coarse drainage layer, the fine filter must
be permeable enough to allow the water to enter the coarse drainage layer
freely under only a small hydraulic gradient (0.5 or less). Assume an average
hydraulic gradient of 0.5 across the fine filter layer and Q2 = 10 cu ft/day
for the amount of water that will enter the first (left) 200 ft of the drain.
From Darcy's law the required minimum permeability of the fine filter is (see
equation 8-2)
k = 0.1 ft/day
8-15
EM 1110-2-1901
30 Sep 86
Figure 8-10. Approximation for estimating reduction

in permeability of narrow size-range aggregate
caused by turbulent flow (courtesy of John Wiley
and Sons155)
Clean washed concrete sand with a permeability of 10 ft/day should allow

seepage from the foundation to enter the coarse drainage layer without
restriction. As stated previously, the inclined and horizontal drains used in
this example would have to meet the filter requirements (see Appendix D) in
addition to the discharge requirements.
d. Location and Geometry. Vertical (or inclined) drains are located

adjacent to and downstream of the core as shown in figure 8-2. The top of the
vertical (or inclined) drain should be above the phreatic surface for maximum
reservoir elevation to prevent seepage flow above the drain. In drawing the
flow net for the dam to use in selecting the height and location for the ver-
tical or inclined drain, a conservative (high) value of anisotropy of perme-
ability of the soil should be used in order to prevent the seepage from flow-
ing over the top of the drain (see figure 8-11). If the dam is located where
earthquake effects are likely (see paragraph 8-6) , the vertical (or inclined)
drain should extend the full height of the dam. The width of the vertical (or
inclined) drain is controlled by the availability of materials and the dis-
charge requirements of the drain. A vertical (or inclined) drain width of
6 ft is the practical minimum for earth-moving and compaction equipment (U. S.
Army Engineer District, Kansas City 1974 and U. S. Army Engineer District,
Philadelphia 1974). When filter or drain material is not available locally and
must be hauled to the site at a substantial cost, a narrow (3 ft or greater)
vertical drain may be constructed by excavating into the core material, back-
filling, and compacting with vibratory equipment (U. S. Army Engineer District,
Mobile 1965, U. S. Army Engineer District, Tulsa 1974). The width of the ver-
tical drain must be sufficient to satisfy the discharge requirements. Hori-
zontal drains are located under the downstream section of the dam and convey
seepage from the vertical (or inclined) drain and underseepage from the
8-16
EM 1110-2-1901
30 Sep 86
CORRECT WAY INCORRECT WAY

Figure 8-11. Use of stringers or finger drains as an alternative
to a continuous horizontal drain (prepared by WES)
foundation to the toe of the dam. The thickness of the horizontal drain must
be sufficient to satisfy the discharge requirements. When filter or drain
material is not available locally and must be hauled to the site at substantial
cost, a thin (2 ft or greater) horizontal drain has been used (U. S. Army Engi-
neer District, Tulsa 1975 and U. S. Army Engineer District, Louisville 1974).
Stringers or finger drains may be used as an alternative to a continuous hori-
zontal drain, as shown in figure 8-11, when the drain material is costly. The
cross-sectional area of the stringer drains must be sufficient to satisfy the
discharge requirements. The stringer drain may be constructed either by
trenching into the embankment or foundation for narrow (6 ft) widths or by
placing the adjacent impervious fill and then the drain material for wider
(50 ft) widths (U. S. Army Engineer District, Kansas City 1974, 1978). In
either case, the side slopes of the stringer drains should be sloped instead of
8-17
EM 1110-2-1901
30 Sep 86
vertical (see figure 8-11) to avoid stress concentrations which could cause
vertical transverse cracks. The stringer drain material must be thoroughly
compacted (see paragraph 8-4c) to ensure that consolidation does not occur upon
saturation leaving an open seepage conduit in the top of the trench, bridged by
the overlying embankment, and susceptible to progressive erosion (Jansen 1980).
The downstream end of horizontal drains and stringer drains must be able to
discharge freely and must be protected against siltation and erosion. This may
be accomplished by providing a weighted filter (riprap overlying bedding mate-
rial) or a toe drain as shown in figure 8-12 (U. S. Army Engineer District,
Mobile 1965 and U. S. Army Engineer District, Tulsa 1975). The toe drain has
the advantage of lower maintenance requirements and preventing the development
of localized wet areas at the surface along the downstream toe of the
embankment.
8-6. Seepage Control Against Earthquake Effects. For dams located where
earthquake effects are likely, there are several considerations which can lead
to increased seepage control and safety.- The core material should have a high
resistance to erosion (Arulanandan and Perry 1983). Relatively wide transition
and filter zones adjacent to the core and extending for the full height of the
dam can be used. Additional screening and compaction of outer zones or shells
will increase permeability and shear strength, respectively. Geometric con-
siderations include using a vertical instead of inclined core, wider dam crest,
increased freeboard, and flatter embankment slopes, and flaring the embankment
at the abutments (Sherard 1966, 1967).
8-18
EM 1110-2-1901
30 Sep 86
a. Weighted filter (courtesy of U. S. Army

Engineer District, Mobile97)
b. Toe drain (courtesy of U. S. Army Engineer District, Tulsa 113)
Figure 8-12. Protection of downstream end of horizontal and

stringer drains
8-19
EM 1110-2-1901
30 Sep 86
CHAPTER 9
SEEPAGE CONTROL IN EARTH FOUNDATIONS
9-1. General. All dams on earth foundations are subject to underseepage.

Seepage control in earth foundations is necessary to prevent excessive uplift
pressures and piping through the foundation (seepage control in earth abut-
ments is given in Chapter 10). The purpose of the project, i.e., long-term
storage, flood control, hydropower, etc., may impose limitations on the allow-
able quantity of seepage. Generally, siltation of the reservoir with time
will tend to diminish underseepage (U. S. Army Engineer Division, Ohio River
1945). Conversely, the use of some underseepage control methods such as relief
wells and toe drains may increase the quantity of underseepage (Sowers 1962).
9-2. Selection of Method for Seepage Control. The methods for control of
underseepage in dam foundations are horizontal drains, cutoffs (compacted back-
fill trenches, slurry walls, concrete walls, and steel sheetpiling (1)), up-
stream impervious blankets, downstream seepage berms, toe drains, and relief
wells. To select an underseepage control method for a particular dam and
foundation, the relative merits and efficiency of different methods should be
evaluated by means of flow nets or approximate methods as described in
Chapter 4 and Appendix B, respectively. As shown in table 9-1, the changes in
the quantity of underseepage, factor of safety against uplift, and uplift pres-
sures at various locations should be determined for each particular dam and
foundation. Since the anisotropy ratio of the foundation has a significant
influence on the results of the underseepage analysis, this parameter should be
varied as appropriate to cover the possible range of
expected field conditions.
9-3. Horizontal Drains. As mentioned previously in Chapter 8, horizontal

drains are used to control seepage through the embankment and to prevent
excessive uplift pressures in the foundation. As shown in figure 9-1, the use
of the horizontal drain significantly reduces the uplift pressure in the foun-
dation under the downstream portion of the dam. The computation of uplift
pressure was illustrated previously in figure 4-15. Figure 9-1 also shows
that the horizontal drain increases the quantity of seepage under the dam.
9-4. Cutoffs.
a. Complete Versus Partial Cutoff. When the dam foundation consists of

a relatively thick deposit of pervious alluvium, the designer must decide
whether to make a complete cutoff (compacted backfill trench, slurry trench,
or concrete wall) or allow a certain amount of underseepage to occur under
controlled conditions (partial cutoff , upstream impervious blanket, downstream
seepage berm, toe trench drain, or relief walls), In some cases, where the
alluvium is not very deep or the water is very valuable, it may be obvious
(1) Steel sheetpiling is not recommended to prevent underseepage but is used

to confine the foundation soil and prevent piping.
9-1
Table 9-1. Selection of Underseepage Control Method
EM 1110-2-1901
30 Sep 86
Changes in i (a)
c
Quantity of F
i
Uplift Pressure
s
Method Under seepage e Under Dam Toe of Dam Downstream of Dam
Horizontal Drain
Complete Cutoff r---------------------,
1 Fill
in table for each particular 1
Partial Cutoff
I dam and foundation. Vary the I
Upstream Impervious Blanket
I permeability anisotropy of the :
Downstream Seepage Berm
Toe Drain ILI foundation
______
(K
1~_=_1,_10, 25, 100) as appropriate.~
_ _ _ _ _ _ _ _ _ _ _ _ ...J
Relief Wells
9-2
(a) Factor of safety against uplift at downstream toe of dam given by the ratio of the
critical hydraulic gradient to the vertical component of the exit gradient.
(prepared by WES)
EM 1110-2-1901
30 Sep 86
Figure 9-1. Influence of horizontal drain on uplift pressure

(prepared by WES)
after relatively little study that a complete cutoff is justified. For

example, the hydropower requirements of the Clarence Cannon Dam, Missouri,
indicated that a complete cutoff was required to sustain power requirements
during periods of little or no rainfall (U. S. Army Engineer District,
St. Louis 1969). In many cases, where the cost of a complete cutoff is great
and where the amount of underseepage without a complete cutoff is problemati-
cal, the decision is not easy. Factors which govern the decision for the type
of underseepage control measure to be used are (Sherard 1968):
(1) Economic comparison of the value of the water or hydropower which

may be lost versus the cost of the complete cutoff.
9-3
EM 1110-2-1901
30 Sep 86
(2) The resistance of the foundation alluvium with respect to potential

progressive backward erosion of leaks or piping. If the foundation contains
layers of fine sand or cohesionless silt, and particularly if these soils are
exposed on the surface of the valley floor or walls, a complete cutoff is more
desirable than if the foundation is basically gravelly (or even coarse sand).
If a large leak develops in a relatively coarse alluvium, in all probability
it will be safe against progressive backward erosion, but even a small con-
centrated leak emerging below the dam in fine cohesionless soils can be
hazardous.
(3) If the tailwater conditions are such that ponds of water exist
downstream of the dam so that underseepage would emerge underwater and could
not be observed, it is desirable to be more conservative in evaluating the
need for a complete seepage cutoff.
(4) The amount of silt and clay sized particles in suspension in the
river water which contributes to siltation of the reservoir with time and
tends to diminish underseepage.
Theory and model tests indicate that it is necessary for a cutoff to penetrate
a homogeneous isotropic foundation at least 95 percent of the full depth before
there is any appreciable reduction in seepage beneath an earth embankment as
shown in figure 9-2 (Telling, Menzies, and Coulthord 1978; and Mansur and
Perret 1949). The effectiveness of the partial cutoff in reducing the quan-
tity of underseepage decreases as the ratio of the width of the dam to the
depth of penetration of the cutoff increases (see figure 9-2). Partial cut-
offs are effective only when they extend down into an intermediate stratum of
lower permeability. This stratum must be continuous across the valley founda-
tion to ensure that three-dimensional seepage around a discontinuous stratum
does not negate the effectiveness of the partial cutoff.
b. Efficiency of Cutoffs. The effectiveness of the cutoff is assessed

either in terms of the flow efficiency (Casagrande 1961)
(9-1)
where
E q = flow efficiency of cutoff
Q o = rate of underseepage without cutoff
Q = rate of underseepage with cutoff
or head efficiency (Lane and Wohlt 1961)
(9-2)
9-4
EM 1110-2-1901
30 Sep 86
where
E H = head efficiency of cutoff
h = head loss between points immediately upstream and downstream of the

cutoff wall at its junction with the base of the dam
H = head loss across the dam
The head efficiency is more widely used because the field performance may be
established from piezometric data taken during construction, before and during
initial filling of the reservoir, and subsequently as frequently as necessary
to determine changes that are occurring and to assess their implications with
respect to safety of the dam, as described in Chapter 13. The flow efficiency
may only be approximated since the rate of underseepage without the cutoff
cannot be directly established and since it is difficult to measure the rate
of underseepage with the cutoff because, except for special cases, only part
of the underseepage discharges at the ground surface immediately downstream of
the dam (Telling, Menzies, and Simons 1978a and Marsal and Resendiz 1971). The
flow efficiency of a compacted backfill partial cutoff in a foundation of
permeable soils of moderate thickness overlying an impervious rock is shown in
figure 9-3. This figure also illustrates the high seepage gradients that occur
along the base of the cutoff and on its downstream face in both the foundation
and embankment zones. Suitable filters must be provided to prevent piping of
soil at faces A-B-C in figure 9-3a and 9-3b (Cedergren 1977 and Klohn 1979).
As shown in figure 9-4, a partial cutoff in a homogeneous isotropic foundation
will lower the line of seepage in the downstream embankment somewhat but exit
gradients at the downstream toe (as reflected by the distance between the
equipotential lines) are reduced only slightly (Cedergren 1973). When the
pervious foundation is cut off by a compacted backfill or slurry trench, (1) the
rate of underseepage may be estimated by (Ambraseys 1963 and Marsal and
Resendiz 1971)
(9-3)
where
3
Q o = rate of underseepage in m /set per running meter of dam
K O = permeability of the foundation in m/sec
(1) This approach neglects the contribution of the filter cake that forms on
the trench walls to the overall slurry trench permeability. When the
permeability of the backfill placed in the trench is high, the overall
slurry trench permeability will be controlled by the filter cake
(D'Appolonia 1980).
9-5
EM 1110-2-1901
30 Sep 86
H = head of water in the reservoir in m
B = width of the base of the dam in m
D = thickness of the foundation in m
k = permeability of the compacted backfill or slurry trench backfill in

m/sec
E = thickness of the cutoff in m
For a concrete wall or steel sheetpiling with defects (openings in the cutoff)
the rate of underseepage per unit length of cutoff is given by (Ambraseys 1963
and Marsal and Resendiz 1971)
(9-4)
where
W = total area of openings in m2
Figure 9-5 compares the rate of underseepage for an impervious upstream

blanket, compacted backfill trench or slurry trench, and concrete wall or
steel sheetpiling with defects. As shown in figure 9-5, the rate of
underseepage loss is the same for an impervious upstream blanket as for a
compacted backfill trench or slurry trench provided B' =
If K is relatively high, assuming gives B' = 49 E . Such compu-

tations allow preliminary cost estimates to be made to determine whether an
impervious upstream blanket is preferable over a compacted backfill trench or
slurry trench. Figure 9-6 can be used to determine the relative magnitudes of
the length of the impervious upstream blanket, the thickness of the compacted
backfill trench or slurry trench, or the area of the defects (openings) in the
concrete wall or steel sheetpiling that would result in the same rate of
underseepage for a given dam.
c. Compacted Backfill Trench. The most positive method for control of

underseepage consists of excavating a trench beneath the impervious zone of the
embankment through pervious foundation strata and backfilling it with compacted
impervious material. The compacted backfill trench is the only method for
control of underseepage which provides a full-scale exploration trench that
allows the designer to see the actual natural conditions and to adjust the
design accordingly, permits treatment of exposed bedrock as necessary, provides
access for installation of filters to control seepage and prevent piping of
soil at interfaces, and allows high quality backfilling operations to be car-
ried out (U. S. Army Engineer District, St. Louis 1969 and Cedergren 1977).
9-6
0~----------~-----------T------------~----------~----------~
L
20
0
J:o D
1-x
n..o
w, 40 ~-----------+------------~------------+-------------+-------~~~
0"'0
lL..
0 0 ..
1-w
9-7
zl-
w~ 60~----------+-----------4------------r-----------+----~~~~
ucr
a:f-
ww
CLZ
w
Q..
20 40 60 80 100
EM 1110-2-1901
PERCENT OF TOTAL SEEPAGE
30 Sep 86
Figure 9-2. Effect of depth of penetration of partial cutoff on seepage reduction for a
homogeneous isotropic foundation (prepared by WES)
EM 1110-2-1901
30 Sep 86
a. Partial cutoff
b. Complete cutoff
c. Relationship between quantity

of seepage and depth of
penetration of partial cutoff
Figure 9-3. Efficiency of a compacted backfill trench partial cutoff

in reducing the quantity of underseepage (courtesy of John Wiley
and Sons155)
9-8
EM 1110-2-1901
30 Sep 86
a. Flow net for no cutoff
b. Flow net for partial cutoff
c. Position of line of seepage for various values

of penetration
Figure 9-4. Effect of depth of penetration of partial cutoff on

the height of the line of seepage in the downstream embankment
and exit gradient at the toe for a homogeneous isotropic
154
foundation (courtesy of John Wiley and Sons )
9-9
EM 1110-2-1901
30 Sep 86
Impervious upstream Compacted back Concrete wall or

blanket fill or slurry steel sheetpiling
trench
Figure 9-5. Rate of underseepage loss for impervious

upstream blanket, compacted backfill trench or slurry
trench, and concrete wall or steel sheetpiling with
defects (courtesy of American Society of Civil
218
Engineers )
Material and compaction requirements are the same as those for the impervious
section of the dam (EM 1110-2-1911). When constructing a complete cutoff (see
para 9-4a), the trench must fully penetrate the pervious foundation and be
carried a short distance into unweathered and relatively impermeable foundation
soil or rock. To ensure an adequate seepage cutoff, the base width should be
at least one-fourth the maximum difference between the reservoir and tailwater
elevations but not less than 20 ft, and should be wider if the foundation
material under the cutoff is considered marginal in respect to imperviousness.
As previously mentioned (see para 9-4b), high seepage gradients occur along the
base of the cutoff and on its downstream face in both the foundation and
embankment zones. Suitable filters must be provided (see Appendix D for design
of filters) to prevent piping of soil at these interfaces. The trench
9-10
EM 1110-2-1901
30 Sep 86
Figure 9-6. Relationships among the length of the

impervious upstream blanket, the thickness of the
compacted backfill trench or slurry trench, or the
area of defects in the concrete wall or steel sheet-
piling, for a given rate of underseepage loss
218
(courtesy of American Society of Civil Engineers )
excavation must be kept dry to permit proper placement and compaction of the
impervious backfill. Dewatering systems of wellpoints or deep wells are
generally required during excavation and backfill operations when below
groundwater levels (TM 5-818-5). Because construction of an open cutoff trench
with dewatering is a costly procedure , the trend has been toward use of the
slurry trench method of construction (EM 1110-2-2300 and Cedergren 1977).
d. Slurry Trench.
(1) Introduction. When the cost of dewatering and/or the depth of the
pervious foundation render the compacted backfill trench too costly and/or
impractical, the slurry trench cutoff may be a viable method for control of
9-11
EM 1110-2-1901
30 Sep 86
underseepage. Using this method, a trench is excavated through the pervious

foundation using a sodium bentonite clay (or Attapulgite clay in saline water)
and water slurry to support the sides. The slurry-filled trench is backfilled
by displacing the slurry with a backfill material that contains enough fines
(material passing the No. 200 sieve) to make the cutoff relatively impervious
but sufficient coarse particles to minimize settlement of the trench forming a
soil-bentonite cutoff (sometimes called American method). Alternatively,
cement may be introduced into the slurry-filled trench which is left to set or
harden forming a cement-bentonite cutoff (sometimes called a grouted diaphragm
wall or Coulis wall or European method). The slurry trench cutoff is not
recommended when boulders, talus blocks on buried slopes, or open jointed rock
exist in the foundation due to difficulties in excavating through the rock and
slurry loss through the open joints. Where a slurry trench is relied upon for
seepage control, the initial filling of the reservoir must be controlled and
piezometers located both upstream and downstream of the cutoff must be read to
determine if the slurry trench is performing as planned. If the cutoff is
ineffective, remedial seepage control measures (see Chapter 12) must be
installed prior to further raising of the reservoir pool (KM 1110-2-2300).
(a) History of Use. The first use of the slurry trench method of con-
struction was by the U. S. Army Engineer District, Memphis, in September 1945,
to form a partial cutoff along the Mississippi River levee on the Arkansas
side of the river just below Memphis, Tennessee (Clay 1976 and Kramer 1946).
The idea for the project probably evolved from the use at that time of puddle
clay trench cutoffs combined with the use of drilling mud for advancing
borings. A paddle wheel mixing device was constructed for making slurry from
native clays. Trenches were dug to a 20-ft depth using a trenching machine and
to a 35-ft depth using a dragline with a 100-ft boom and 2-cu-yd bucket.
Backfill was mixed in windrows at the site from hauled-in clay gravel and
native materials and pushed into the trench by a bulldozer when the length of
the trench was equal to about twice the trench depth. BG Hans Kramer foresaw
the use of the slurry trench method for the construction of cutoffs for earth
dams. It is amazing that after 38 years, the technique is still about the same
as it was when first developed by the Memphis District. A soil-bentonite
cutoff was constructed under the Kennewick Levee adjacent to the Columbia River
as part of the McNary Dam Project in Washington by the Walla Walla District in
1952 (Jones 1961). The first application of a soil-bentonite slurry trench
cutoff for control of underseepage at a major earth dam was at Wanapum Dam on
the Columbia River in Washington in 1959 (La Russo 1963). Subsequently, soil-
bentonite cutoffs have been used for control of underseepage at a number of
dams as shown in table 9-2. The cement-bentonite slurry trench cutoff was
first used to tie into the abutment zones at the Razaza Dam on the Euphrates
River in Iraq in 1969 (Soletanche 1969). Subsequently, cement-bentonite cut-
offs were installed as remedial seepage control measures through the embank-
ment and foundation of four existing dams in Mexico from 1970 to 1972
(Soletanche 1970, 1971, 1971-1972, 1972). As shown in table 2, the first
cement-bentonite cutoff in the United States was constructed at the Tilden
Tailings Project to store tailings from the Tilden Mine in Michigan in 1976
(Meier and Rettberg 1978). The first cement-bentonite cutoff constructed at a
dam on a river retaining a reservoir in the United States was completed in
1978 at the Elgo Dam (formerly the San Carlos Dam) in Arizona (Anonymous 1978
and Miller and Salzman 1980).
9-12
EM 1110-2-1901
30 Sep 86
Table ?-2. Comparison of Slurry Trench Cutoffs
Max. Hal<.
Differential
!late Head Width De~th
Pro1ect Location Owner Constructed Foundation Uaterial ft __!!._ __!!._ Location ~ Reference
Soil-Bentonite Slurrv Trench
Kennewick Levee, Columbia River .. Corps of Engineers 1952 Sandy or silty gravel with 15 22 Center of 2. 5 Jones 1961
t-tcNa.ry Dam Wash. z:ones of open gravel dam
Wou>aj>ua n- Col.-hi.a Uve.r. Pllfllie UUlUy 1962 S<md, g1"3"ftls ...,d. g,;ave!.ly 88 10 80 Center o£ 8.8 L.a Russo 1963
wasn. District 2 .. Grant sands underlain by open da10
County, Wasb. wrk gravels
Wells Dam Columbia River, l'ublic Utility 1964 Pervious gravels 70 80 Center of 8.8 Jones 1967
Wash. District 1. dam
Douglass County,
Wasb.
Yards Creek Hew Jersey Public Service 1964 Sands~ gravels, cobbles, 55 8 40 Center of 6.9 Jones 1967
Lo""r Reservoir Electricity and and boulders daa
Gas
Comanche Dam - Moke1UII!IIIe East Bay 1966 Opper stratum of clayey 45 95 Upstream 5.6 Anton and
Dike 2 River, CaUf. )!unicipal District silts, silts, and clayey of toe of Dayton 1972
sands .. Lower stratum of da..
sand over zone of gravel
\lest Point Dat'l Chattabocchie Corp~ of Eng1neers 1966 Upper stratum of alluvial 61 60 Upstream 12.2 u.s. Army
River, Ca. soil~ alternating layers toe of dam Engr~ Dist,
.and Ala.,. of clay. si~t, sanO, a.ncl Savannah

gfa'Vel. Lower stratum. of 1968, 1979
residual soil, brown
oHty sand
Saylorville Des: !ok>ines Corps of Engineers 1969 Surface zone of sandy clay. 93 65 Upstrf'am 11.6 Harza 1965,
llaa River, Iowa Lower st.rattm of sand and toe of dam u.s. Army
gravel Engr. Dist.
Rock Island
1978
Mill Site ferron Creek. Soil Conservation 1969 Sandy gravel 73 68 Upstrf!am 9.1 H.anoon 1972
llaa Utab Service toe o:!. dam
Plum Creek Plum Creek., Soil Conservation 1973 Sands, gravelly sands. 54 40 Ce.ntet:' of 10.8 Knabach and
llaa Wis. Service silty gravels, silty sands dam Dingle 1914
Upper Big Upper Big Soil Cortservet ion 1978 Sands and gravels 62 40 Center of 15.5 Bloom, Dynes,
&lue Daa Blue R.iver, Service dam Glol!sett
Ind. 1979
Addicks lla,.(a) Buffalo Bayou, Co-rps of Engineers 1979 Silty aands 30 70 Varies 10.0 u.s. Anny
Tex~ Engr .. Dist,
Galveston
Barker Dall(a) Buffalo Bayou, Corps of Engineers 1979 Sandy clay, clayey sand, 24 55 Cenr:er of 8.0
ailty sand 1983 (sallie
Tex .. for Barker Da11)
Celaent-Beatonite Slurrl Trench

Sand interspersed witn gravel 100(b) Upstream 51).0 Meier and
Tilden Tilden !line, Cleveland-Cliffs 1976 80
Tailings Mieb. Iron Co11pany toe of dam ll.ettberg
1918
Elgo Dam San Carlos Sat'l Carlos Apac.he 1978 Sands and gravels 66 68 Center of 22.0 tliller and
(formerly lU.ver, Ari•- Tribe dam Salzman 1980
San Carlos
Diu&)
(a) Slurry trench c.ucof[ ins called aa reaedial seepage control for exlatina da11 (aee Chapter 12).
(b) Planrted maximum differential head wben ta1.11ng;a dam is at final height (45 ft in February 1983}.
9-13
EM 1110-2-1901
30 Sep 86
(b) Patents. The soil-bentonite slurry trench cutoff was covered by

United States Patent No. 2,757,514 dated August 7, 1956, "Method of Forming an
Impermeable Wall in the Terrain," in the name of Harold T. Wyatt. This patent
expired in 1973. The cement-bentonite slurry trench cutoff is covered by
United States Patent No. 3,759,044 dated September 18, 1973, "Method of Earth
Wall Construction Using Cementitious Bentonite Mud," in the name of Claude
Caron and Jean Hurtado, both of France, assignor to Soletanche, Paris, France.
(c) Comparison of Soil-Bentonite and Cement-Bentonite Slurry Trench

Cutoffs. A comparison of soil-bentonite and cement-bentonite slurry trench
cutoffs is given in table 9-3. The soil-bentonite slurry trench cutoff is
generally the most economical if the cost of backfill is not prohibitive. For
deep cutoffs where the foundation is prone to failure during excavation, the
cement-bentonite slurry trench cutoff is more applicable.
(d) Location of Cutoff. Normally, the slurry trench should be located

under or near the upstream toe of the dam (EM 1110-2-2300). An upstream
location provides access for future treatment provided the reservoir could be
drawn down and facilitates stage construction by permitting placement of a
downstream shell followed by an upstream core tied into the slurry trench. For
stability analysis, a soil-bentonite slurry trench cutoff should be considered
to have zero shear strength and exert only a hydrostatic force to resist
failure of the embankment (U. S. Army Engineer District, Savannah 1968). If
the slurry trench is located under a central core, consolidation of the slurry
trench backfill combined with arching of the core material immediately above
the slurry trench may result in the opening of a cavity under the dam with
possible leakage along the contact. If a central location for the slurry
trench is dictated by other factors, some possible benefits are obtained by
flaring the top of the trench to provide a transition between the cutoff and
the core. Also, the slurry trench can be constructed and allowed to settle
before placement of the embankment (Jones 1967 and Jansen 1968). When the
groundwater table is located some distance beneath the ground surface, it is
usually more economical to excavate a conventional open trench with stable side
slopes with the trench bottom a few feet above the highest level of ground
water expected during the construction period, as was done at West Point Dam,
Alabama and Georgia (U. S. Army Engineer District, Savannah 1968). The bottom
of the open trench provides a working level from which the slurry trench may be
constructed. Also, this prevents the problem of significant amounts of slurry
being lost into the excavated trench above the ground-water table. If the
ground-water table is located near the ground surface, compacted impervious
fill should be placed in order to raise the level of the slurry trench to
maintain the level of slurry in the trench a sufficient distance above the
ground-water level (Jones 1967).
(e) Stability of the Trench. In cohesionless soils the penetration of

the slurry into the wall of the slurry trench excavation forms a relatively
impervious filter cake on which the hydrostatic pressure of the slurry can
act. The depth of penetration ranges from 1 to 3 in. in sand, 3 to 6 in. in
sand and gravel, and up to 12 in. in gravel, depending on the gradation. The
main stabilizing force supporting the slurry trench excavation is the hydro-
static pressure exerted on the trench walls. For a slurry trench excavated in
a homogeneous clay, remaining open only for a few days to permit placement of
9-14
Table 9-3. Comparison of Soil-Bentonite and Cement-Bentonite Slurry Trench Cutoffs(a)
Item Soil Bentonite Slurry Trench Cutoff Cement-Bentonite Slurry Trench Cutoff
1. Excavation Long slope of backfill requires trenching con- Construction sequence is more flexible to meet site
tinuously in one direction. constraints.
Can accommodate interruptions in construction. Excavation for each panel should progress uninter-
rupted so each panel is completed before the slurry
begins to set.
Difficult to work with coarse gravels, with Easier to remove coarse gravels with clam shell
backhoe and/or dragline, which settle out and mounted on a rigid Kelly bar, from trench trench
accumulate on the trench bottom reducing the bottom. Better suited for excavation in areas
certainty of a tight seal. prone to failure.
Wider trench required to prevent segregation dur- Narrower trench (2 - 3 ft) may be used.
ing backfilling (> 3 ft) and prevent piping
failure of the sl~rry trench backfill into the
adjacent soil downstream of the cutoff (1 ft
width for each 10 ft of differential head)
2. Slurry Mixture of bentonite and water used to support Mixture of cement, bentonite, and water used to
excavation during trenching. Slurry displaced by support ~xcavation foe panel and later sets to form
backfill is desanded and reused. Upon completion permanent backfill. No disposal or desanding of
of trench, unused slurry must be disposed of in slurry required.
an environmentally acceptable fashion.
9-15
' 3. Backfill Mixture of slurry plus excavated material from None (see 2). Therefore•not dependent on availa-
trench and/or imported select backfill materials bilityor quality of soil for backfill and more
.suitable for work in confined areas.
4. Properties of completed cutoff:

7
a. Permeability = 10- em/sec (~ 1 percent bentonite) = 10- 6 em/sec
b. Strength Little field data available - assumed zero for 15 - 20 lbs/sq in. unconfined compressive strength.
design purposes. Will support construction loadings or future
structures ftfter several weeks.
c. Consolidation Some consolidation with time. No significant consolidation with time.
5. Susceptible to attack by No Yes - cement may be.

sulphates or other chemicals
in groundwater
6. Relative cost Generally lower if cost of backfill is not Generally higher due to cost of cement. More
EM 1110-2-1901
prohibitive (see 3) competitive as depth of trench increases due to
narrower trench width required.
7. Patent applicable No (expired 1973) Yes (United States Patent No. 3,759,044 dated
30 Sep 86
September 18, 1973)
(a) Courtesy o f Eng i neer1ng

· construct1on
· I n t erna t lana
· 1 , Inc . 132 • 249 , '~rl· can
~·~ Soc1· ety of C1·v1·1 Engineers 22 2,
- - • 1 • • • 221
EM 1110-2-1901
30 Sep 86
backfill material, the factor of safety against instability is (Nash and Jones
1963)
(9-5)
where
F = factor of safety
Cu = undrained cohesion
H = depth of the trench
= unit weight of the soil
= unit weight of the slurry
For a slurry trench excavated in dry cohesionless soil (Nash and Jones 1963)
(9-6)
where angle of internal friction.
For a slurry trench excavated in a saturated cohesionless soil with the

ground-water table and slurry level in the excavation both at the ground
surface (Nash and Jones 1963)
(9-7)
where
= effective unit weight of the soil
= effective unit weight of the slurry
= effective angle of internal friction
For arbitrary levels of ground water and slurry in cohesionless soil, as shown
in figure 9-7a, a slightly conservative (neglects arching effect of short
9-16
EM 1110-2-1901
30 Sep 86
trenches and stabilization of the soil adjacent to the trench face due to
slurry penetration and gelation) estimate of the slurry density required to
ensure stability of the trench is (Morgenstern and Amir-Tahmasseb 1965)
(9-8)
where
n = defined in figure 9-7a
= unit weight of the slurry
= unit weight of the soil
= angle of inclination of the wedge of soil at the point of slipping,

in practice assumed to be equal to 45" +
effective angle of internal friction
m = defined in figure 9-7a
Equation 9-6 may be solved by use of the nomograph shown in figure 9-7b
(Duguid et al. 1971).
(2) Slurry. The slurry has three basic functions in slurry trench
construction (Ryan 1977):
• To hold the trench open and maintain a stable excavation.
• To be fluid enough to permit passage of the excavating equip-

ment and to allow placement of the backfill (for the cement-bentonite slurry
trench, there is no backfill).
• To form a filter cake to enhance the low permeability of the

completed trench.
(a) Materials. As a general rule sodium montmorillonite in powder form

(Wyoming-type bentonite) is used for slurry trench construction. However,
when salt water is present Attapulgite clay is used to avoid flocculation
(Spooner et al. 1982). Specifications for both bentonite and Attapulgite are
given by the American Petroleum Institute (American Petroleum Institute 1981).
Each shipment of bentonite or Attapulgite should be checked for compliance with
the specifications. At Saylorville Dam, Iowa, changes in slurry properties
were traced to lower quality bentonite which was mined from different beds
(U. S. Army Engineer District, Rock Island 19788). No chemically treated
9-17
EM 1110-2-1901
30 Sep 86
a. Slurry trench typical section
b. Nomograph for slurry

trench stability
Figure 9-7. Determination of slurry

trench stability in cohesionless
soil (courtesy of National Research
168
Council of Canada )
9-18
EM 1110-2-1901
30 Sep 86
bentonite should be used for slurry trench construction. The pH of the water
used for mixing with the bentonite should equal 7.0 ± 1.0. Water hardness
should not exceed 50 ppm (parts per million). Total dissolved solids should
not exceed 500 ppm. The amount of oil, organics, or other deleterious sub-
stances should be limited to no more than 50 ppm each (Stanley Consultants,
Inc., and Woodward-Clyde Consultants 1977). If the use of poor quality water
cannot be avoided, it will require more bentonite and longer mixing times to
achieve the desired properties.
(b) Mixing. The methods used to prepare the hentonite slurry vary with
project size and layout. Always add the bentonite to the water, never the
water to the bentonite. For small jobs the batch system is used where
specific quantities of water and bentonite are placed in a tank and mixed at
high speeds with a circulationpump or paddle mixer and the slurry is dis-
charged into the trench. Mixing is usually complete in a matter of minutes
for the 2- to 5-cu-yd batch produced by this method. The most commonly used
method is the flash or Venturi mixer and circulation ponds which is well
adapted for bulk handling of large slurry volumes. A flash mixer introduces
dry bentonite into a turbulent water jet which discharges into a low speed
circulation pond. When Marsh Funnel viscosity readings stabilize, the slurry
is stored in a second pond prior to using the trench (Spooner et al. 1982 and
D'Appolonia 1980).
(c) Properties. Tests of bentonite quality must be conducted for each

rail car or truck load delivered. The minimum acceptable viscosity of a slurry
made with the bentonite is 40 seconds Marsh funnel viscosity at 65° F. The
fresh slurry shall have a minimum Marsh funnel viscosity of 40 seconds at 65° F
and a pH of from 7 to 10, a bentonite content of from 3 to 7 percent by dry
weight (depending on the grade of bentonite), a unit weight of from 1.0 to
1.04 g/cm3 (about 65 lb/ft3), and the filtrate or water loss shall not be
greater than 20 at 100 lb/in.2 x cm3 in 30 minutes (Spooner et al. 1982 and
U. S. Army Engineer District, Savannah 1968). The slurry in the trench should
be sampled at least twice daily with samples taken from the top of the trench
and at 10-ft vertical, 50-ft horizontal intervals along the trench center line.
During all stages of construction the minimum acceptable viscosity of the
slurry shall be 40 seconds Marsh funnel viscosity at 65° F. The minimum
in-trench slurry unit weight is based upon trench stability considerations (see
Equation 9-8 and Figure 9-7). The maximum in-trench slurry density is 85/ft3
to avoid buildup of sediment beyond the slurry capacity to hold it in suspen-
sion in the trench during excavation (Clough 1978).
(d) Quality Control Testing. In order to mix and maintain a proper

slurry to hold the trench open during excavation and form a filter cake for
the soil-bentonite slurry trench, quality control testing must be performed.
The property, frequency, standard (if any), and specified value for slurries
and their components are given in table 9-4. The quality of the mixing water
used can influence the slurry trench characteristics. If the specific values
for mixing water quality are not met, the bentonite will flocculate and settle
out and not form the filter cake on the sides of the trench. Poor quality
mixing water will increase the set time for cement-bentonite slurry trench
cutoffs. The bentonite is tested to be sure it will have the minimum viscos-
ity required to keep the soil in suspension. The slurry is tested both after
9-19
30 Sep 86
EM 1110-2-1901
Table 9-4. Summary of Quality Control Tests for Slurry Trench Cutoff
Category or llaterial Property Frequency Standard Specified Value

Mixing Water pH Once per month API RP 13B(a) 7.0 ~ 1.0
Hardness Once per month API RP 13B 2 50 ppm
Total dissolved solids Once per month API RP 13B 2 500 ppm
Oil, organics, etc. Once per month API RP 13B < 50 ppiit each
Bentonite Viscosity Each railroad car or truck load API RP 13A > 40 seconds at 65°F
Bentonite slurry (after Viscosity Twice per day API RP 13B > 40 seconds at 65°F
mixing) pH Twice per day API RP 13A 7 - 10
Bentonite content Twice per day None(b) 3 - 7 percent
3
Unit weight Twice per day API RP 13B 1.01 - 1.04 g/cm
3
Filtrate or water loss Twice per day API RP 13B > 20 cm at 100 psi in 30 min
Bentonite slurry (in trench) Viscosity Twice per day API RP 13B > 40 seconds at 65°F
Unit weight Twice per day API RP 13B minimum(c)_ 85 pcf
Cement Slump Twice per day ASTM C-143 (d) 2 - 5 in.

9-20
Cement-bentonite slurry Viscosity Twice per day API RP 13B 40-50 seconds at 65°F
(after mixing) pH Twice per day API RP 13A 10 - 13
Bentonite content Twice per day None 4 - 7 percent
1.03 - 1.04 g/c~
3
Unit weight Twice per day API RP 13B
3
< 165 cm at 100 psi in 30 min
Filtrate or water loss Twice per day API RP 13B
Cement/water ratio Twice per day None 0.17 - 0.20
(a) American Petroleum Institute specification.

(b) Can be checked by evaporating the water from a known volume of slurry and weighing the remaining solids
(c) Minimum in trench slurry unit weight is based upon trench stability considerations.
(d) American Society for Testing and Materials Standard.
EM 1110-2-1901
30 Sep 86
mixing and in the trench to determine that it is dense enough to stabilize the
trench, but not so dense as to cause the backfill to settle too loosely, and
that it has sufficient viscosity to maintain cuttings in suspension (Stanley
Consultants, Inc., and Woodward-Clyde Consultants 1976).
(3) Excavation, Mixing, and Backfilling.
(a) Excavation. The preferred method of trench excavation depends upon

the required depth of the slurry trench cutoff, the nature of the subsurface
materials, and access to the trench at the ground surface. It is important to
ensure that the equipment used can maintain a continuous excavation line to
the total depth required. At depths less than about 50 ft, backhoes are gen-
erally the most rapid and least costly excavation method. Modified backhoes
with an extended dipper stick, modified engine, and counter-weighted frame can
excavate to about 80 ft deep. Draglines with weighted (> 10,000 lb) buckets,
which have been used in the depth range 60 to 80 ft, have been replaced by more
efficient extended backhoes. The clamshell bucket can excavate to depths in
excess of 150 ft. The clamshell may be mechanically operated attached to a
crane or hydraulically operated attached to a Kelly bar. On larger jobs, the
backhoe may be used to excavate the first 50 ft followed by a clamshell bucket
to excavate the primary and secondary panels to the impervious zone. Regard-
less of the equipment used, it should be capable of excavating a trench of the
desired width in a single pass in order to obtain a fairly consistent trench
width. The bucket used should be nonperforated to allow retention and removal
of sand particles from the trench. The continuity of the trench is tested by
passing the bucket or clamshell of the excavating tool vertically and horizon-
tally along each segment of the trench before it is backfilled. Whatever
excavation method is used, it is important that good communications are main-
tained with the operator of the excavation equipment since abnormalities in
the trench excavation are usually noticed first by the equipment operator
(D'Appolonia 1980; Spooner et al. 1982; Bloom, Dynes, and Glossett 1979; Case
International Company 1982; and Winter 1978). Soil-bentonite slurry trench
cutoffs are excavated in a continuous trench as shown in figure 9-8a, while
cement-bentonite slurry trench cutoffs are excavated in a continuous trench or
in short sections or panels as shown in figure 9-8b. The cement-bentonite
slurry begins to harden within 2 to 3 hours after mixing. Alternate panels
are excavated under a cement-bentonite slurry and then allowed to partially
set. Intervening panels are excavated also under a cement-bentonite slurry
and a portion of the initial panel ends are removed to ensure continuity
between adjacent panels. Construction delays can cause problems in setup of
cement-bentonite slurry trench cutoffs because continued agitation of the
cement-bentonite slurry (more than 24 hours) reduces the ability of the cement
to set (Spooner et al. 1982).
(b) Bottom Treatment. The aquiclude used for the slurry wall founda-
tions should be continuous, and relatively free of fractures and other pervious
zones. The cutoff wall should extend a minimum of 2 ft into clay (or 1 ft into
rock) to prevent weathered zones, desiccation cracks, or other geological
features from permitting seepage under the cutoff (Spooner 1982). As the
trench is excavated, heavier soil particles such as sand and gravel fall to the
bottom of the trench. The amount of sand accumulation on the trench bottom
depends upon the coarseness of the strata being excavated as well as the
9-21
EM 1110-2-1901
30 Sep- 86
a. Soil bentonite cutoff
b. Cement bentonite cutoff
Figure 9-8. Construction procedure for soil-bentonite

and cement-bentonite cutoffs (courtesy of American
222
Society of Civil Engineers )
excavation technique used. Although this sand layer may not have a direct
effect on trench stability, it may adversely affect the permeability of the
slurry trench cutoff wall (Spooner 1982). An air lift pump should be used to
remove the sand and gravel particles from the trench bottom prior to backfill-
ing. When the slurry trench is keyed into a soil aquiclude after the trench
bottom has been cleaned thoroughly, a minimum of one split-spoon sample shall
be taken every 50 ft along the length of the trench to determine if additional
excavation is required (Winter 1978).
(c) Backfill Mixing and Placement. A minimum of one day is required

between trench excavation and backfilling in order to develop a low permeabil-
ity filter cake on the trench walls (D'Appolonia 1980). Stockpiled material
from the trench excavation and/or material from borrow areas are mixed and
blended by windrowing, disc harrowing, bulldozing, or by blading to remove
lumps of clay, sand, or gravel. The backfill is then sluiced with slurry
(mixing with water shall not be permitted) and just prior to placement has a
consistency of a wet concrete with a slump of 5 in. ± 1 in. tested in
9-22
EM 1110-2-1901
30 Sep 86
accordance with ASTM C-143 (1) (Winter 1978 and Ryan 1976). The backfill is
placed continuously from the beginning of the trench in the direction of the
excavation to the end of the trench. Free dropping of the backfill material
through the slurry would produce segregation and is not allowed. Depending on
the steepness of the excavated slope, it may be necessary to lower the initial
backfill to the bottom of the trench with a crane and clamshell bucket until a
slope at the angle of repose of the backfill has been formed from the bottom of
the trench to the top of the excavation. The toe of the backfill slope is kept
to within 50 to 150 ft of the leading edge of the excavation to minimize the
open length of the slurry-supported trench while allowing enough space behind
the excavation for cleaning the trench bottom. Additional backfill is placed
by a bulldozer in such a manner that the backfill enters the trench and slides
progressively down the slope of the previously placed backfill and produces a
slope ranging from 1V on 5H to 1V on 10H. The slope of the backfill shall be
measured with soundings starting at the toe of the backfill in the bottom of
the trench and progressing up the backfill slope at 25-ft horizontal intervals.
A set of soundings shall be made at least for every 25 ft horizontal advance-
ment of backfill placement. Once the natural slope of the backfill is estab-
lished during initial placement of the backfill, the slope should remain nearly
the same. If the slope, or a portion of the slope, suddenly gets steeper, it
could be an indication that sediment is being trapped or that the backfill has
a pocket of relatively clean material (slurry not mixed in properly or was
washed out). If the slope suddenly gets flatter, it could indicate that a
pocket of slurry was trapped in the backfill or that the backfill does not
contain sufficient sand or coarser material (Stanley Consultants, Inc., and
Woodward-Clyde Consultants 1976; Winter 1978; and Ryan 1976).
(d) Temperature During Construction. The mixing and placing of backfill

shall be limited to days when the air temperature is not less than 20" F. Even
though the surface of the slurry trench freezes overnight, there will be no
difficulty breaking through the surface ice and continuing excavation during
the day. Frozen backfill or pieces of ice must never be placed in the trench
(U. S. Army Engineer District, Rock Island 1978a and Jones 1967).
(e) Protection of Top of Trench. The top of the completed slurry

trench cutoff should be immediately protected with a temporary 2- to 3-ft-
thick blanket of moist impervious fill material to prevent drying of the
backfill and formation of shrinkage cracks along which paths of seepage could
easily develop. The layer is temporary and is removed once the embankment
construction is started (Jones 1967; Stanley Consultants, Inc., and Woodward-
Clyde Consultants 1977).
(4) Soil-Bentonite Slurry Trench Cutoff.
(1) American Society for Testing and Materials standard. If a desirable back-
filled slope (1V on 5H to 1V on 10H) cannot be maintained in the trench
with a 5 in. ± 1 in. slump, the slump may be altered to meet construction
conditions. Such was the case at the soil-bentonite slurry trench cutoff
constructed at W. G. Huxtable Pumping Plant, Marianna, Arkansas (U. S.
Army Engineer District, Memphis 1978).
9-23
EM 1110-2-1901
30 Sep 86
(a) Design Considerations. The primary design parameters are blowout

requirements, permeability, strength, and compressibility. The backfill mate-
rial must not blow out into the surrounding pervious foundation under the
maximum differential hydraulic head that will act on the slurry trench. The
-7
permeability is usually sufficiently low (-10 cm/sec for > 1 percent bento-
nite) to reduce the seepage through the slurry trench cutoff to an acceptable
value. Under most conditions, the only strength requirement for the slurry
trench cutoff is to approximate the strength of the surrounding ground. The
compressibility of the slurry trench cutoff, once consolidated under its own
weight (usually within 6 months after placement), should be compatible with the
compressibility of the surrounding ground to minimize differential movement of
the dam and resultant stress concentrations in the embankment or its foundation
(Ryan 1976 and Xanthakos 1979).
(b) Blowout Requirements. Once the slurry trench is installed, the dam
has been constructed, and the reservoir filled, there is a substantial differ-
ential head acting on the slurry trench (see table 9-2 for typical values).
Depending upon the characteristics of the backfill material and pervious foun-
dation, the hydraulic gradient acting across the slurry trench may be suffi-
cient to cause blowout or piping of backfill material into the surrounding per-
vious foundation. This is especially critical when the foundation contains
openwork gravel where the piping process could result in the formation of
channels and cavities that may breach the slurry wall. Based upon laboratory
tests conducted on widely graded gravel containing no sand, the blowout gradi-
ent ranges from 25 to 35, depending on the properties of the backfill mate-
rial (La Russo 1963 and Nash 1976). The factor of safety against blowout is
(9-9)
where
F = factor of safety against blowout
i = allowable hydraulic gradient from laboratory blowout tests

allowable
i = actual hydraulic gradient existing on slurry trench
actual
Substituting for the actual hydraulic gradient
(9-10)
where
= maximum differential hydraulic head acting on the slurry trench
w = slurry trench width
9-24
EM 1110-2-1901
30 Sep 86
and using a factor of safety of 3 and an allowable hydraulic gradient of 30 in

equation 9-9 gives
(9-11)
If the pervious foundation contains openwork gravel, the width of soil-

bentonite slurry trench required to prevent blowout failure may be estimated
from equation 9-11. Further refinements on the trench width would require con-
ducting laboratory blowout tests (as described by Xanthakos 1979).
(c) Permeability. For design purposes the permeability of the soil-

bentonite slurry trench cutoff is based on the backfill material only
(Xanthakos 1979). The permeability of the slurry trench is a function of both
the filter cake that forms on the trench walls and the backfill material. The
contribution of the filter cake and the backfill depends on the relative
permeability and thickness of the two materials. The horizontal permeability
of the soil-bentonite slurry trench is (D'Appolonia 1980)
(9-12)
where
k = permeability of slurry trench
tb = backfill thickness
k b = backfill permeability
t = filter cake thickness
kc = filter cake permeability
The permeability of the backfill material can be determined in a laboratory

permeability test (EM 1110-2-1906). The thickness of the backfill is selected
in design (see figure 9-6). The ratio kc/tc is determined from the filter
press test (American Petroleum Institute 1982) using various formation cake
pressures as shown in figure 9-9a. For a range of practical applications, the
ratio kc/tc varies from 5 x 10-9/sec to 25 x 10-9/sec as shown in fig-
ure 9-9b. Figure 9-10 shows the permeability of a soil-bentonite slurry trench
cutoff wall 80 cm (about 2-1/2 ft) thick for various values of backfill perme-
ability and ratios of kc/tc . As shown in figure 9-10, the slurry trench
9-25
EM 1110-2-1901
30 Sep 86
a. Schematic of filter press test apparatus
b. Relationship among filter cake permeability, filter cake formation

pressure, and time
Figure 9-9. Determination of filter cake permeability (courtesy of American

163
9-26
EM 1110-2-1901
30 Sep 86
Figure 9-10. Permeability of a soil-bentonite slurry trench cutoff

wall 80 cm thick for various values of backfill and filter cake
permeability (courtesy of American Society of Civil Engineers 163)
9-27
EM 1110-2-1901
30 Sep 86
permeability is controlled by the backfill when the backfill permeability is

low and by the filter cake when the backfill permeability is high. Also, the
-6
slurry trench permeability has an upper limit of about 10 cm/sec even for
very permeable backfill due to the thin low permeability filter cake
(D'Appolonia 1980).
(d) Shear Strength. Soil-bentonite slurry trench cutoffs are difficult

to sample because of their soft nature and very little data are available on
the shear strength of soil-bentonite slurry trench backfill material. For
design purposes, in conducting the stability analysis of the embankment and
foundation, the shear strength of the backfill material is assumed, to be zero.
However, the shear strength of the backfill material does increase with time
due to consolidation and thixotropy. At time of placement, the backfill mate-
rial will stand on a slope ranging from 1V on 5H to 1V on 10H. This improves
to about 1V on 2H with time (Ryan 1976 and D'Appolonia 1980). The results of
shear strength tests (see table 9-5) on undisturbed samples taken from the
soil-bentonite slurry trench at Saylorville Dam, Iowa, show that the undrained
shear strength of the slurry backfill about a year after placement was 0.10 to
0.12 tons/sq ft (U. S. Army Engineer District, Rock Island 1978b).
(e) Compressibility. The compressibility of the soil-bentonite slurry

trench backfill material depends primarily on the percentage of granular
particles in the gradation as shown in figure 9-11. Low permeability and low
compressibility are contradictory requirements because the plastic fines
required for low permeability result in higher compressibility. Relatively
low compressibility results when there is sufficient granular material in the
backfill to allow grain-to-grain contact between the granular particles
(D'Appolonia 1980).
(f) Mix Design. The gradation of the backfill for the soil-bentonite
slurry trench is selected by conducting permeability, shear strength, and com-
pressibility tests on a range of materials including soil to be excavated from
the trench. Such a procedure was followed in the mix design for the backfill
of the soil-bentonite slurry trench installed for remedial seepage control at
Addicks Dam, Texas (U. S. Army Engineer District, Galveston 1977c). The allow-
able range set on the gradation of the backfill should produce a material which
contains enough fines to reduce the seepage through the slurry trench cutoff to
an acceptable level and sufficient coarse particles to approximate the strength
and compressibility of the surrounding ground. If sufficient fines are not
present in material excavated from the trench, borrow sources should be identi-
fied or alternatively a higher bentonite content specified for the backfill.
If sufficient coarse particles are not present in material excavated from the
trench, approved sources should be identified for obtaining natural sound,
hard, durable sand and gravel. Crushed limestone, dolomite, or other crushed
calcareous materials should not be used. The maximum particle size of the
gravel shall be 3 in. and the material should be well graded.
(5) Cement-Bentonite Slurry Trench Cutoff.
9-28
Table 9-5. Summary of Shear Strength Data from Post-Construction Testing of Backfill Material from Soil-Bentonite
Slurry Trench Cutoff at Saylorville Dam, Towa(a)
Triaxial Com~reasion Test(b)

Atterberg Dry Moisture Time Ela~sed,.months g R
Boring Depth Unified Limits Density Content Construction Sampling 0 c •• c'
No. Station ft Soil Classification LL PL PI lb/cu ft % to Sam~ling to Testing ~ tons/sg ft ~ tons/sg ft
STU-1 58+90 26.7-28.9 Gravelly clayey
sand, SC 33 14 19 ll8.2 13.3 8.5 5 4.5 0.075
STU-I 58+90 39.2-41.4 sc
Clayey sand, 33 14 19 108.5 17.5 8.5 5 0 0.130
STU-2 64+00 14.5-16.6 Clayey sand, sc 36 15 21 103.5 21.0 8 5 0 0.075
STU-1 58+90 49.2-51.4 Clayey sand, sc 41 17 24 107.9 18.9 8.5 19 14.5 0.10
STU-2 64+00 32.o-34.3 Clayey sand, sc 33 14 19 114.1 14.9 8 19 19 0
sand, SC 34 15 19 120.1 16.9 ll.S 7.5 0 0.10

sand, SC 37 15 22 113.1 17.7 11.5 7.5 0 0.12
STU-7 87+80 17.o-l9.0 Gravelly clayey

sand, sc 32 17 15 113.5 17.6 11 6 0 0.12
Sandy clay, CL 31 18 l3 117 .o 13.1 ll.5 9.5 14.5 0
STU-6 82+90 35.4-35.6
Clayey sand, SC 28 14 14 112.9 19.2 11 11.5 17 0
26.6-28.8
9-29
STU-7 87+80
EM 1110-2-1901
30 Sep 86
( ) 106
a From U. s. Army Engineer District, Rock Island
(b) Conducted on undisturbed soil specimens 5 in. in diameter by 11 in. high; unconsolidated-undrained (Q) and consolidated-undrained with
·pore pressure measurements (R) shear tests.
EM 1110-2-1901
30 Sep 86
e PLASTIC FINES
0 NON-PLASTIC FINES
•
/
~
/
e / 1-D COMPRESSION
/e
0 / STRESS INCREMENT
0.5 TO 2.0 kg /cm2
0 ~~--~--~--~--~~~~--~--~--~--~~
0 .02 .04 .06 .08 .10 .12
COMPRESSION RATIO, Cc/ I+ e 0
80
70
/.
.~
60 0
50
0
/
40
0./ • ISOTROPIC COMPRESSION
30
20
:;t .
10
0
/ STRESS INCREMENT
0.5 T02.0 kg/cm2
0
0 .02 .04 .06 .08 .10 .12
COMPRESSION RATIO, Cc/ I+ e 0
Figure 9-11. Relationship between fines content and compres-

sibility of a soil-bentonite slurry trench backfill (courtesy
of American Society of Civil Engineers163)
9-30
EM 1110-2-1901
30 Sep 86
(a) General. If backfill for the slurry trench is not available or is

prohibitive in cost or if the cutoff is deep and the foundation is prone to
failure during excavation, the cement-bentonite slurry trench cutoff may be
more applicable (see table 9-3).
(b) Design Considerations. The primary design parameters are continu-

ity, set time, resistance to hydraulic pressure, permeability, shear strength,
and compressibility.
(c) Continuity. When cement-bentonite slurry trench cutoffs are con-

structed in panels rather than in a continuous trench, there is a possibility
for unexcavated portions of the trench to remain between the panels. To pre-
vent this the clamshell bucket is moved both vertically and horizontally
throughout each slot at the completion of slot excavation. Also, when the
connecting area between the initial and subsequent panels is excavated, a por-
tion of the adjacent set panels is removed to ensure that all intervening soil
has been excavated (Spooner et al. 1982).
(d) Set Time. The set time is important because of the construction
technique employed. After the cement-bentonite slurry in the first set of
panels has set, the areas between them can be excavated. A normal cement-
bentonite mixture begins to set after a few hours and has a consistency simi-
lar to lard after 12 hours. The second day the cement-bentonite slurry can be
walked on and final set is normally taken at 90 days (Ryan 1977).
(e) Resistance to Hydraulic Pressure. Once the slurry trench has been
completed, the embankment constructed, and the reservoir filled, there is a
substantial differential head acting on the slurry trench (see table 9-2 for
typical values). The time between completion of the slurry trench and reser-
voir filling is generally sufficiently long (> 90 days) to allow the cement-
bentonite slurry trench to develop its final set. The resistance of the
cement-bentonite material to withstand gradients comparable to those which
will exist in the field should be tested in the laboratory by subjecting
intact specimens which have developed full set to hydraulic pressure and
measuring the increase (if any) of permeability with time (Spooner et al. 1982
and Jefferis 1981).
(f) Permeability. Although there is some buildup of concentration near

the sides of the cement-bentonite slurry trench, the cement-bentonite does not
form a low permeability cake. The permeability of the slurry trench is a
function of the concentrations of cement, bentonite, sand, and gravel (sus-
pended during the excavation process) in the completed wall (Ryan 1977). The
amount of sand and gravel in the cement-bentonite trench cutoff may range from
10 to 60 percent by dry weight, depending on the foundation material and
method of construction, and generally increases with depth (Dank 1981). If
the trench is excavated under a conventional bentonite slurry which is then
replaced by a cement-bentonite slurry, the sand and gravel content will be low
(10 to 18 percent was measured on undisturbed samples taken from the San
Lorenzo Dam, El Salvador; Dank 1981). Alternatively, if the trench is exca-
vated under a cement-bentonite slurry which is left in the trench to set up and
form the cutoff, the sand and gravel content will be relatively high. Also, if
the trench is excavated under a cement-bentonite slurry, the slurry loss into
9-31
EM 1110-2-1901
30 Sep 86
the surrounding ground will be higher than normal and in some instances as
great as 100 percent of the trench volume (Xanthakos 1979). For design
purposes, specimens should be prepared from the cement-bentonite with varying
percentages of sand and gravel, cured for 28 days under consolidation pressures
existing in the field, and laboratory permeability tests conducted
(EM 1110-2-1906).
(g) Shear Strength. Cement-bentonite slurry trench cutoffs are more

easily sampled and tested than are soil-bentonite slurry walls. Also, speci-
mens of cement-bentonite may be cast in the laboratory and tested. The
results of shear strength tests (see table 9-6) on undisturbed samples taken
from the cement-bentonite slurry trench at Tilden Tailings Project, Michigan,
show the unconfined compressive strength about 6 months after placement
increased with depth ranging from 0.88 to 1.43 tons/sq ft (Dank 1981).
(h) Compressibility. Very little data are available on the compres-

sibility of cement-bentonite slurry trench material (Millet and Perez 1980).
The compressibility should decrease with increase in cement to water ratio
(provided the bentonite is fully hydrated with water prior to adding the
cement) and with increase in sand content (once the concentration of suspended
sand and gravel is sufficient to allow grain-to-grain contact between the
granular materials).
(i) Mix Design. The cement-bentonite slurry trench mixture consists of

water, bentonite, cement, set retarders as necessary, and sand and gravel
entering the trench as a by-product of the excavation. The bentonite should
be fully hydrated with water prior to adding the cement (Millet and Perez
1980). A retarder of the lignosulphite group may be added in small amounts
(0.1 percent) to delay the initial set to avoid hardening of the mix in the
panel before the excavation is completed (Xanthakos 1979). When low perme-
ability is required, the bentonite content of the slurry should be increased
(in the range from 3 to 6 percent by dry weight). Increased sand and gravel
in the slurry will result in an increase in permeability (Dank 1981). The
cement, sand, and gravel content are the chief factors in controlling the
strength and deformability characteristics of the slurry mix (see table 9-6 and
figure 9-12). Generally, the higher the cement to water ratio, the higher the
strength, and more brittle (lower failure strain) the cement-bentonite slurry
mix (Millet and Perez 1980). By varying the bentonite and cement quantities,
flexibility can be designed into the cement-bentonite slurry trench cutoff.
This is especially important if the dam is located at a site where strong
earthquake shocks are likely. The cement-bentonite slurry mix proportions
should be selected by conducting permeability, shear strength, and compres-
sibility tests on a range of materials including soil to be excavated from the
trench. Varying proportions of water, bentonite, cement, sand and gravel
(representing aggregate entering the trench during the excavation process)
should be tested to select a design mix which will reduce the seepage through
the slurry trench cutoff to an acceptable level and approximate the strength
and compressibility of the surrounding ground.
(6) Failure Mechanisms of Cutoffs
9-32
Table 9-6. Summary of Shear Strength Data from Post-Construction Testing
(a)
of Cement-Bentonite Slurry Trench Cutoff at Tilden Tailings Project, Michigan
Sand Time Elapsed Unconfined Triaxial Cornpressiou Tesib)

Content Moisture Construction Compressive
Boring Depth Dry Density By Volume Content to Sampling Strength
~ _g_ lb/cu ft % months
1 6.0 26.5 2.5 271.6 6 0.92
1 10.7 29.4 11.6 151.4 6 0.94
1 15.9 ]2.1 13.2 141.0 6 0.97
l 21.0 37.0 11.4 138.8 6 1.02
1 31.0 )7.5 6.2 129.0 6 1.02 15.5 0.99
1 4J..O 40.0 115.8 6 24.5 0.66

9-33
2 6.9 21.4 0.0 236.0 6 0.88
2 10.8 43.0 6.2 201.0 6 1.00
2 16.0 38.3 11.4 134.5 6 1.14 29.6 0.36
2 21.0 39.0 13.4 117.1 6 1.30
(a) From Northwestern University

(b) Conducted on undisturbed soil specimens; consolidated-drained (S) shear test
EM 1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
100
I I rl
80 f-
Bentonite Type I
e SPV 200
r.
• UG 180 I
60 - I
.,
I
I
-
40
• I
~
/
.... 20
0"
1:
41
...
l<
.,
\1'1
0
-..."'"'
>
41
0 O.l'J o. 20 0. 30
Cement-Water Ratio C/W
0.40 0.50 0.60
0.
E 140
u
0
...c: I I I I
.....
...."' ...
...... \ LEGEND
....., .0 t\.
X
..... 120 • Slurry prepared in the laboratory -
.
~:\
c: o Slurry sampled during constrUction
:::::1
r;,?
......,Cl
........E
.....
:::::1
100
•,\ ' '''
~
~
.,>.
Q
I
80
'~
.
....
a!
' ........
', ~ • .........
60
i' .•""'
'' ~
.......
J
--
00
0
.._
-
40
' ~ :~ 1
• t-0 •
20
' '-,.,
~"-,
.............
•nf~>.ZS"%
.....
0
0 5 10 15 20 25 30
28-oay Triaxial Compressive Strain at Failure

c:f'
Figure 9-12. Strength and deformation characteristics of

cement-bentonite slurries (courtesy of American Society
of Engineers223)
9-34
EM 1110-2-1901
30 Sep 86
(a) Introduction. Several mechanisms can affect the functioning of

slurry walls and cause failure. Failure may occur during excavation of the
trench, upon first filling of the reservoir with the resulting rise in differ-
ential head acting across the slurry wall, or at some future time due to
adverse chemical substances in the soil and ground water. Specific failure
mechanisms include trench collapse, gaps (or windows) in the slurry wall,
inadequate aquiclude key-in, blowout or piping of backfill material into the
surrounding pervious foundation, and chemical destruction of the slurry wall
(Spooner et al. 1982).
(b) Trench Collapse. Trench collapse is caused by instability of the

trench walls during excavation and before backfilling (for soil-bentonite
slurry trench) or setup (for cement-bentonite slurry trench). Causes of trench
collapse include failure to maintain the minimum differential head between the
top of the slurry and the top of the ground water and/or too low unit weight of
slurry (see figure 9-7). Drop in the slurry level in the trench may be caused
by contact with gravel, fissures, etc., during excavation, while rise in the
ground-water level may be caused by surface runoff into cracks adjacent to the
trench, particularly following heavy rainfall. Too low unit weight of the
slurry may be caused by the cessation of agitation by excavation equipment over
the week end (Spooner et al. 1982). Such a set of circumstances contributed to
the collapse of a portion of one wall of a soil-cement bentonite slurry trench
at Duncan Dam in Canada (Duguid et al. 1971).
(c) Gaps (or Windows) in the Slurry Wall. Trench collapse or improper
placement of backfill can create gaps (or windows) which result in zones of
higher permeability as well as variations in wall thickness and strength
(Spooner et al. 1982). The continuity of the trench should be tested before
backfilling by passing the bucket or clamshell of the excavating tool verti-
cally and horizontally along each segment of the trench. As mentioned pre-
viously, for soil-bentonite slurry trenches irregularities in the backfill
slope are indications that pockets of clean material (slurry not mixed in
properly or was washed out) or slurry were trapped in the backfill or that the
backfill does not contain sufficient sand or coarse material.
(d) Inadequate Aquiclude Key-In. As discussed previously, inadequate

aquiclude key-in will permit seepage under the cutoff. Inadequate key-in can
result from variations in trench depth, insufficient aquiclude penetration,
trench collapse, or the presence of boulders (Spooner et al. 1982).
(e) Blowout or Piping of Backfill Material. As mentioned previously,

blowout or piping of backfill material into the surrounding pervious founda-
tion is especially critical for soil-bentonite slurry trenches when the foun-
dation contains openwork gravel. The required width of the slurry trench to
prevent blowout (factor of safety of 3) in openwork gravel may be estimated
from Equation 9-11.
(f) Chemical Destruction of the Slurry Wall. Chemical substances in

the foundation soil and ground water can affect the durability of the slurry
wall once it is constructed. If salt water is present in the construction
area, appapulgite may be used instead of bentonite. Permeation of a soil-
bentonite or cement-bentonite slurry wall by polluted ground water generally
9-35
EM 1110-2-1901
30 Sep 86
leads to increased permeability. The bentonite may become entrained in the

solution and the cement may become slurry solubilized as a solution channel is
created through the slurry wall and into the foundation. Chemicals may also
prevent the slurry from forming an adequate filter cake along the sides of the
soil-bentonite slurry trench (Spooner et al. 1982). Where polluted ground
water is present, long-term permeability tests should be conducted using the
specific soil-bentonite backfill materials or cement-bentonite mix from the
site permeated by the actual pollutant in designing the slurry trench cutoff
(Spooner et al. 1982 and D'Appolonia 1980).
(7) Instrumentation and Monitoring.
(a) Introduction. Whenever a slurry trench is used for control of

underseepage, the initial filling of the reservoir must be controlled and
instrumentation monitored to determine if the slurry trench is performing as
planned. If the slurry trench cutoff is ineffective, remedial seepage control
measures must be installed prior to further raising of the reservoir pool
(EM 1110-2-2300).
(b) Parameters of Interest. There are two parameters of interest with

regard to slurry trench cutoffs for control of underseepage. These are the
drop in piezometric head from upstream to downstream across the trench during
reservoir filling and the differential settlement between the top of the
slurry trench and the overlying compacted embankment material.
(c) Efficiency of Slurry Trench Cutoff. To evaluate the head effi-

ciency (see equation 2) of the slurry trench cutoff, the head loss is deter-
mined between points immediately upstream and downstream of the slurry trench
cutoff wall at its junction with the base of the dam. The head loss is
established from piezometer readings taken during construction, before and
during initial filling of the reservoir, and subsequently as frequently as
necessary to determine changes that are occurring and to assess their impli-
cations with respect to safety of the dam (see Chapter 13). Equal numbers of
piezometers are normally placed on each side of the slurry trench cutoff.
Piezometers should be installed at two or more locations along the length of
the slurry trench depending upon the foundation conditions at the site.
Pneumatic piezometers installed upstream and downstream of the soil-bentonite
slurry trench at West Point Dam, Alabama and Georgia, showed a near-horizontal
piezometric surface existed across the trench prior to filling the reservoir.
Piezometer readings taken after reservoir filling indicated a drop in piezom-
eter head from upstream to downstream across the slurry trench, confirming the
effectiveness of the cutoff (U. S. Army Engineer District, Savannah 1979).
Open-tube piezometers installed upstream and downstream of the soil-bentonite
slurry trench at Addicka Dam, Texas, indicated a drop in piezometric head
across the slurry trench (U. S. Army Engineer District, Galveston 1983).
(d) Differential Settlement. The differential settlement between the

top of the slurry trench and the overlying compacted embankment material is
important because a separation of materials in this region could result in
piping at the interface between the embankment and the foundation. Settlement
plates placed on top of the soil-cement bentonite slurry trench at West Point
Dam, Alabama and Georgia, indicated a uniform total settlement of approximately
9-36
EM 1110-2-1901
30 Sep 86
0.5 ft throughout the trench. Excavation of a portion of the trench prior to

filling the reservoir showed no void between the slurry trench backfill and the
overlying compacted fill (U. S. Army Engineer District, Savannah 1968 and
1979).
e. Concrete Wall.
(1) Introduction. When the depth of the pervious foundation is exces-

sive (< 150 ft) and/or the foundation contains cobbles, boulders, or cavernous
limestone, the concrete cutoff wall may be an effective method for control of
underseepage. Using this method, a cast-in-place continuous concrete wall is
constructed by tremie placement of concrete in a bentonite slurry supported
trench. Two general types of concrete cutoff walls, the panel wall and the
element wall have been used, as shown in figures 9-13 and 9-14, respectively.
Since the wall in its simpler structural form is a rigid diaphragm, earth-
quakes could cause its rupture; therefore, cutoff walls should not be used at
a site where strong earthquake shocks are likely (U. S. Army Engineer
District, Pittsburgh 1965).
(2) History of Use. Conventional (excavated without bentonite slurry)

concrete cutoff walls were widely used prior to 1925. Since they require
about the same excavation and dewatering as compacted backfill trenches and
the wall itself is far more expensive than compacted soil, the popularity of
conventional concrete cutoff walls has declined (Sowers 1962 and Sherard et al.
1963). The method of excavating trenches supported by bentonite for the con-
struction of cast-in-place concrete cutoff walls was used for the first time
in 1951 at the Volturno-Garigliano hydroelectric plant on the Volturno River
at Venafro, near Naples, Italy (Veder 1963, Veder 1975 and Franke 1954). Since
the 1950's, concrete cutoff walls constructed by tremie placement of concrete
in a bentonite slurry supported trench have been used for projects throughout
the world. The deepest concrete cutoff wall to date was constructed at
Manicouagan 3 Dam in Quebec, Canada, in 1972, where two parallel concrete
walls, 2 ft thick and 10 ft apart, extended 430 ft deep (Anonymous 1972). A
comparison of concrete cutoff walls constructed at Corps of Engineers dams is
given in table 9-7.
(3) Sequence of Construction and Location of Wall. Normally the

embankment is constructed first, followed by the concrete cutoff wall located
upstream of the toe of the dam as was done at Kinzua Dam (formerly Allegheny
Dam) and tied into the core of the dam with an impervious blanket (U. S. Army
Engineer District, Pittsburgh 1965). The upstream location minimizes the pos-
sibility of compressive failure of the concrete cutoff wall due to negative
skin friction as the foundation settles under the weight of the embankment (as
would occur is the cutoff wall is located under the center of the dam). Con-
structing the embankment first, followed by the concrete cutoff wall, minimizes
the possibility of rupture of the concrete cutoff wall due to horizontal move-
ment of the foundation as the embankment is constructed. For remedial seepage
control of existing dams (see Chapter 12) where it is not practical to draw
down the reservoir and primary consolidation of the foundation has been
complete, a central location for the concrete cutoff wall may be feasible.
9-37
EM 1110-2-1901
30 Sep 86
Figure 9-13. Construction procedure for concrete cutoff wall at

Kinzua Dam (formerly Allegheny Dam), Pennsylvania (after U. S.
103
Army Engineer District, Pittsburgh )
(4) Design Considerations. The primary design parameters are perme-

ability, strength, and compressibility. (1) The permeability is usually
-10
sufficiently low (=10 cm/sec for water-cement ratio of 0.6) to reduce the
seepage through the concrete cutoff wall to an acceptable value (Xanthakos
1979). The concrete cutoff wall is generally stronger (>3,000 psi compressive
strength) than the surrounding foundation soil and introduces a heterogeneous
zone (in the form of a rigid diaphragm) in the foundation. The compressibility
of the concrete cutoff wall is sufficiently low that the wall is essentially
rigid with respect to the surrounding foundation soil (Xanthakos 1979).
(a) Permeability. For workable concrete mixes used in concrete cutoff

walls (see table 9-8), the permeability increases rapidly for water-cement
ratios higher than 0.5. For a concrete mix with maximum coarse aggregate size
(1)
The workability of the concrete, discussed under Mix Design, is of primary
importance with respect to tremie placement of the concrete.
9-38
EM 1110-2-1901
30 Sep 86
a. Excavation procedure for primary and secondary elements
b. Interlocking of primary
and secondary elements
Figure 9-14. Construction procedure for concrete cutoff wall at Wolf

Creek Dam, Kentucky (courtesy of American Society of Civil
160
Engineers )
9-39
Table 9-7. Comparison of Corps of Engineers Concrete Cutoff Walls
30 Sep 86
EM 1110-2-1901
t.;utorr
Max. Max. (a)
Dif- Devi-
fer- at ion Max.
Date Embankment- ential Max. from Head
Con- Foundation Head Width Depth Ver- Cutoff
Project Location structed Material ft. ft ....!!__ Location tical Width Reference
Kinzua (formerly Allegheny River 1964 Silts, 140 2.5 185 Upstream 1:370 56.0 U. S. Army
Allegheny) Dam Pennsylvania sands and of toe Engineer District,
gravels, of dam Pittsburgh 1965
boulders
Wolf Creek Dam(b) Cumberland River 1978 Embankment 187 2.0 300 Center of 1:600 93.5 U. S. Army
Kentucky (200ft.); dam Engineer District,
cavernous Nashville 1975;
limestone Fetzer 1979
(100 ft.)
Mill Creek Dam Mill Creek 1981 Silt 51 2.0 170 Upstream 1:100 25.5 U. S. Army
Washington overlying toe of Engineer District,
cot)glomerate dam Walla Walla 1980a,b
Clemson Lowerb Hartwe 11 Lake 1982 Embankment 44 2.0 85 Center of 1:100 22.0 u. s. Army
9-40
Diversion Dam Georgia and (50 ft.); dam Engineer District,_

South Carolina silty sand Savannah 1981
and sand
(35 ft)
(a)
(b) Deviation from vertical in any direction at any depth.
Concrete cutoff installed as remedial seepage control for existing dam (see Chapter 12).
Table 9-8. Comparison of Concrete Mix Proportions for Corps of Engineers Concrete Cutoff Walls
28 Day Compressive Strength

Aggregate lb/sg in.
Cement(a) Fly Ash Water Water/Cement lb/cu ;ld Slump
Project lb/cu ~d lb/cu ~d lb/cu xd + Fl~ Ash Coarse Fine in. SEecified Obtained
Kinzua (formerly 564 0 292 0.52 1,652 1,345 7 3,000 3.843
Allegheny Dam)
Wolf Creek Dam 564 123 275 0.40 1,619 1,369 6.5-7.5 3,000 4,685
Mill Creek Dam 299 l34(b) 260 0.60 1,628 1,587 8 1,500 1,970
Clemson Lower 507 110 334 0.54 1,540 1,310 60.9.0 1,300
Diversion Dam
9-41
EM 1110-2-1901
30 Sep 86
(a)
(b) Weights are saturated surface dry.
Fly ash percentage replaced by solid volume was raised from 35 to 45 before wall was completed.
EM 1110-2-1901
30 Sep 86
of 3/4 in. and a water cement ratio of 0.6, the permeability is usually lower
-10
than 10 cm/sec (Xanthakos 1979). The permeability of a concrete cutoff
wall is influenced by cracks in the finished structure and/or by void spaces
left in the concrete as a result of honeycombing or segregation (see Equa-
tion 9-4 and figure 9-5). The joints between panels are not completely
impermeable but the penetration of bentonite slurry into the soil in the
immediate vicinity of the joint usually keeps the flow of water very small
(Hanna 1978). Measured head efficiency for concrete cutoff walls from
piezometric data generally exceeds 90 percent (Telling, Menzies, and Simons
1978b). At Kinzua Dam (formerly Allegheny Dam), the measured head efficiency
was 100 percent, i.e., the head just downstream of the concrete cutoff wall was
of the magnitude established by vertical seepage through the upstream
connecting blanket (Fuquay 1968).
(b) Strength. The compressive strength for concrete cutoff walls is

generally specified to exceed 3,000 lb/sq in. (see table 9-8). Therefore, the
concrete cutoff wall is generally stronger than the surrounding foundation
soil. The most important factor influencing the strength of the concrete is
the water-cement ratio. The concrete's fluidity, i.e., ability to travel
through the tremie and fill the excavation, also depends upon the water-cement
ratio. Too low a water-cement ratio would decrease flowability and increase
compressive strength. Too high a water-cement ratio would promote segrega-
tion. A good balance is achieved with a water-cement ratio near 0.5 which
results in a 28-day compressive strength exceeding 3,000 lb/sq in. (see
table 9-8). Cement continues to hydrate and concrete continues to increase in
compressive strength, at a decreasing rate, long after 28 days (Winter and
Nilson 1979).
(c) Compressibility. The concrete cutoff wall is essentially rigid and

has low compressibility compared to the surrounding foundation soil. The
modulus of elasticity for concrete cutoff walls may be approximated from
(Winter and Nilson 1979)
(9-13)
where
= modulus of elasticity in lb/sq in.
W = unit weight of concrete in lb/cu ft
= compressive strength of concrete in lb/sq in.
(5) Mix Design. In addition to strength, workability is an important

requirement for the concrete mix. The mix must not segregate during place-
ment. Too high a water-cement ratio or too low a cement content (with a good
water-cement ratio) will tend to segregate. Natural well rounded aggregate
increases flowability and allows the use of less cement than an angular
9-42
EM 1110-2-1901
30 Sep 86
manufactured aggregate. Since the concrete is poured into the trench through
tremie pipes and displaces the bentonite slurry from the bottom of the exca-
vation upward, the concrete must have a consistency such that it will flow
under gravity and resist mixing with the bentonite slurry. Admixtures may be
used as required to develop the desired concrete mix characteristics. Fly ash
is often used to improve workability and to reduce heat generation. The unique
problems inherent at each project require studies to develop an adequate con-
crete mix (Holland and Turner 1980). Some typical concrete mixes used in Corps
of Engineers concrete cutoff walls are given in table 9-8. The placement
techniques used for the concrete are of equal importance in assuring a satis-
factory concrete cutoff wall.
(6) Excavation and Placement of Concrete. Temporary guide walls are

constructed at the ground surface to guide the alignment of the trench and
support the top of the excavation. Typically, a cross section, 1 ft wide and
3 ft deep, is sufficient for most concrete cutoff walls. In order to ensure
continuity between panels and provide a watertight joint to prevent leakage,
an appropriate tolerance is placed on the maximum deviation from the vertical
(see table 9-7). The same general requirements apply to the slurry used to
keep the trench open for concrete cutoffs. As stated previously, two general
types of concrete cutoff walls, the panel wall, and the element wall have been
used. The panel wall is best suited for poorly consolidated materials and soft
rock can be installed to about a 200-ft depth. The element wall has the
advantage of greater depth (430 ft deep at Manicouagan 3 Dam in Quebec,
Canada), better control of verticality, the ability to penetrate hard rock
using chisels and/or nested percussion drills, and the protection of the
embankment with casing when used for remedial seepage control. However, the
element wall is more costly and has a slower placement rate than the panel
wall. Both types of concrete cutoff walls open short horizontal sections of
the embankment and/or foundation at a time, which limits the area for potential
failure to a segment that can be controlled or repaired without risking
catastrophic failure of the project. The concrete cutoff wall penetrates the
zone(s) of seepage with a rigid, impermeability barrier capable of withstanding
high head differentials across cavities with no lateral support. The concrete
must be placed at considerable depth through bentonite slurry in a continuous
operation with as little contamination, honeycomb, or segregation as possible.
The bottom of the excavation must be cleaned so that a good seal can be
obtained at grade. Fresh bentonite slurry is circulated through the excavation
to assist in the cleaning and lower the density of the slurry to allow the
concrete to displace the slurry easier once placement begins. The tremie
procedure used to place the concrete is straightforward in theory and yet often
in practice causes more problems with the final quality of the concrete cutoff
wall than any other factor. The tremie system consists of a hopper, tremie
pipe, and a crane or other lifting equipment to support the apparatus. The
hopper should be funnel shaped and have a minimum capacity of 0.5 cu yd. The
size of the tremie pipe depends upon the size of aggregate used in the concrete
mix. For 3/4-in. maximum diameter coarse aggregate, a 10-in.-diam tremie pipe
9-43
EM 1110-2-1901
30 Sep 86
should be used. (1) The dry tremie is placed in the hole with a metal plate and
rubber gasket wired to the end of the tremie. The tremie pipe is lifted,
breaking the wires and allowing the concrete flow to begin. Concrete is added
to the hopper at a uniform rate to minimize free fall to the surface in the
pipe and obtain a continuous flow. The tremie apparatus is lifted during
placement at a rate that will maintain the bottom of the pipe submerged in
fresh concrete at all times and produce the flattest surface slope of concrete
that can practically be achieved. The flow rate (foot of height per hour) and
surface slope of the concrete shall be continuously measured during placement
with the use of a sounding line. A sufficient number of tremies should be
provided so that the concrete does not have to flow horizontally from a tremie
more than 10 ft. As soon as practical, core borings should be taken in
selected panels through the center of the cutoff wall to observe the quality of
the final project. Unacceptable zones of concrete such as honeycombed zones,
segregated zones, or uncemented zones found within the cored panels or elements
should be repaired or removed and replaced. One means of minimizing such
problems at the start of a job is to require a test section in a noncritical
area to allow changes in the construction procedure to be made early in the
project (Hallford 1983; Holland and Turner 1980; and Gerwick, Holland, and
Komendant 1981).
(7) Treatment at Top of Concrete Cutoff Wall. As mentioned previously,

normally the concrete cutoff wall is located under or near the upstream toe of
the dam and tied into the core of the dam with an impervious blanket. If a
central location for the concrete cutoff wall is dictated by other factors,
the connection detail between the top of the concrete cutoff wall and the core
of the dam is very important. Generally, the concrete cutoff wall extends
upward into the core such that, the hydraulic gradient at the surface of the
contact does not exceed 4 (Wilson and Marsal 1979). Various precautions (see
figure 9-15) have been taken to prevent the top of the concrete cutoff wall
from punching into the core of the dam and causing the core to crack as the
foundation settles on either side of the rigid cutoff wall under the weight of
the embankment. The bentonite used at the connection between the concrete
cutoff wall and the core of the dam (see figure 9-15) is intended to create a
soft zone to accommodate differential vertical settlements of the core around
the concrete cutoff wall. Also, saturation of the bentonite is intended to
produce swelling which will provide for a bond between the core and the con-
crete cutoff wall to prevent seepage (Radukic 1979).
(8) Failure Mechanisms of Concrete Cutoff Walls. Several mechanisms

can affect the functioning of concrete cutoff walls and cause failure. As
mentioned previously, the wall in its simpler structural form is a rigid
diaphragm and earthquakes could cause its rupture. For this reason concrete
cutoff walls should not be used at a site where strong earthquake shocks are
(1)
At Wolf Creek Dam concrete problems (areas of segregated sand or coarse
aggregate, voids, zones of trapped laitance, and honeycombed concrete)
occurred for tremie-placed 26-in. -diam cased primary elements. This must
be considered in future projects which involve tremie-placed elements of
small cross-sectional areas (Holland and Turner 1980).
9-44
EM 1110-2-1901
30 Sep 86
a. Forked connection b. Plastic impervious cap
c. Piston connection d. Double wall connection
Figure 9-15. Connections between concrete cutoff wall

182
and core of dam (courtesy of ICOS )
9-45
EM 1110-2-1901
30 Sep 86
likely. Concrete cutoff walls located under or near the toe of the dam are
subject to possible rupture from horizontal movements of the foundation soil
during embankment construction. This effect can be minimized by constructing
the dam embankment prior to the concrete cutoff wall. As mentioned previously,
concrete cutoff walls located under the center of the dam are subject to pos-
sible compressive failure due to negative skin friction as the foundation
settles under the weight of the embankment. The probability of this occurring
would depend upon the magnitude of the negative skin friction developed at the
interface between the concrete cutoff wall and the foundation soil and the
stress-strain characteristics of the concrete cutoff wall. Also, as previously
mentioned, a centrally located concrete cutoff wall may punch into and crack
the overlying core material unless an adequate connection is provided between
the concrete cutoff wall and the core of the dam.
(9) Instrumentation and Monitoring. Whenever a concrete cutoff wall is

used for control of underseepage, the initial filling of the reservoir must be
controlled and instrumentation monitored to determine if the concrete cutoff
wall is performing as planned. If the concrete cutoff wall is ineffective,
remedial seepage control measures must be installed prior to further raising
the reservoir pool. When the embankment is constructed first, followed by the
concrete cutoff wall located upstream of the toe of the dam, as was done at
Kinzua (formerly Allegheny Dam), the parameters of interest are the drop in
piezometric head from upstream to downstream across the concrete cutoff wall,
differential vertical settlement between the upstream impervious blanket and
the top of the concrete cutoff wall, and vertical and horizontal movement of
the concrete cutoff wall due to reservoir filling. If a central location for
the concrete cutoff wall is dictated by others factors, the parameters of
interest are the drop in piezometric head from upstream to downstream across
the cutoff wall, differential vertical settlement between the core of the dam
and the top of the concrete cutoff wall, and vertical and horizontal movement
of the concrete cutoff wall due to construction of the embankment and reser-
voir filling. Instrumentation data should be obtained during construction,
before and during initial filling of the reservoir, and subsequently as fre-
quently as necessary to determine changes that are occurring and to assess
their implications with respect to the safety of the dam (see Chapter 13).
The head efficiency for concrete cutoff walls is evaluated in the same manner
as described previously for slurry trench cutoffs. As previously mentioned,
measured head efficiency for concrete cutoff walls generally exceeds
90 percent.
f. Steel Sheetpiling.
(1) Introduction. Steel sheetpiling is rolled steel members with

interlocking joints along their edges. Sheetpiling is produced in straight
web, arch web, and Z sections in a graduated series of weights joined by
interlocks to form a continuous cutoff wall as shown in figure 9-16. Steel
sheetpiling is not recommended for use as a cutoff to prevent underseepage
beneath dams due to the low head efficiency. Steel sheetpiling is frequently
used in conjunction with concrete flood control and navigation structures to
confine the foundation soil to prevent it from piping out from under the
structure (EM 1110-2-2300 and Greer, Moorhouse, and Millet 1969).
9-46
EM 1110-2-1901
30 Sep 86
STRAIGHT ARCH Z
a. Sections
b. Interlocking of sections
Figure 9-16. Steel sheetpiling installation (from U. S. Army Engineer

57
Waterways Experiment Station )
9-47
EM 1110-2-1901
30 Sep 86
(2) History of use. Steel sheetpiling was first used by the Corps of
Engineers to prevent underseepage at Fort Peck Dam, Montana (U. S. Army Engi-
neer District, Omaha 1982). The steel sheetpiling, driven to Bearpaw shale
bedrock with the aid of hydraulic spade jetting, reached a maximum depth of
163 ft in the valley section (see table 9-9). An original plan to force grout
into the interlocks of the steel sheetpiling was abandoned during construction
as impractical. Steel sheetpiling was used as an extra factor to prevent pip-
ing of foundation soils at Garrison Dam, North Dakota (U. S. Army Engineer
District, Omaha 1964). At Garrison Dam, underseepage control was provided for
by an upstream blanket and relief wells and the contribution of the steel
sheetpiling to reduction of underseepage was neglected in the design of the
relief wells. Steel sheetpiling and an upstream blanket were installed for
control underseepage at Oahe Dam, South Dakota. Relief wells were installed
for remedial seepage control to provide relief of excess hydrostatic pressures
developed by underseepage (U. S. Army Engineer District, Omaha 1961).
(3) Efficiency of Steel Sheetpiling Cutoffs. The efficiency of steel

sheetpiling cutoffs is dependent upon proper penetration into an impervious
stratum and the condition of the sheeting elements after driving. When the
foundation material is dense or contains boulders which may result in ripping
of the sheeting or damage to the interlocks (see figure 9-17), the efficiency
will be reduced (Guertin and McTigue 1982). Theoretical studies indicate that
very small openings in the sheeting (< 1 percent of the total area) will cause a
substantial reduction in the cutoff efficiency (from 100 to 10 percent effi-
ciency) as shown in figure 9-18 (Ambraseys 1963). The measured head efficiency
for steel sheetpiling cutoffs installed at Corps of Engineers dams is given in
table 9-9. The effectiveness of the steel sheetpiling is initially low, only
12 to 18 percent of the total head was lost across the cutoff as shown in
table 9-9. With time, the head loss across the steel sheetpiling increased to
as much as 50 percent of the total head. This increase in effectiveness is
attributed to migration of fines and corrosion in the interlocks and reservoir
siltation near the dam.
9-5. Upstream Impervious Blanket. (1)
a. Introduction. When a complete cutoff is not required or is too

costly, an upstream impervious blanket tied into the impervious core of the
dam may be used to minimize underseepage. Upstream impervious blankets should
not be used when the reservoir head exceeds 200 ft because the hydraulic
gradient acting across the blanket may result in piping and serious leakage.
Downstream underseepage control measures (relief wells or toe trench drains)
are generally required for use with upstream blankets to control underseepage
and/or prevent excessive uplift pressures and piping through the foundation.
Upstream impervious blankets are used in some cases to reinforce thin spots in
natural blankets. Effectiveness of upstream impervious blankets depends upon
their length, thickness, and vertical permeability, and on the stratification
and permeability of soils on which they are placed (EM 1110-2-2300, Barron
1977 and Thomas 1976).
(1) The blanket may be impervious or semipervious (leaks in the vertical

direction).
9-48
Table 9-9. Comparison of Corps of Engineers Steel Sheetpiling Cutoff Walla
Max Head Efficiency of

Date Foundation Depth Cutoff, %
Project Location Constructed Materials ...!!..__ Location Pur~se Initial With TiM Reference
Ft. Peck Dam Missouri River, 1940 Gravels, aa~ds, 163 Center of daa Prevent 12 lo(a) Lane and Wohlt 1961
Montana and·silts under seepage
Garrison Dam I.Ussouri River, 1954 Sand with 110 Upstream Prevent piping 18 50(b) u. s. Army Engineer
North Dakota gravel lenses portion of District, Ollaha
dam 1964
Oahe Dam Missouri River, 1958 Sands, gravels, 75 Upstream Prevent piping 16 22(c) u. s. Army Engineer
South Dakota and silts portion of District, Olasha
dam 1961
(a) After 17 years (relief wells were installed in 1942 aa remedial underseepage control)
(b) After 10 years (upstream blanket and relief wells were constructed prior to impoundment)
(c) After 4 years (upstream blanket constructed prior to impoundment and relief wells installed in 1962 as remedial underseepage control)
9-49
EM 1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
Figure 9-17. Sources of leakage associated with steel sheetpile

cutoffs (from U. S. Department of Transportation 41)
9-50
EM 1110-2-1901
30 Sep 86
Figure 9-18. Cutoff efficiency versus open space ratio for

129
imperfect cutoffs (courtesy of Butterworths, Inc. )
b. Design Considerations. In alluvial valleys, frequently soils consist

of fine-grained top stratum of clay, silt, and silty or clayey sand underlain
by a pervious substratum of sand and gravel. As stated previously, the top
stratum or blanket may be impervious or semipervious (leaks in the vertical
direction). The substratum aquifer or pervious foundation is generally
anisotropic with respect to permeability so the flow is horizontal. For this
condition, shown in figure 9-19, the basic assumptions for the design of up-
stream impervious blankets are:
(1) Flow through the blanket is vertical.
(2) Flow through the pervious foundation is horizontal.
(3) All flows are laminar and steady state.
9-51
EM 1110-2-1901
30 Sep 86
a. Continuous blanket and aquifer
b. Discontinuous upstream blanket, continuous aquifer
L1 = Effective length of upstream natural blanket

L 2 = Length of embankment base
L 3 = Effective length of downstream natural blanket
L o = Length of discontinuous upstream blanket
h = Net head to dissipate
Z = Thickness of natural blanket
d = Thickness of aquifer
kb = Permeability coefficient of blanket
k f = Permeability coefficient of aquifer
= Submerged unit weight of blanket
ho = Pressure head under blanket at downstream toe of dam
h C = Critical head under blanket at downstream toe of dam
F h = Factor of safety relative to heaving at downstream toe
= Unit weight of water (63.4 pcf)
qf = Rate of discharge through aquifer with unit length normal to the
section
Figure 9-19. Upstream impervious blanket (from U. S. Department of

72
Agriculture )
9-52
EM 1110-2-1901
30 Sep 86
(4) The dam (or core of a zoned embankment) is impervious.
(5) Both the blanket and substratum have a constant thickness and are
horizontal.
When the top stratum or pervious foundation consists of several layers of

different soils, they must be transformed into a single stratum with an
effective thickness and permeability (see procedure given in U. S. Army
Engineer Waterways Experiment Station 1956a). For the upstream impervious
blanket shown in figure 9-19, the effective length of the upstream blanket is
(9-14)
where
L1 = effective length of upstream blanket
kf = horizontal permeability of pervious foundation
k b R = vertical permeability of upstream blanket
Z b R = thickness of upstream blanket
d = thickness of pervious foundation
The effective length of the downstream blanket is
(9-15)
where
L 3 = effective length of downstream blanket
kbL = vertical permeability of downstream blanket
Z b L = thickness of downstream blanket
Upstream blankets should be designed so that under maximum reservoir

conditions the pressure head under the blanket at the downstream toe of the
dam and the rate of discharge through the pervious foundation are acceptable.
The pressure head under the blanket at the downstream toe of the dam (see
figure 9-19) is
9-53
EM 1110-2-1901
30 Sep 86
(9-16)
where
h o = pressure head under the blanket at the downstream toe of the dam
h = net head to dissipate
L2 = length of impervious core or dam base
The critical pressure head under the blanket at the downstream toe of the dam
is
(9-17)
where
hc = critical pressure head under the blanket at the downstream toe of

the dam
= submerged unit weight of downstream blanket soil
The factor of safety against uplift or heaving at the downstream toe of the dam
is
(9-18)
where Fh is the factor of safety against uplift or heaving at the downstream

toe of the dam. Generally dams are designed without relying upon natural
downstream blankets because it is difficult to assure the continuity and the
existence of the blanket throughout the life of the structure. Also, down-
stream seepage control measures (relief wells or trench drains) are generally
used with upstream blankets to reduce uplift or heaving at the downstream toe
of the dam. However, for the exceptional case where the dam is designed with a
natural downstream blanket and with no downstream seepage control measures
(relief wells or trench drains), upstream blankets should be designed so that
the factor of safety against uplift or heaving at the downstream toe of the dam
is at least 3. The rate of discharge through the pervious foundation per unit
length of dam (see figure 9-19) is
9-54
EM 1110-2-1901
30 Sep 86
(9-19)
where qf is the rate of discharge through the pervious foundation per unit
length of dam. The acceptable rate of discharge or underseepage depends upon
the value of the water or hydropower lost, availability of downstream right-
of-way, and facility for disposal of underseepage. The following procedure is
used to determine the length of an upstream blanket when there is a downstream
blanket present (see figure 9-19b):
(a) Determine L1 from equation 9-14 using a conservative value of

kf/kbR , i.e., the highest probable ratio.
(b) Determine L3 from equation 9-15 using a conservative value of

kf/kbL , i.e., the highest probable ratio.
(c) Determine ho , hc , and Fh from equations 9-16, 9-17, and 9-18,

respectively. If Fh < 3.0 , the blanket thickness of the upstream blanket may
be increased, the permeability of the upstream blanket material may be
decreased by compaction, or downstream seepage control measures may be used.
(d) Determine the rate of discharge through the pervious foundation per
unit length of dam from equation 9-19. If the rate of discharge is excessive,
a reduction can be obtained by increasing the thickness of the upstream blanket
or reducing the permeability of the upstream blanket material by compaction.
When these methods are used, steps 1 to 4 are repeated before going to step 5.
(e) If the rate of discharge is acceptable, calculate the factor
(9-20)
where c has the units of 1/ft .
(f) Enter figure 9-20 with c and L1 and obtain Lo , which is the
distance from the upstream toe of a homogeneous impervious dam or the imper-
vious core section of a zoned embankment to where a discontinuity in the up-
stream blanket will have no effect on the uplift at the downstream toe of the
dam or rate of discharge through the pervious foundation. This is the point
beyond which a natural blanket may be removed in a borrowing operation.
Also, Lo would represent the distance upstream from the toe of the dam to
which a streambed should be blanketed to ensure the continuity of a natural
upstream blanket. If there is no downstream blanket the pressure head under
the blanket at the downstream toe of the dam will be zero (see equation 9-16)
9-55
EM 1110-2-1901
30 Sep 86
Figure 9-20. Effective lengths of upstream and downstream imper-

72
vious blankets (from U. S. Department of Agriculture )
and the following procedure is used to determine the length of the upstream
blanket:
• Assume several values of Lo (length of the upstream blanket from

the upstream toe of a homogeneous impervious dam or the impervious core
section of a zoned embankment).
9-56
EM 1110-2-1901
30 Sep 86
• Calculate c from equation 9-20 using the design thickness and

permeability rates for the constructed blanket and pervious foundation. Note
that c has units of 1/ft.
• Enter figure 9-20 with the assumed values of Lo and the

calculated values of c to obtain the corresponding value of Ll for each
assumed value of Lo .
• Calculate qf from equation 9-19 (L3 = 0 for no downstream

blanket) using the values of L1 obtained from figure 9-20.
• Plot qf versus Lo . The curve will indicate a rapid decrease in

qf with increasing values up to a point where the curve flattens out indicat-
ing an optimum length. The upstream blanket can be terminated at any point
where the desired reduction in rate of discharge through the pervious founda-
tion per unit length of dam is achieved (Talbot and Nelson 1979).
c. Materials and Construction. At sites where a natural blanket of

impervious soil already exists, the blanket should be closely examined for
gaps such as outcrops of pervious strata, streambeds, root holes, boreholes,
and similar seepage paths into the pervious foundation which, if present,
should be filled or covered with impervious material to provide a continuous
blanket to a distance Lo from the upstream toe of the dam. Also, as
previously stated, upstream borrow areas should be located greater than the
distance L o from the upstream toe of the dam so as not to reduce the effec-
tiveness of the natural blanket. Figure 9-21 shows the influence of gaps in
the upstream blanket on relative seepage and uplift at the toe of the dam.
That portion of the upstream blanket placed beneath the embankment to tie into
the impervious core should be composed of the same material and compacted in
the same manner as the core. Upstream of the embankment, the blanket is con-
structed by placing impervious soil in lifts and compacted only by movement of
hauling and spreading equipment, or to whatever additional extent is necessary
for equipment operation. Exposed clay blankets can shrink and crack after
placement. If such cracks penetrate the blanket, they will reduce the effec-
tiveness of the blanket. Thus it may become necessary to sprinkle the surface
of the blanket to help retain moisture until a permanent pool is impounded. In
higher reaches of abutments which are infrequently flooded by the reservoir, a
thicker blanket may be required so that cracks will not fully penetrate the
blanket. In colder climates, the blanket thickness should be increased to
account for the loosening of the upper part of the blanket by frost action
which substantially increases the permeability.
d. Reservoir Siltation. For some reservoirs, appreciable siltation

occurs which may both increase the thickness of and lengthen the upstream
blanket. Although the siltation may reduce the rate of discharge through the
pervious foundation with time, it is not a factor to be counted upon in design
because the upstream blanket must function adequately following initial fill-
ing of the reservoir prior to the occurrence of siltation.
9-57
EM 1110-2-1901
30 Sep 86
a. Cross section of dam
b. Flow net with incomplete blanket (X/L = 0.1)
c. Relative seepage d. Uplift at toe
Figure 9-21. Effect of gap in upstream blanket on relative seepage

155
and uplift at toe (courtesy of John Wiley and Sons )
9-6. Downstream Seepage Berms.

costly, and it is not feasible to construct an upstream impervious blanket, a
downstream seepage berm may be used to reduce uplift pressures in the pervious
foundation underlying an impervious top stratum at the downstream toe of the
dam. Other downstream underseepage control measures (relief wells or toe
trench drains) are generally required for use with downstream seepage berms.
Downstream seepage berms can be used to control underseepage efficiently where
the downstream top stratum is relatively thin and uniform or where no top
stratum is present, but they are not efficient where the top stratum is
relative thick and high uplift pressures develop. Downstream seepage berms
may vary in type from impervious to completely free draining. The selection
of the type of downstream seepage berm to use is based upon the availability
of borrow materials and relative cost of each type.
b. Design Considerations. When the top stratum or pervious foundation

consists of several layers of different soils, they must be transformed into a
single stratum with an effective thickness and permeability (see procedure
given in U. S. Army Engineer Waterways Experiment Station 1956a). Where a
9-58
EM 1110-2-1901
30 Sep 86
downstream natural blanket is present, the downstream seepage berm should have
a thickness so that the factor of safety against uplift or heaving at the down-
stream toe of the dam is at least 3 and width so that the factor of safety
against uplift at the downstream toe of the seepage berm is at least 1.5.
Formulas for the design of downstream seepage berms where a downstream natural
blanket is present are given in figure 9-22. If there is no downstream natural
blanket present, the need for a downstream seepage berm will be based upon
Bligh's creep ratio.
(9-21)
where
c B = Bligh's creep ratio
Xl = effective length of upstream blanket
L 2 = length of dam base
X = width of downstream seepage berm
h = net head on dam
Minimum acceptable values of Bligh's creep ratio are given in table 9-10. If
the creep ratio is greater than the minimum value, a downstream seepage berm is
not required. (1) If the creep ratio is less than the minimum value, the width
of the downstream seepage berm should be made such that the creep ratio is
above the minimum value shown in table 9-10. The thickness of the downstream
seepage berm at the toe of the dam will be determined so that the factor of
safety against uplift or heaving at the downstream toe of the dam is at
least 3. The pressure head beneath the downstream seepage berm at the landside
toe of the levee is
(9-22)
where
h o = pressure head under the seepage berm at the downstream toe of the
dam
(1)
A downstream seepage berm may be required to correct other problems such
as excessive seepage gradients under the dam (could be detected by check-
ing the rate of underseepage).
9-59
EM 1110-2-1901
30 Sep 86
d = thickness of pervious foundation
X1 = effective length of upstream natural blanket (taken equal to 0.43d

where no upstream natural blanket exists)
The rate of discharge through the pervious foundation per unit length of dam
is
(9-23)
where
q f = rate of discharge through the pervious foundation per unit length

of dam
kf = horizontal permeability of pervious foundation
As stated previously, the acceptable rate of discharge or underseepage depends

upon the value of the water or hydropower lost, availability of downstream
right-of-way, and facility for disposal of underseepage. Downstream seepage
berms should have a minimum thickness of 10 ft at the dam toe and a minimum
thickness of 5 ft at the berm toe. The computed thickness of the berm should
be increased 25 percent to allow for shrinkage, foundation settlements, and
variations in the design factors. Downstream seepage berms should have a
slope of 1V on 50H or steeper to ensure drainage (U. S. Army Engineer Waterways
Experiment Station 1956a).
c. Materials and Construction. As previously stated, the selection of

the type of material used to construct the downstream seepage berm is based
upon the availability of borrow materials and relative cost of each type. A
berm constructed of impervious soil should be composed of the same material as
the impervious core. That portion of the downstream impervious seepage berm
placed beneath the embankment to tie into the impervious core should be com-
pacted in the same manner as the core. Downstream of the embankment, the
impervious seepage berm is constructed by placing impervious soil in lifts and
compacting only by movement of hauling and spreading equipment, or to whatever
additional extent is necessary for equipment operation. Semipervious material
used to construct downstream seepage berms should have an in-place vertical
permeability equal to or greater than that of the upstream natural blanket and
are compacted in the same manner as described previously for impervious mate-
rial. Material used in a sand berm should be as pervious as possible, with
-4
a minimum in-place vertical permeability of 100 x 10 cm per sec. Downstream
seepage berms constructed of sand should be compacted to an average in-place
relative density of at least 85 percent with no portion of the berm having a
relative density less than 80 percent. As proper functioning of a downstream
seepage berm constructed of sand depends upon its continued perviousness, it
should not be constructed until after the downstream slope of the earth dam has
9-60
X
PeRVIOUS SUBSTRATUM
NOTATIONS
BERM=~
X = REQUIRED BERM WIDTH
c = h0 = HEAD AT LANDS IDE TOE OF DAM WITHO\.)T
s + )(3 t = REQUIRED THICKNESS OF BERM AT TOE OF DAM
h~ = HEAD AT LANDSIDE TOE OF DAM WITH BERM
k 1 = VERTICAL PERMEABILITY OF BERM
SUBMERGED UNIT WEIGHT OF TOP STRATUM i 0 = ALLOWABLE UPWARD GRADIENT AT LANDSIDE TOE OF DAM
F = FACTOR OF SAFETY AGAINST UPLIFT AT TOE
SUBMERGED UNIT WEIGHT OF BERM
9-61
i 1 = ALLOWABLE UPWARD GRADIENT AT TOE OF BERM OF DAM

ha "' ALLOWABLE HEAD AT TOE OF BERM 0 8 = FLOW INTO BERM PER FT OF DAM
UNIT WEIGHT OF WATER =
1 zt
1
FORMULAS
IMPERVIOUS BERM (k, = 01 SEMIPERVIOUS BERM lk 1 - k0 ci SAND BERM PERVIOUS BERM WITH COLLECTOR
X1 = x{~- ~-s Xsp -

-A + V A 2 - 24(2 + t) (1 + SC- ':!_)
ha_
2C 12 + t)
, Hx
3 h = h = -3 -
h , =H (:-x- +X
-) WHERE IN·
0 0 s + )(3
o S + x3 +X
A= 6 + 3 SC (r + 1)
1i
1+--
F 1w
h;,- 1 Z1
0 =~ (1 -~,;
EM 1110-2-1901
0
USE F;;. 1.5 t=---
1 + i0 B S + x3
30 Sep 86
Figure 9-22.
Design of downstream seepage berms where a downstream natural blanket is
120
present (from U.S. Army Engineer Waterways Experiment Station )
EM 1110-2-1901
30 Sep 86
Table 9-10. Minimum Bligh's Creep Ratios for Dams

Founded on Pervious Foundations (a)
Minimum Bligh's
Material Creep Ratio
Very fine sand or silt 18

Fine to medium sand 15
Coarse sand 12
Fine gravel or sand and gravel 9
Coarse gravel including cobbles
(a) From U. S. Army Engineer Waterways Experiment
Station120
become covered with sod and stabilized so that soil particles carried by sur-
face runoff and erosion will not clog the seepage berm. If it is necessary to
construct the downstream seepage berm at the time the earth dam is built or
before it has become covered with sod, an interceptor dike should be built at
the intersection of the downstream toe of the dam and the seepage berm to pre-
vent surface wash from clogging the seepage berm. A free-draining downstream
seepage berm is one composed or random fill overlying horizontal sand and
gravel drainage layers with a terminal perforated collector pipe system (U. S.
Army Engineer Waterways Experiment Station 1956a).
9-7. Relief Wells.

costly, relief wells installed along the downstream toe of the dam may be used
to prevent excessive uplift pressures and piping through the foundation.
Relief wells increase the quantity of underseepage from 20 to 40 percent
depending upon the foundation conditions. Relief wells may be used in combi-
nation with other underseepage control measures (upstream impervious blanket
or downstream seepage berm) to prevent excessive uplift pressures and piping
through the foundation. Relief wells are applicable where the pervious foun-
dation has a natural impervious cover. The well screen section (see fig-
ure 9-23), surrounded by a filter if necessary, should penetrate into the
principal pervious stratum to obtain pressure relief, especially where the
foundation is stratified. The wells, including screen and riser pipe, should
have a diameter which will permit the maximum design flow without excessive
head losses but in no instance should the inside diameter be less than
6 in. Filter fabrics should not be used in conjunction with relief wells (see
Appendix D). Even in nearly homogeneous stratum, a penetration of less than
50 percent results in significant rise in pressure midway between adjacent
wells, or requires close spacing. Relief wells should be located so that
9-62
EM 1110-2-1901
30 Sep 86
V-lWTCH WEIR
METAL WELL GUARD -
VARIABLE- TO TOP OF
MINIMUM WATER TABLE
SAND BACKFILL
RISER PIPE TOF' OF

GRAVEL FILTER
0 0"
• 0
0 o°• TOP OF
•:o~ WELL SCREEN
•• 0
BLANK PIPE
THROUGH VERY
FINE SAND STRATA
o•.
0 ..
-~+---- GRAVEL FILTER

0 0 •
e• 0 o
• 0 .o
...
• 0
•• 0
...
0 ••
0 PERFORATED OR
••• SLOTTED SCREEN
~l L.:.'-'·'-.::::::__;;._.o BOTTOM PLUG
MIN
Figure 9-23. Typical relief well (after EM 1110-2-1913)
9-63
EM 1110-2-1901
30 Sep 86
their tops are accessible for cleaning, sounding for sand, and pumping to
determine discharge capacity. Relief wells should discharge into open ditches
or into collector systems outside of the dam base which are independent of toe
drains or surface drainage systems. Experience with relief wells indicates
that with the passage of time the discharge of the wells will gradually
decrease due to clogging of the well screen and/or reservoir siltation. A
comprehensive study of the efficiency of relief wells along the Mississippi
River levee showed that the specific yield of 24 test wells decreased 33 per-
cent over a 15-year period. Incrustation on well screens and in gravel filters
was believed to be the major cause (Montgomery 1972). Therefore, the amount of
well screen area should be designed oversized and a piezometer system installed
between the wells to measure the seepage pressure, and if necessary additional
relief wells should be installed (EM 1110-2-2300, U. S. Army Engineer Waterways
Experiment Station 1956a, Singh and Sharma 1976).
b. History of Use. The first use of relief wells to prevent excessive

uplift pressures at a dam was by the U. S. Army Engineer District, Omaha, when
21 wells were installed from July 1942 to September 1943 as remedial seepage
control at Fort Peck Dam, Montana. The foundation consisted of an impervious
stratum of clay overlying pervious sand and gravel. Although a steel sheetpile
cutoff was driven to shale, sufficient leakage occurred to develop high hydro-
static pressure at the downstream toe that produced a head of 45 ft above the
natural ground surface. This uplift pressure was first observed in piezometers
installed in the pervious foundation. The first surface evidence of the high
hydrostatic pressure came in the form of discharge from an old well casing that
had been left in place. Since it was important that the installation be made
as quickly as possible, 4- and 6-in. well casings, available at the site, were
slotted with a cutting torch and installed in the pervious stratum with solid
(riser) pipe extending to the surface. Wells were first spaced on 250-ft
centers and later intermediate wells were installed making the spacing 125 ft.
The hydrostatic pressure at the downstream toe was reduced from 45 to 5 ft and
the total flow from all wells averaged 10 cu ft per sec (U. S. Army Engineer
District, Omaha 1982). The first use of relief wells in the original design of
a dam was by the U. S. Army Engineer District, Vicksburg, when wells were
installed during construction of Arkabutla Dam, Mississippi, completed in June
1943. The foundation consisted of approximately 30 ft of impervious loess
underlain by a pervious stratum of sand and gravel. The relief wells were
installed to provide an added measure of safety with respect to uplift and
piping along the downstream toe of the embankment. The relief wells consisted
of 2-in. brass wellpoint screens 15 ft long attached to 2-in. galvanized
wrought iron riser pipes spaced at 25-ft intervals located along a line 100 ft
upstream of the downstream toe of the dam. The top of the well screens was
installed about 10 ft below the bottom of the impervious top stratum. The well
efficiency decreased over a 12-year period to about 25 percent primarily as a
result of clogging of the wells by influx of foundation materials into the
screens and/or the development of corrosion or incrustation. However, the
piezometric head along the downstream toe of the dam, including observations
made at a time when the spillway was in operation, has not been more than 1 ft
above the ground surface except at sta 190+00 where a maximum excess hydro-
static head of 9 ft was observed (U. S. Army Engineer Waterways Experiment
Station 1958). Since these early installations, relief wells have been used at
9-64
EM 1110-2-1901
30 Sep 86
many dams to prevent excessive uplift pressures and piping through the
foundation.
c. Design Considerations.
(1) General. The factors to be considered in determining the need for

and designing a relief well system include characteristics of the landside top
stratum; permeability, stratification, and depth of the pervious foundation in
which. seepage is to be controlled; net head acting on the dam; dimensions of
the relief well system being considered; allowable factor of safety with
respect to uplift at the downstream toe of the dam; and allowable rate of dis-
charge through the pervious foundation. Some factors, like the net head act-
ing on the dam, can be determined with good accuracy. Other factors like
permeability and stratification are more difficult to assess. The design of
the relief wells should be based on the best estimate of permeability values
and then subjected to a sensitivity analysis using several values of perme-
ability to ensure that the adopted design is adequate to intercept seepage and
lower uplift pressures to the required extent allowing for the likelihood that
the values of permeability used in design lake precision (Kaufman 1976). The
area between the dam abutments is divided into reaches where geologic and soil
conditions are assumed uniform within the reach (see figure 9-24). Generally,
the design procedure for relief wells consists of determining the head which
would exist along the downstream toe of the dam without relief wells, compar-
ing this head to that desired for a given factor of safety, and designing a
relief well system to reduce the head to the desired value. There is no
unique solution because there is an infinite number of well systems (radius,
penetration, spacing, etc.) which reduce the head to the given value. The
objective is to select one which is economical, has reasonable dimensions, and
can produce the desired results. Usually the designer selects the radius and
penetration and then determine the required spacing of the well system. This
becomes an iterative procedure wherein the designer assumes a value of well
spacing, computes the head between wells and repeats this for several trial
spacings until a spacing is found that produces the desired head along the
downstream toe of the dam. The cost of the well system is determined and then
a design can be prepared for a different penetration to determine if some
economy can be achieved by changing the penetration of the system. Fully pene-
trating relief wells are often used in aquifers up to about 75 ft thick. For
larger depths of pervious strata, it is usually more economical to have well
systems with 50 percent or greater penetration at closer spacing. The equa-
tions for relief well design depend upon the values of the source of seepage
and seepage exit length as shown in figure 9-25. The source of seepage is
assumed to be a line source parallel to the well system and the dam axis. The
location of the source of seepage depends upon the thickness and vertical
permeability of any natural top stratum upstream of the dam and any impervious
blanket constructed upstream of the dam, the permeability and thickness of the
pervious foundation, and the presence of any borrow pits and/or major erosion
features which reduce the thickness of the top stratum (see procedure to
evaluate the source of seepage given in U. S. Army Engineer Waterways Experi-
ment Station 1956a). The value of the seepage exit (X3 in figure 9-25)
depends upon the thickness and permeability of the top stratum downstream from
the toe of the dam, the thickness and permeability of the pervious substratum,
9-65
EM 1110-2-1901
30 Sep 86
Figure 9-24. Profile of typical design reaches for relief

well analysis (prepared by WES)
and the presence of any geologic features and/or man-made features which would
result in an open or blocked seepage exit. The procedure for computation of
the seepage exit distance, rate of discharge through the pervious foundation
per unit length of dam, and pressure head without relief wells is given in
figure 9-26. Generally relief wells have diameters of 6 to 18 in. and screen
lengths of 20 to 100 ft, depending on the requirements. Some types of screens
used for wells are slotted or perforated steel pipe, perforated steel pipe
wrapped with steel wire, slotted wood stave pipe, and slotted plastic pipe.
Riser pipe usually consist of the same material as the screen but does not
contain slots or perforations. The open area of a well screen should be suf-
ficiently large to maintain a low entrance velocity (< 0.1 ft per sec) at the
design flow in order to minimize head losses across the screen and reduce the
incrustation and corrosion rates. The entrance velocity is calculated by
dividing the expected or desired yield of the relief well by the area of open-
ings in the screen (Johnson Division, Universal Oil Products Co. 1972). Fil-
ter packs around relief wells are usually 6 to 8 in. thick and must meet the
criteria specified in Appendix D. Head losses within the relief well system
consist of entrance head loss, friction head loss in the screen and riser
pipes, and velocity head loss as shown in figure 9-27. The effective well
radius is that radius which would exist if there were no hydraulic head loss
into the well. For a well without a filter, the effective well radius is
one-half the outside diameter of the well screen. Where a filter has been
placed around the well, the effective well radius is the outside radius of the
well screen plus one-half of the thickness of the filter.
9-66
EM 1110-2-1901
30 Sep 86
Figure 9-25. Design of relief wells (prepared by WES)
(2) Effective Well Penetration. In a stratified foundation, the effec-

tive well penetration usually differs from that computed from the ratio of the
length of well screen to the total thickness of the aquifer. The procedure
for determining the required length of well screen to achieve an effective
penetration in a stratified aquifer is as follows. Each stratum of the per-
vious foundation is first transformed into an isotropic layer (Leonards 1962)
(9-24)
where
d = transformed layer thickness
d' = actual layer thickness
k H = horizontal permeability of layer
k v = vertical permeability of layer
9-67
EM 1110-2-1901
30 Sep 86
Seepage per unit length of dam •••• Q

s
Head beneath top stratum of

downstream toe of dam •.•.••.•.•••• h
0
=
A factor •••••••••••••••••••••••••• c =
CASE l, 1
3
= ...
X ., -
1
3 c
h =h t-cx
X 0
(t = 2. 718)
1
x3 =c tanh(cL )
3
cosh· c(L
3
- x}
h
X = h
0 cosh(cL )
3
h
0
h
x=L3 = cosh (ct )
3
CASE 3, OPEN EXIT
sinh c{L - x)
3
hx = ho sinh (cL )
3
hx=L =0
3
NOTE: x can be COI:If·t:.ted ~r-on. formulas for x 1 by
1
inserting the ~ength of upstrerum blanket L1 fO~ L
in the appropn.ute exnressi<"'r, when upstrenm condl.tiods
are similar to the above do~nstreP~ conditions.
Figure 9-26. Computation of rate of discharge and pressure heads for semi-
pervious downstream top stratum and no relief wells (from U. S. Army
120
Engineer Waterways Experiment Station )
9-68
TYPE OF PIPE c
STEEL \NEW) I ?5 O.fi7
STEEL IAVG CONDITION! ,, a· 0.83
WOOD STAVE RISER 130 0.71
fOTAL. HYDRAULIC HEAD LOSS IN A WELL (Hw) IS WOOD STAVE SCF<EEN 100 .00
CORRUGATED M£TAL 70 .92
Jr
~ r"
~"<HERE He ENTRANCE HEAD LOSS (SCREEN AND FILTER)
t- Jr
ESTIMATE FROM CURVE a "- 0
Ul •
H HEAD LOSS IN SCREENED SECTION OF WELL "'I
0_j
' fSTIMATE FROM CURVE b FOR A DISTANCE OF ONE- w
•
HI\LF THE SCREEN LENGTH

..
on.
-
wn.
H, Ht.~D L 055 WITHIN THE RISER AND CONNECTIONS. I U.
EST tMA 1 E FROM CURVE b z 0

Of-
f="-
uo
-0
ll -
H VELOCITY HEAD LOSS. ESTIMATE FROM CURVE c. "-
v
10 100 1,000 10,000
THL VAL.UE OF 11 MUST BE SUBTRACTED FROM THE COMPUTED
DISCHARGE, GPM
V•\LUE OF- h"' TO OBTAIN THE LII"T OR WATER LEVEL IN A WELL.
BASED ON HAZEN-WILLIAMS EQUATION WITH C = 100; MULTIPLY
LOSSES BY I \00/C)I.BS FOR VALUES OF C OTHER THAN 100
9-69
(b)
\0.0
3.0~-------.----.----.----.---.----,----,----,----,
s.o
w f-
u "- 3.0
z ..::
<t
Jr f- !
';; "- 1.0
LJj I 2,0 "'"'0 0.5
z • _j
WI
w-
Jr Ul ..
0
w
0.3
~0 I
.J
u }- 0.1
"7 {J t .0 DEVELOPED FROM----- 1-
;;: "w PUMPING TEST u 0.05
rr r DATA. 16-IN.-DIAM 0
w .J
WELL,\00-SQ-IN SCREEN w 0.03
EM 1110-2-1901
f-
_j
WITH 5/32-IN SLOTS, >
6-IN.·THICK FILTER
Co - - - J. - - - -- _j ____..L____j__ _..L__ _ i_ _.J....____j_ _....l
0 1C 20 30 40 so
30 Sep 86
Wl:LL DISCHARGE PER FT OF SCREEN, GPM DISCHARGE, GPM
(a) (c)
205
Figure 9-27. Hydraulic head loss in relief wells (after Leonards )
EM 1110-2-1901
30 Sep 86
The transformed permeability of each layer of the pervious foundation is
(9-25)
where is the transformed permeability of layer. The thickness of the

transformed, homogeneous, isotropic pervious foundation is
(9-26)
where D is the thickness of pervious foundation. The effective permeability

of the transformed pervious foundation is
(9-27)
where k is the effective permeability of transformed pervious foundation.

The effective well screen penetration into the transformed pervious foundation
is
(9-28)
where
w = effective well screen penetration into transformed pervious

foundation
= actual well screen length
The percent penetration of the well screen into the transformed pervious
foundation is
9-70
EM 1110-2-1901
30 Sep 86
(9-29)
where is the actual pervious foundation thickness.
(3) Factor of Safety. The factor of safety against uplift or heaving

at the downstream toe of the ham, based upon the critical gradient, is
(9-30)
where
Fh = factor of safety against uplift or heaving at the downstream toe

of the dam
i C R = critical upward hydraulic gradient under the top stratum at the

downstream toe of the dam
iO = allowable upward hydraulic gradient under the top stratum at the

downstream toe of the dam
= submerged unit weight of downstream top stratum soil
ha = allowable pressure head under the top stratum at the downstream

toe of the dam
Z b L = thickness of downstream top stratum
The factor of safety against uplift or heaving at the downstream toe of the dam
(1)
provided by the relief well system should be at least 1.5 .
(4) Infinite and Finite Relief Well Systems. Formulas for the design
of relief wells are based on the assumption that the flow is laminar,
(1) Relief wells should be designed to reduce the excess head to zero to pre-
vent upward seepage from occurring beneath the downstream top stratum.
9-71
EM 1110-2-1901
30 Sep 86
artesian, and continuous and that a steady-state condition exists. Relief well
systems are considered to be infinite or finite in length. The term infinite
is applied to a system of wells that conforms approximately to the following
idealized conditions:
(a) The wells are equally spaced and identical in dimensions.
(b) The pervious foundation is of uniform depth and permeability along

the entire length of the system.
(c) The effective source of seepage and the effective line of

downstream exit are parallel to the line of wells.
(d) The boundaries at the ends of the relief wells are impervious,
normal to the line of wells, and at a distance equal to one-half the well
spacing beyond the end wells of the system.
If these conditions exist, the flow to each well and the pressure distribution
around each well are uniform for all wells along the line. Therefore, there
is no flow across planes centered between wells and normal to the line, hence
no overall longitudinal component of flow exists anywhere in the system. The
term infinite is applied to such a system because it may be analyzed mathe-
matically by considering an infinite number of wells; the actual number of
wells in the system may be from one to infinity. Normally, a line of relief
wells below a dam extending entirely across a valley and terminating at rela-
tively impervious valley walls should be designed as an infinite line. A
finite system of wells in any system that does not approximate the idealized
condition for the infinite system. Whenever a major and abrupt change in the
character of the system such as penetration or well spacing might result in
an appreciable component of flow parallel to the line of wells, the use of
design procedures for finite systems will be used (see U. S. Army Corps of
Engineers 1963).
(5) Drawdown to Infinite Line of Fully Penetrating Relief Wells with

Impervious Top Stratum. Where the flow to an infinite line of fully pene-
trating relief wells is from an infinite line source and the top stratum is
assumed to be completely impervious, (1) as shown in figure 9-28. The drawdown
produced by an equivalent continuous slot is
(9-31)
(1) Also applicable-when the top stratum is semipervious provided the well
system is located in a drainage ditch and the head is kept below the
ground surface on the downstream side of the dam to prevent any seepage
upward through the top stratum. Under these conditions, the downstream
top stratum acts as if it is impervious.
9-72
EM 1110-2-1901
30 Sep 86
where
H - he = drawdown produced by flow from continuous slot
Q w = discharge from equivalent continuous slot
L = distance from line source of seepage to wells
k = effective permeability of transformed pervious foundation
D = thickness of transformed pervious foundation
a = well spacing
However, an additional head occurs because of converging flow at the wells.

This head loss is a function of well flow, well spacing and penetration, well
radius, and thickness and permeability of the pervious foundation. For fully
penetrating wells
(9-32)
where
= head loss at well due to converging flow (see figure 9-28)
r w = effective radius of well (outside radius of well screen plus

one-half of the thickness of the filter)
The total drawdown at the well, neglecting hydraulic head losses in the well,
is that at the slot plus that due to the well
(9-33)
Substituting equations 9-31 and 9-32 into equation 9-33
(9-34)
9-73
EM 1110-2-1901
30 Sep 86
a. Plan view of wells
b. Section A-A c. Section B-B
Figure 9-28. Flow to an infinite line of fully penetrating

relief wells from an infinite line source of seepage (after
Leonards 2 0 5 )
The head midway between wells will exceed the head at the well by
(9-35)
where is the head increase midway between wells. The drawdown midway
between wells is
(9-36)
At a distance downstream from the well system, the head will exceed that at
the well by
9-74
EM 1110-2-1901
30 Sep 86
(9-37)
where is the head increase downstream of wells. The drawdown downstream

of the wells is
(9-38)
(6) Drawdown to Infinite Line of Partially Penetrating Relief Wells

with Impervious Top Stratum. For an infinite line of partially penetrating
relief wells where the flow is from an infinite line source and the top stratum
is assumed to be impervious (or semipervious as previously described) the head
loss at the partially penetrating well due to converging flow is
(9-39)
where is the average uplift factor (obtained from figure 9-29). The
total drawdown at the partially penetrating well, neglecting hydraulic head
losses in the well, is that at the slot plus that due to the well
(9-40)
The head midway between partially penetrating wells will exceed the head at
the well by
(9-41)
where
is the midpoint uplift factor (obtained from figure 9-29). The draw-
down midway between partially penetrating wells is
(9-42)
9-75
EM 1110-2-1901
30 Sep 86
a.lrw
~i
.
-
~ '
t· ~
I - 0.5
o.l
- - - -
I
·- I
I
~ru
\()0 - - -
•.!!.
0
'!-0 It i
10
iS /!'l
l
I
- -- I
2' >---
,o ! -- E
Q I I CD
....
...."'"'
~!l !
--
- --
f-j I > a>

E
•" KEY
;
w
z ,
~
41
w I I
>
9-76
Q_
.I I "
a>
_J
.J
cJ
~
t[
II
IIj ---- -f-
I
20
.... ·~ '1 j
I
2 EXAMPLE.
II
w
<)
a: ,.;: Where Q /rw :100, 9av~~ I 0, 9'6-= 1.0
w
"- Step 1.· Locate Sav: I 0 on ca/rw line a 100 and mark point
2>
- Step 2. Locate 0 /Q = 1.0 on Ola. hne for 9av
~; Step ·3 Place tr•onqle edc;~e on obova point$ located by steps I ond 2.
/
Step 4 Slide tnonqle parallel to intttol position to po.nt 0/a : 1.00 on Q/a line for 9m
Reod off value at 9m on a/rw l1ne = 100.
--
-1 - - -
-
3.0
Step 5.
Step 6.
em: 1.05
Slide tnongle parallel to 1nil!ol posttion to pole potnt on 0/a line of Oav
Step 7. Read off percent screen penetrotton = 46% on alrw ""1000
g g
N
qlr,.
Figure 9-29. Nomograph for obtaining uplift factors for design of partially penetrating
120
relief wells (from U.S. Army Engineer Waterways Experiment Station )
EM 1110-2-1901
30 Sep 86
(7) Drawdown to Infinite Line of Relief Wells with Semipervious Top

Stratum. Where the top stratum is semipervious, the need for relief wells is
evaluated by determining the piezometric grade line without relief wells using
blanket formulas given in figure 9-26. As stated previously, the factor of
safety against uplift or heaving at the downstream toe of the dam, as deter-
mined from equation 9-30, should be at least 1.5. If relief wells are
required, the spacing for an infinite line of relief wells for a given pene-
tration is determined using a procedure of successive trials and the nomograph
given in figure 9-29. The required well spacing is affected by hydraulic head
losses in the well which are estimated from figure 9-27. The procedure for
computing the well spacing is as follows:
(a) Compute the allowable pressure head under the top stratum at the
downstream toe of the dam from
(9-43)
where
h a = allowable pressure head under the top stratum at the downstream

toe of the dam
= submerged unit weight of downstream top stratum soil
Z b l = thickness of downstream top stratum
Fh = factor of safety against uplift or heaving at the downstream toe

of the dam
(b) Assume that the net head in the plane of the wells equals the
allowable pressure head under the top stratum at the downstream toe of the dam
and compute the net seepage gradient toward the well line
(9-44)
where
net seepage gradient toward the well line
h = net head acting on the dam
H a v g = net head in the plane of the wells
9-77
EM 1110-2-1901
30 Sep 86
S = distance from line of relief wells to effective source of seepage

entry (see procedure in U. S. Army Engineer Waterways Experiment
Station 1956a)
= distance from line of relief wells to effective seepage exit (see

X3
procedure in figure 9-26)
Setting Havg = ha in equation 9-44 gives
(9-45)
(c) Assume a well spacing and compute the flow from a single well
(9-46)
where
Qw = flow from a single well
kf = effective permeability of transformed pervious foundation
D = transformed thickness of pervious foundation
a = well spacing
= average uplift factor (obtained from figure 9-29)
(d) Estimate the total hydraulic head loss in the well from
figure 9-27.
(e) Compute the net average head in the plane of wells above the total
head loss in the well including elevation head loss (see figure 9-25) from
(9-47)
where
h a v g = net average head in the plane of wells above the total head loss
in the well including elevation head loss
9-78
EM 1110-2-1901
30 Sep 86
H a v g = net head in the plane of wells
H W = total head loss in the well including elevation head loss
(f) Substitute values obtained from and ha v g from equation 9-45

and 9-47, respectively, and solve for the average uplift factor
(9-48)
where is the average uplift factor.
(g) Find from figure 9-29 using the values of a used in

equation 9-48 and the corresponding a/rw and D/a values.
(h) The first trial well spacing is that of value a for which
from step (f) equals from step (g).
(i) Find from figure 9-29 for the first trial well spacing and the
corresponding values of a/rw and D/a .
(j) If repeat steps (c) to (i) using the first trial well
spacing in lieu of the spacing originally used in step (c), and determine the
second trial well spacing. This procedure should be repeated until relatively
consistent values of a are obtained on two successive trials. Usually the
second trial spacing is sufficiently accurate.
If in step (j), a modified procedure is used for the second trial:
(k) Assume Hm = ha and compute QW from equation 9-46 using the

value of AM obtained in step (b) and the first trial well spacing from
step (h).
(1) Estimate Hw from Qw of step (k) and figure 9-27.
(m) Compute the net head beneath the top stratum midway between the
wells above the total head loss in the well including elevation head loss (see
figure 9-25) from
(9-49)
where
hm = net head beneath the top stratum midway between the wells above the
total head loss in the well including elevation head loss
9-79
EM 1110-2-1901
30 Sep 86
= net head beneath the top stratum midway between the wells
Hm
= total head loss in the well including elevation head loss
HW
(n) Using from steps (h) and (i), respectively, compute
ha v g from
(9-50)
where h a v g is the net head in the plane of wells.
(o) Using Hw and h a v g from steps (1) and (n), respectively, and
compute H a v g f r o m
(9-51)
(p) Compute from equation 9-44 using H a v g from step (o).
(q) Using hm and from steps (m) and (p), respectively, compute
for various values of a from
(9-52)
where is the midpoint uplift factor.
(r) Find from figure 9-29 for the values of a used in step (q)
and the corresponding a/r w and D/a values.
(s) The second trial well spacing is that value of a which from
step (q) equals from step (r).
(t) Find from figure 9-29 for the second trial well spacing and
the corresponding values of a/rw and D/a .
(u) Determine the third trial well spacing by repeating steps (k) to
(t) using the second trial well spacing in lieu of the spacing originally
assumed in step (k), and in step (n) using the values of and from
steps (s) and (t), respectively, instead of those from steps (h) and (i). This
procedure should be repeated until relatively consistent values of a are
obtained on two successive trials. Normally, the third trial is sufficiently
accurate.
9-80
EM 1110-2-1901
30 Sep 86
(8) Drawdown to Finite Line of Relief Wells. In a short, finite line of

relief wells, the heads midway between wells exceed those for an infinite line
of wells both at the center and near the ends of the well system as shown in
figure 9-30. Note that the pressures between wells, or midpoint pressures, are
lower at the center of the well system and gradually increase towards the end
of the line. With an infinite line of wells, the heads midway between wells
are constant along the entire length of the well line. Numerous well systems
may be fairly short, and for these it will be necessary to reduce the well
spacing computed for an infinite line of wells so that heads midway between
wells will not be more than the allowable pressure head under the top stratum
at the downstream toe of the dam. The ratio of the head midway wells at the
center of finite well systems to the head between wells in an infinite line of
wells, for various well spacings and seepage exit lengths, is given in
figure 9-31. The spacing of relief wells in a finite line should be the same
as that required in an infinite line of wells to reduce the head midway
between wells to ha divided by the ratio of from figure 9-31. In
I
any finite line of wells of constant penetration and spacing, the head midway
between wells near the ends of the system exceeds that at the center of the
system. Thus at the ends of both short and long well systems, the relief
wells should generally be made deeper to provide additional penetration of the
pervious substratum so as to obtain the same head reduction as in the central
part of the well line. The above-mentioned procedures for designing finite
relief well systems, although approximate, are usually sufficient. More exact,
but more complex, procedures are available (see U. S. Army Corps of Engineers
1963).
Figure 9-30. Variation of pressure relief along a finite line of relief

wells (after EM 1110-2-1905)
9-81
4.0
Assumptions:
EM 1110-2-1901
30 Sep 86
~ .!. x3
(1) Penetration of pervious aquifer by ve11 screen = 100%
1 50 250 {2) Semipervious top stratum infinite in landvard extent
{3) Effective ve11 radius = 1.0 ft
3.~ 2 10() 500-
Note:
3 150 '500
(l)H = Head midvay betveer. vells in infinite vell line
4 200 500 m...
(2)Hm = Head midvay betveen vells at center of a finite

:Lt"tl
Q 2.~
n vell line
....
..{
IX'
(3) a= Well specing in feet
9-82
(4) x
3
= Exit length in feet
l.O r-----~~r\--------t---------T---~----T---------+---------+---------+---------J
200 400 600 800 1000 1200 I 00 1600

LENGTH Of WELL REACH IN fEET
Figure 9-31. Ratio of head midway between relief wells at center of a finite well system to
head midway between wells in an infinite system (from U.S. Army Engineer Waterways Experi-
.
ment s tat1on
120)
EM 1110-2-1901
30 Sep 86
d. Installation. While the specific materials used in the construction

of relief wells and methods of installation differ, relief wells are basically
very similar. They consist of a boring to facilitate the installation, a
screen or slotted pipe section to allow the entrance of ground water, a filter
to prevent entrance and ultimate loss of foundation material, a riser pipe to
conduct the water to the ground surface, a check valve to prevent backflooding
and entrance of foreign material detrimental to the installation, backfill to
prevent recharge of the formation by surface water, a bottom plug to prevent
inflow of soil, (1) a V-notch weir at the top of the relief well to facilitate
measurement of flow, and a cover and some type of barricade protection to pre-
vent vandalism and damage to the top of the well by maintenance crews, live-
stock, etc. (see figure 9-23). Following development of the relief well, a
pumping test should be conducted to determine the specific yield of the well
and the amount of sand infiltration. Information from the pumping test is
used to determine the acceptability of the well and for evaluating any changes
in performance or loss of efficiency with time. Procedures for installation,
development, and pumping tests are given in EM 1110-2-1913. A guide specifi-
cation for relief wells is available.
e. Monitoring. As mentioned previously, the discharge of relief wells

gradually decreases with time due to clogging of the well screen and/or reser-
voir siltation. Piezometers should be installed between relief wells to
determine the seepage pressure in the main pervious strata. Relief wells
should be sounded for sand and pumped to determine their discharge capacity
under varying reservoir levels (see Chapter 13). A trend toward fall in
relief well discharge accompanied by a fall in piezometric levels indicates a
decrease in underseepage due to reservoir siltation and is favorable. How-
ever, a decrease in relief well discharge accompanied by a rise in piezometric
levels indicates clogging of the relief wells and immediate rehabilitation
and/or replacement of the wells or installation of additional wells is
required (Singh and Sharma 1976). The operation, maintenance, and rehabili-
tation of relief wells is discussed in Chapter 14.
9-8. Trench Drain.

costly, a trench drain may be used in conjunction with other underseepage con-
trol measures (upstream impervious blanket and/or relief wells) to control
underseepage. A trench drain is a trench generally containing a perforated
collector pipe and backfilled with filter material (see figure 9-32). Trench
drains are applicable where the top stratum is thin and the pervious founda-
tion is shallow so that the trench can penetrate into the aquifer. The exis-
tence of moderately impervious strata or even stratified fine sands between the
bottom of the trench drain and the underlying main sand aquifer will render
the trench drain ineffective. Where the pervious foundation is deep, a trench
drain of practical depth would only attract a small portion of underseepage,
and detrimental underseepage would bypass the drain and emerge downstream of
the drain, thereby defeating its purpose. Trench drains may be used in
(1) For partially penetrating relief wells, the bottom plug should be such
that future screen extension will be possible,
9-83
EM 1110-2-1901
30 Sep 86
conjunction with relief well systems to collect seepage in the upper pervious
foundation that the deeper relief wells do not drain. If the volume of seepage
is sufficiently large, the trench drain is provided with a perforated pipe. A
trench drain with a collector pipe also provides a means of measuring seepage
quantities and of detecting the location of any excessive seepage (U. S. Army
Engineer Waterways Experiment Station 1956a, EM 1110-2-1911, EM 1110-2-1913,
and Cedergren 1977).
b. Location and Geometry. Trench drains are generally located at the

downstream toe of the dam as shown in figures 9-32a and 9-32c, but are some-
times located beneath the downstream slope of the dam as shown in figure 9-32b.
Trench geometry will depend on the volume of expected underseepage, desired
reduction in uplift pressure, construction practicalities, and the stability of
the material in which the trench is to be excavated. Trenches with widths as
small as 2 to 6 ft have been used. However, narrow trench widths require
special compaction equipment (EM 1110-2-1913).
c. Design Considerations. The maximum head at the base of an imper-

vious top stratum downstream of a toe trench drain overlying a homogeneous,
isotropic, pervious foundation may be computed from figure 9-33. The distance
to the source of seepage may be evaluated using the procedure given in U. S.
Army Engineer Waterways Experiment Station 1956a. If the pervious foundation
is stratified, it is transformed into an isotropic layer, as described pre-
viously (see Equations 9-24 to 9-27) prior to using figure 9-33. The' factor of
safety against uplift or heaving at downstream toe of the dam provided by the
trench drain should be at least 1.5. If the downstream top stratum is semi-
pervious, seepage into the trench, and the maximum head landward of the
trench, will be somewhat less than that computed from figure 9-33 giving a
slightly conservative design. When there is no downstream top stratum, seep-
age flow into the trench can be estimated from a flow net analysis (U. S. Army
Engineer Waterways Experiment Station 1956a).
d. Construction. A trench drain usually contains a perforated pipe,

surrounded by filter gravel, and backfilled with sand as shown in figure 9-34.
Materials in trench drain must satisfy the filter gradation criteria given in
Appendix D. As filter materials are placed, they must be protected from
contamination resulting from inwash that might occur during a rainfall. The
same control procedures are used for trench drains as those used in construc-
tion of pervious fill in the main embankment (EM 1110-2-1911).
9-9. Concrete Galleries. Internal reinforced concrete galleries have been

used in earth and rockfill dams built in Europe, for grouting drainage, and
monitoring of behavior. Galleries have not been constructed in embankment
dams built by the Corps of Engineers to date. Some possible benefits to be
obtained from the use of galleries in earth and rockfill dams are as follows
(Sherard et al. 1963):
a. Construction of the embankment can be carried out independently of

the grouting schedule.
9-84
EM 1110-2-1901
30 Sep 86
a. Trench drain at downstream toe of dam
b. Trench drain under downstream slope of dam
c. Trench drain used in conjunction with relief wells
Figure 9-32. Trench drains to control underseepage (from

EM 1110-2-1913)
9-85
EM 1110-2-1901
30 Sep 86
Figure 9-33. Design of toe trench drains for homogeneous, isotropic,

pervious foundation, and for an impervious downstream top stratum
120
(from U. S. Army Engineer Waterways Experiment Station )
b. Drain holes drilled in the rock foundation downstream from the grout
curtain can be discharged into the gallery and observations of the quantities
of seepage in these drain holes will indicate where foundation leaks are
occurring.
c. Galleries provide access to the foundation during and after reservoir

filling so that additional grouting or drainage can be installed, if required,
and the results evaluated from direct observations.
d. The additional weight of the overlying embankment allows higher

grout pressures to be used.
e. Galleries can be used to house embankment and foundation instrumen-

tation outlets in a more convenient fashion than running them to the down-
stream toe of the dam.
f. If the gallery is constructed in the form of a tunnel below the rock

surface along the longitudinal axis of the dam, it serves as an exploratory
tunnel for the rock foundation.
9-86
EM 1110-2-1901
30 Sep 86
Figure 9-34. Trench drain with collector

pipe (from EM 1110-2-1913)
The minimum size cross section recommended for galleries and access shafts is 8
by 8 ft to accommodate drilling and grouting equipment. A gutter located along
the upstream wall of the gallery along the line of grout holes will carry away
cuttings from the drilling operation and waste grout from the grouting opera-
tion. A gutter and system of weirs located along the downstream wall of the
gallery will allow for determination of separate flow rates for foundation
drains (EM 1110-2-3502, and Blind 1982).
9-87
EM 1110-2-1901
30 Sep 86
CHAPTER 10
SEEPAGE CONTROL THROUGH EARTH ABUTMENTS ADJACENT TO STRUCTURES
AND BENEATH SPILLWAYS AND STILLING BASINS
10-1. Through Earth Abutments. Earth and rock-fill dams, particularly in

glaciated regions, may have pervious material, resulting from filling of the
preglacial valley with alluvial or morainal deposits followed by the down-
cutting of the stream, in one or both abutments (Twelker 1957). Seepage
through the pervious abutment(s) is combined with through seepage and under-
seepage to determine the total seepage loss. As mentioned previously, the pur-
pose of the project, i.e., long-term storage, flood control, hydropower, etc.,
will determine the allowable seepage loss. Seepage control through earth
abutments is provided by extending the upstream impervious blanket in the
lateral direction to wrap around the abutment up to the maximum water surface
elevation, by placing a filter layer between the pervious abutment and the dam
downstream of the impervious core section, and, if necessary, by installing
relief wells at the downstream toe of the pervious abutment. At the North
Branch of Kokosing Dam, Ohio, the left abutment is an outwash terrace consist-
ing of sands and gravels with layers of silt and clay as shown in figure 10-1.
Seepage control through the pervious abutment was provided by a 5-ft-thick
impervious upstream blanket which wrapped around the left abutment, a filter
layer between the pervious abutment and the dam downstream of the impervious
core (see figure 10-2), and three fully penetrating relief wells at the down-
stream toe of the pervious abutment (U. S. Army Engineer District, Huntington
1969).
10-2. Adjacent to Outlet Conduits. When the dam foundation consists of com-
pressible soils, the outlet works tower and conduit should be founded upon or
in stronger abutment soils or rock. When conduits are laid in excavated
trenches in soil foundations, concrete seepage cutoff collars shall not be
provided solely for the purpose of increasing seepage resistance since their
presence often results in poorly compacted backfill around the conduit. Col-
lars, with a minimum projection from the conduit surface, will be used over
conduit joints to protect against joint displacements resulting from differen-
tial movement on yielding foundations. Excavations for outlet conduits in
soil foundations shall be wide enough to allow for backfill compaction
parallel to the conduit using heavy rolling compaction equipment. Equipment
used to compact along the conduit should be free of framing that prevents its
load transferring wheels or drum from working against the structure. Excavated
slopes in soil for conduits should be no steeper than 1V to 2H to facilitate
adequate compaction and bonding of backfill with the sides of the excavation.
Drainage layers should be provided around the conduit in the downstream zone of
embankments without pervious shells. A concrete plug shall be used as backfill
in rock cuts for cut-and-cover conduits within the core zone to ensure a water-
tight bond between the conduit and vertical rock surfaces. The plug, which can
be constructed of lean concrete, should be at least 50 ft long and extend up to
the original rock surface. In embankments having a random or an impervious
downstream shell, horizontal drainage layers should be placed along the sides
and over the top of conduits downstream of the impervious core. Where outlet
structures are to be located in active seismic areas, special attention must be
given to the possibility of movement along-existing or possibly new faults
(EM 1110-2-2300).
10-1
EM 1110-2-1901
30 Sep 86
EL. 1169 TOP Of DAM
------------------------------------
1150
[
--------c------ 1100
10-2
I
I
L SANDS &
GRAVELS\
ASSUMED SEEPAGE
CHANNEL LIMITS
END SEEPAGE
CUTOff
1050
----------t------ T1LL"\.
25+00 ' 15+00
Figure 10-1. Profile of left pervious abutment at North Branch of Kokosing Dam, Ohio
87
(from U. S. Army Engineer District, Huntington )
EM 1110-2-1901
30 Sep 86
Figure 10-2. Seepage control of left pervious abutment at North Branch

of Kokosing Dam, Ohio (from U.S. Army Engineer District, Huntington 87 )
10-3. Beneath Spillways and Stilling Basins. Adequate drainage should be

provided under floor slabs for spillways and stilling basins to reduce uplift
pressures. For soil foundations, a drainage blanket under the slab with
transverse perforated pipe drains discharging through the walls or floor is
generally provided, supplemented in the case of stratified foundations by deep
well systems. Usually drainage of a slab on rock is accomplished by drain
holes drilled in the rock with formed holes or pipes through the slab. The
drainage blanket is designed to convey the seepage quickly and effectively to
the transverse collector drains. It is designed as a graded reverse filter
with coarse stones adjacent to the perforated drain pipe and finer material
adjacent to the concrete structure to prevent the migration of fines into the
drains. Outlets for transverse drains in the spillway chute discharge through
the walls or floor at as low an elevation as practical to obtain maximum pres-
sure reduction. Wall outlets should be 1 ft minimum above the floor to pre-
vent blocking by debris. Cutoffs are provided at each transverse collector
10-3
EM 1110-2-1901
30 Sep 86
pipe to minimize buildup of head in case of malfunction of the pipe drain.

Drains should be at least 6 in. in diameter and have at least two outlets to
minimize the chance of plugging. Outlets should be provided with flat-type
check valves to prevent surging and the entrance of foreign matter in the
drainage system. For the stilling basin floor slab, it may be advantageous to
place a connecting header along each wall and discharge all slab drainage into
the stilling basin just upstream from the hydraulic jump at the lowest practi-
cal elevation, in order to secure the maximum reduction of uplift for the
downstream portion of the slab. A closer spacing of drains is usually
required than in the spillway chute because of greater head and considerable
difference in water depth in a short distance through the hydraulic jump.
Piezometers should be installed in the drainage blanket and deeper strata, if
necessary, to monitor the performance of the drainage systems. If the drains
or wells become plugged or otherwise noneffective, uplift pressures will
increase which could adversely affect the stability of the structure
(EM 1110-2-2400 and EM 1110-2-2300).
10-4
EM 1110-2-1901
30 Sep 86
CHAPTER 11
SEEPAGE CONTROL IN ROCK FOUNDATIONS AND ABUTMENTS
11-1. General Considerations.
a. The choice of seepage control methods to use in rock foundations and

abutments is dependent on a number of factors. Characterization of the founda-
tion or abutment and identification of potential seepage paths is essential.
Before any method of seepage control is implemented, the area must be
thoroughly explored and tested to assure that the method chosen will apply to
the general conditions as well as the conditions locally encountered and will
serve the intended purpose. In many cases, a combination of methods can be
used to the best advantage for rock foundations or abutments. The use of dif-
ferent control methods becomes particularly important when there is a change in
the character of the foundation from one location to another, or a change in
seepage characteristics between the foundation and the abutment.
b. Seepage should be cut off or controlled by drainage whenever economi-

cally possible. Safety, however, must be the governing factor for selecting a
seepage control method. It should be noted that the possibility exists for
control measures to cause substantial increases in seepage rather than
decreases. Such increases are normally accompanied by reductions in uplift
pressures and are therefore desirable if the increased seepage produces no
detrimental side effects. In the final choice of a seepage control method, or
methods, economic factors must be recognized and evaluated.
11-2. Cutoff Trenches.
a. No cutoff is 100 percent impervious and therefore the reduction in

seepage from cutoff trenches is a relative matter. Cutoff trenches are nor-
mally employed where the character of the foundation is such that the con-
struction of a satisfactory or effective grout curtain is not practical. Such
trenches, when constructed, are normally backfilled with compacted impervious
material, bentonite slurry, or neat cement.
b. Construction of trenches in rock foundations and abutments normally

involves blasting using the presplit method with primary holes deck-loaded
according to actual foundation conditions. After blasting, excavation is nor-
mally accomplished with a backhoe. Cutoff of seepage within the foundation is
obtained by connecting an impervious portion of the foundation to the imper-
vious portion of the structure by backfilling the trench with an impervious
material. In rock foundations, as in earth foundations, the impervious layer
of the foundation, in some cases, may be sandwiched between an upper and a
lower pervious layer, and a cutoff to such as impervious layer would reduce
seepage only through the upper pervious layer. However, where the thicknesses
of the impervious and upper pervious are sufficient, the layers may be able to
resist the upward seepage pressures existing in the lower pervious layer and
thus remain stable. Cutoff of seepage within the abutment is normally
obtained by extending the cutoff from above the projected seepage line to an
impervious layer within the abutment. The type of backfill material is nor-
mally dictated by condition of the foundation or abutment, economics, and
degree of cutoff required.
11-1
EM 1110-2-1901
30 Sep 86
11-3. Abutment Impervious Blankets.
a. Impervious blankets overlying the upstream or riverside face of per-

vious abutments, or foundations, are effective in reducing the quantity of
seepage and to some extent will reduce uplift pressures and gradients down-
stream. An impervious blanket may be used for earthen or rock abutments;
however, a filter material is normally required with rock abutments.
b. The construction of impervious blankets is particularly adaptable to

treating exposed pervious areas of abutments which are adjacent to the main
structure. In cases where a natural impervious blanket exists on the abutment,
ranging in depth from a few feet to many feet, full advantage should be taken
of the existing material. Upstream borrow along the abutment should be con-
trolled to prevent excessive excavation of the natural impervious top blanket.
Conversely, localized areas which are thin and weak should be reinforced by
the addition of additional impervious material.
c. Blankets may sometimes give adequate control of seepage water for low
head structures, but for high head structures it is usually necessary to incor-
porate a downstream drainage system as a part of the overall seepage-control
design. The benefits derived from abutment impervious blankets are due to the
dissipation of a part of the reservoir head through the blanket. The propor-
tion of head dissipated is dependent upon the thickness, length, and effective
permeability of the blanket in relation to the permeability of the adjacent
soil, or rock.
11-4. Drainage and Grouting Galleries and Tunnels.
a. Foundation galleries and tunnels in concrete gravity dams provide an

exit for foundation drains and convenient facilities for rehabilitation work,
or supplemental grouting, if required. The depth of drainage galleries, or
exit elevation of the drains, with respect to the tailwater, controls the
uplift downstream of the drains or wells. Generally, the lower the elevation
of the gallery, the more reduction in head, or uplift, is experienced. If the
depth of the drainage gallery is located and the gallery discharges at the
elevation of the tailwater, the magnitude of the uplift downstream of the wells
is normally very modest. The uplift, however, is controlled by well spacing,
well efficiency, and other seepage control measures, such as grout curtains and
cutoffs, in addition to the elevation of the drainage gallery and the elevation
at which it discharges. In special cases where the drainage gallery is very
deep, i.e., below tailwater, it is possible to actually create a negative up-
lift on the base of a dam. In such cases some of the seepage pumped from the
drainage gallery flows through the foundation to the drains from downstream.
b. Grouting curtains are frequently centered along drainage galleries or

tunnels. Remedial or supplemental' grouting may be performed from within
drainage tunnels. The additional grouting may be performed vertically or in
inclined, or sloping, boreholes. Also, should excessive uplift pressures
become evident, additional grouting to widen or deepen the curtain may be
performed or additional drainage wells may be installed from the gallery to
relieve excess pressures.
11-2
EM 1110-2-1901
30 Sep 86
11-5. Grouting of Foundations and Abutments.
a. The Corps‘ grouting methods have become less standardized in recent

years (Albritton, Jackson, and Banget 1984). Typically, however, a combination
of a line of drainage holes and a grout curtain provide an efficient and effec-
tive seepage control method. A properly installed grout curtain in the founda-
tion of a structure not only provides a substantial reduction in seepage but
also reduces the uplift pressures downstream of the curtain. Conveniently, for
the great majority of dam sites, irrespective of type of rock, strike and dip
of strata, and faulting conditions, the pervious zone which requires seepage
control is relatively shallow and lends itself readily to control by grouting.
A comprehensive coverage of drilling methods, as well as grouting methods, is
presented in EM 1110-2-3506.
b. Grouting operations in foundations and abutments are not limited to

construction of grout curtains. Grouting may be used for foundation repair and
filling cavities or voids in limestone or carbonate formations. The require-
ment for a single- or multiple-line grout curtain is dictated by the condition
and integrity of the foundation as determined from preconstruction exploration.
c. Grouting of steep abutment slopes has the potential for causing dif-
ficulties and possibly can do more harm than good. Care must be exercised when
grouting in abutments to avoid displacements within the rock mass. Even rela-
tive low grouting pressures can cause joint opening and decrease the integrity
of the abutment.
d. In general the efficiency of a grouting operation for controlling

seepage in well graded sediments is proportional to the width of the curtain.
In rock foundations and abutments the seepage control accomplished by grouting
is not a function of width, however, but is dictated by the effectiveness of
sealing seepage paths and open joints identified in the exploration program.
The effectiveness of a grouting operation may be evaluated by pre- and post-
grouting pressure injection tests for evaluating the water take and the foun-
dation or abutment permeability.
11-6. Surface Treatment of Foundations and Abutments.
a. Surface treatment of foundations and abutments is essential to ensure

intimate contact of backfill materials with sound rock. Once a foundation is
exposed by excavation, the method of treatment and potential protection against
piping of embankment or abutment materials will be dictated by the conditions
encountered. All cavities, or caves, should be cleaned and plugged with con-
crete, both upstream and downstream from the core trench. All openings, fis-
sures, joints, etc., should be cleaned after excavation and treated with dental
concrete, where possible. Treatment with dental concrete stops major deterio-
ration of materials that weather rapidly during construction, and helps to
prevent lateral piping of the embankment material into the adjacent foundation.
b. If a well-developed, highly solutioned joint system, or similar con-

dition is encountered in the abutment, thick concrete walls may be placed
against the abutment. Prior to placement of concrete walls the abutment face
must be set back and cleaned, with presplitting being occasionally required.
11-3
EM 1110-2-1901
30 Sep 86
Concrete walls offer the advantage of either filling or blocking cavities,

affording a reasonable condition for treatment by grouting, and provide an
abutment "tie-in" for fill placement with ideal conditions for maximum compac-
tion of the adjacent embankment.
11-4
EM 1110-2-1901
30 Sep 86
CHAPTER 12
REMEDIAL SEEPAGE CONTROL
12-1. General Considerations. This chapter assumes that a seepage problem

with an existing structure has been identified and defined by methods dis-
cussed in Chapter 13, or by other observations. The next step is to decide on
a remedy, design and install the remedial measure, and monitor its performance
to determine if the problem has been satisfactorily addressed. Several fac-
tors, including consequences of continued detrimental seepage, the geotechnical
environment (embankment, foundation, abutment), and economy, will determine the
type and degree of remedial seepage control. Some of the more critical conse-
quences include:
a. Breaching of the embankment or loss of support to structural members

due to piping.
b. Breaching of the embankment from slope instability induced by loss of

material and/or strength due to seepage.
c. Loss of significant amounts of reservoir water.
d. Maintenance problems or loss of useful areas due to seepage on the

downstream slope or areas downstream of the embankment.
12-2. Remedial Methods.
a. Factors Affecting Choice of Methods, Several methods of reducing

undesirable seepage are discussed in this chapter; most have been previously
addressed in Chapters 8-11, which described methods and appropriate settings
for each. The remedial designer, while possibly having more advanced technol-
ogy available than the original designer, must work with existing conditions.
The embankment and its foundation, abutments, and seepage control measures may
form a complicated structure through which seepage occurs. This can make pre-
cise detection and remedial control difficult or impossible. Remedial action
may range from continued or additional monitoring to rebuilding or abandonment
of the dam. Choice of remedial method(s) will depend on several factors, which
include:
(1) Geotechnical environment.
(2) Risk.
(3) Degree of correction required.
(4) cost.
b. Effects of Methods on Other Structure Elements. The remedial

designer must also consider the interplay of the remedial measures with other
dam elements. For example:
(1) Effect of excavation for drains, cutoff trenches, slurry trenches,

etc., on embankment stability.
12-1
EM 1110-2-1901
30 Sep 86
(2) Difficulty of tying remedial measure to existing seepage control

elements.
(3) Possibility of hydraulic fracturing when grouting.
c. Monitoring. In all cases, pre- and post-remedial monitoring of seep-

age is essential to determine the effectiveness of remedial action. Since
Chapters 8-11 describe control measures in detail, this chapter will just point
out primary considerations in the choice of remedial measures and give examples
of their use. These examples are provided for general guidance only, since
efficient use of remedial measures is very dependent upon geotechnical charac-
teristics of the particular site's as-built configuration, reservoir uses, and
pool history.
12-3. Storage Restriction. The most direct method to alleviate a seepage

problem is to lower the reservoir and restrict pool levels in order to stop or
reduce seepage and its effects. This is often done during problem identifica-
tion. If piezometer and seepage quantity measurement devices are in place at
this time, the effect of this remedy will be experimentally determined.
Considerations in storage reduction include:
a. Reduction of downstream inundation area and level should breaching

occur.
b. Effects of pool lowering on water supply, flood control, power gen-

eration, navigation, recreation, and environment.
Normally, lowering and restriction of the reservoir pool is not an acceptable

long-term solution, but this depends on restriction levels and purpose of the
reservoir. Care must be taken in lowering the reservoir since rapid drawdown
can lead to instability of the upstream slope. Of course, risk of upstream
slope failure would normally be a preferred alternative to breaching of the dam
and release of a full reservoir.
12-4. Grouting. Grouting is a common, long-used remedy for seepage. Its

effectiveness is dependent upon being able to rather specifically locate the
leaking area and fill the culprit openings without damage to the embankment.
Possible damage includes cracking of impermeable cores or other impermeable
areas of the embankment, foundation, or abutments, and clogging of drains. If
grouting results in sealing of the foundation just downstream of or beneath the
downstream portion of the dam, uplift pressures may increase beneath the
embankment or seepage may be forced up into the downstream portion of the
embankment. Pore pressure instrumentation should be in place to monitor such
changes before grouting begins. This must be considered in design of remedial
controls. Because of the many variables in grouting, it is highly desirable
to have an experienced contractor and field engineer. In many cases, post-
grout drilling may be warranted to determine if the grout has thoroughly pene-
trated the desired area. Information about grout properties and grouting is
given in Chapters 9 and 11. Several case histories follow which provide
general examples.
12-2
EM 1110-2-1901
30 Sep 86
a. 95-ft-High Earthfill Dam (Ley 1974). Upon initial filling, an earth-

fill dam with a foundation and abutments of volcanic tuffs and breccias exhibi-
ted leakage at one of the downstream embankment-abutment contacts and out onto
the downstream slope. Inspection revealed leakage from open fractures in the
volcanic rock and drains were installed in the areas of seepage. The seepage
was stable for a number of years. Subsequent evaluation of the embankment for
seismic safety resulted in a need to reduce foundation seepage quantities and
piezometric levels within the embankment. A grouting program employed a low
viscosity chemical grout in order to penetrate any permeable layers in the
embankment where seepage might be occurring. Grout holes were split-spaced for
120 ft along the dam crest from the left abutment. If significant circulation
water was lost during drilling, the hole was grouted. Initial spacing was
12 ft with 14 of 23 holes taking low pressure grout (0-5 psi at the collar of
the hole). This low pressure was to prevent embankment heave. Most of the
take was well into the left abutment with holes spaced as close as 2-1/2 ft and
being deepened in stages and further grouted. Gel time varied from 2 to
18 minutes and final depth of holes varied from 24 to 84 ft. Total take was
3,200 gal with seepage being reduced 90 percent, but with little reduction in
piezometric levels. Grouting can reduce seepage quantities significantly but
still not alleviate high piezometric pressures, particularly in tight or fine-
grained materials since any continuous void or pore space can transmit upstream
heads.
b. 140-ft-High Earth Dam (Ley 1974). In the left abutment, gypsum had
apparently formed in the bedding planes and fractures of folded and faulted
shale and siltstone. After water was impounded, leakage, carrying dissolved
gypsum, occurred from the abutment. Settlement and gradual increase in seep-
age also indicated that gypsum was being removed from the formation. Built in
1915, the dam underwent a grouting program from 1930-1933, resulting in place-
ment of about 35,000 cu ft of grout in a series of holes along the dam crest,
the left abutment, and at the bottom of the hill forming the left abutment.
This program reduced seepage quantities by 75 percent. Approximately 30 years
later, over 32,000 cu ft of cement-bentonite grout (colored with iron oxide to
distinguish from previously placed grout) was placed in 137 holes to again
reduce seepage and replace material removed by solution. Cores indicated good
penetration with most seams from hairline to 1/8 in. thick. Seepage was
greatly reduced. Other geologic materials may also be dissolved when subjected
to seepage. In a similar manner, silt and clay in limestone cavities may also
be removed by seepage. Grouting may only be a temporary solution to a seepage
problem if solution of a soluble foundation continues after grouting.
c. 70-ft-High Earthfill Dam (Ley 1974). Seepage of 130-140 gal/minute

was discovered downstream and attributed to foundation leakage. Installation
of drains downstream of the dam allowed collection, metering, and return of
water to the reservoir. Drilling for a grouting program, undertaken some years
later to reduce seepage losses, revealed loose, sugarlike, decomposed granite
30-40 ft below the dam foundation. The grouting was unsuccessful in reducing
seepage. Subsequently, the bottom and right side of the reservoir were
covered with an impervious blanket of 40 tons of bentonite mixed with native
material. After mixing of the bentonite with native soil to a depth of 3 in.,
the surface was rolled with a rubber-tired roller. Surface drainage provisions
prevented runoff from eroding the blanket during partial pool. Seepage, after
12-3
EM 1110-2-1901
30 Sep 86
blanketing, decreased 50 percent. In this case, attacking the seepage problem

further upstream (at the reservoir) proved more efficient than trying to seal
an underlying seepage path.
d. Fontenelle Dam (Gebhart 1974). A 165-ft zoned embankment, Fontenelle

Dam, almost failed when a leak of up to 20 cu ft/second developed at the down-
stream contact with the right abutment. Much of the embankment was eroded
before drawdown was effective in stabilizing the embankment. Fortunately, out-
let capacity allowed lowering the reservoir 3-4 ft per day. The source of
leakage was not specifically determined, but an extensive grouting of founda-
tion rock (calcareous sandstone, siltstone, and carbonaceous shale) was suc-
cessful in preventing a recurrence of the problem. A 90- by 140-ft cement
grout blanket was placed upstream from the original grout cap. Grout curtains
were extended beneath the dam beyond the abutments. Over 200,000 cu ft of
grout was used.
e. Hills Creek Dam (Jenkins and Bankofier 1972). Hills Creek Dam, con-
structed by the Portland District, has a maximum height of 338 ft and consists
of a central impervious core with gravel and rock shells. Minor seepage occur-
red near the left abutment during first filling, but decreased with time.
Seepage markedly increased in extent and volume after 6 years of normal opera-
tion. Vertical drains placed in the downstream shell as an initial remedial
measure were not effective in lowering water levels in the downstream shell and
seepage continued to increase. An investigation to determine the seepage
source continued during the remedial action. Initially it was thought that
leakage was through the upstream blanket into the foundation and abutment, but
further observations indicated flow was through the core or core-foundation
contact. Grouting , which injected 4,500 sacks of cement, most in a 1-1/2:1 mix
at zero psi, resulted in elimination of almost all seepage. Four 42-in. bucket
auger holes, as well as several smaller borings, were drilled to inspect
grouting of the core and foundation. The main source of seepage was at a point
along the core foundation contact where a haul road had crossed the abutment.
Twelve years later, seepage is still negligible. Frequently, the source of
seepage is not obvious. The engineer must consider all possibilities and,
after choosing and installing a remedial measure, try to understand what post-
remedial monitoring is indicating. The extent of the engineer's knowledge of
the foundation, embankment materials, and construction history will greatly
influence the accuracy of his analysis of the seepage problem. Often available
foundation and construction information will not be adequate and further
geotechnical investigation will be required.
12-5. Upstream Impervious Blanket.
a. If it is determined that sealing of the reservoir bottom and sides

immediately upstream of the embankment will be useful in reducing undesirable
seepage quantities and pressures beneath the embankment, an upstream impervious
blanket may be employed. If successful and economically feasible, this is one
of the most efficient measures since the source of water, the reservoir, is
controlled upstream of the embankment and its foundation. This generally
requires removal of reservoir water, though some small reservoirs have been
sealed by placement of materials through water. Sources of fine-grained mate-
rial and, in some cases, filter materials are required. The impervious
12-4
EM 1110-2-1901
30 Sep 86
materials are usually placed on the reservoir bottom. If sloped areas such as
the reservoir sides of upstream embankment slope are to be sealed, considera-
tion must be given to protection against wave attack and erosion from runoff.
Additionally, fine-grained materials placed on the upstream embankment slope
may be removed during drawdown because of low saturated strength and high
saturated weight. If seepage can also go through the upstream portion of the
embankment and then into the foundation an upstream blanket will be less effec-
tive and another remedy may be necessary, e.g., cutoff beneath dam, fig-
ure 12-1. The nature of reservoir bottom materials must be considered. Any
large voids must be filled with a stable material such as compacted soil,
Figure 12-1. Possible problem if existing and remedial seepage

control measures are not properly coordinated (prepared by WES)
stabilized soil, concrete etc. High gradients will likely exist through the
blanket during high reservoir levels, particularly close to the embankment. It
may be necessary to place a filter material before placing the blanket to pre-
vent piping of the blanket material into the foundation. The extent of the
blanket is determined by analysis and will depend on several factors, including
extent of desired decrease in seepage quantities and pressures and blanket
material available (quantity and permeability) (EM 1110-2-1913 and Barron
1977). Man-made liners have provided a seal for reservoirs with pervious
foundations when fine-grained materials were not economically available. They
are usually rather expensive, require relatively smooth surface for placement,
and coverings (normally soil) to protect them from puncture in stressed areas
and deteriorating exposure to sunlight. Joining of sections is one of the most
critical and difficult aspects of man-made liners. Field seams, especially
under difficult field conditions and with other than highly experienced person-
nel, can be an appreciable source of leakage. Quality control of seaming
should be strict. One example of the use of an impervious upstream blanket was
given in paragraph 12-3c; another is provided below:
b. An impervious upstream blanket connected to a sloping impervious core

was placed during the construction of Tarbela Dam on the Indus River in
Pakistan (Lowe 1978). The blanket material consisted of sandy silt mixed with
a sandy silt angular boulder gravel. The blanket lay over an alluvium of
cobble gravel choked with fine sand. The blanket, which was to increase the
length of seepage path and not necessarily to reduce seepage quantities, met
12-5
EM 1110-2-1901
30 Sep 86
the piping criteria, D15 (alluvium) < 5D85 (blanket), Appendix D. As the
reservoir emptied after first filling, several sinkholes and cracks were noted
in the blanket. Sinkholes ranged from 1 to 15 ft in diameter and 4 to 6 ft in
depth. It was felt that uneven settlement during the first reservoir filling
caused tension and compression cracks in the blanket which allowed considerable
seepage into the underlying sand-choked gravel. In areas where the sand was
less dense, the seepage moved the sand down to form a layer in the lower part
of the gravel. This created open work gravel just beneath the blanket, and
fines from the blanket moved into and through this open layer forming the sink-
holes. Sinkholes were filled with filter material and mounded over with
blanket material. Typically, the blanket mounds were approximately 15 ft high
and extended 30 to 35 ft beyond the sinkhole edge. After filling of the reser-
voir, sinkholes were located by side-scan sonar and filled with a mixture of
filter material and silt from self-propelled bottom dump barges. Each sinkhole
generally received 50 barge loads of material. Sinkholes continued to be dis-
covered and covered over another 3-4 years after the initial remedial action.
Siltation on the reservoir blanket and filling of sinkholes have reduced
seepage about one half.
12-6. Downstream Berm. Berms control seepage by increasing the weight of the
top stratum so that the weight of the berm plus top stratum is sufficient to
resist uplift pressure. If of low permeability, they will reduce seepage, but
increase uplift pressures beneath the downstream toe of the dam since they
force seepage to exit further downstream of the dam. If pervious, they must
be designed as a filter or with an underlying filter to prevent upward migra-
tion of line particles from the foundation materials beneath them. Again, a
seepage analysis must be made to determine the resisting load required of the
berm. Downstream slope stability of the embankment will normally increase
because of the resistance to sliding provided by the berm. Huntington District
has employed berms as remedial measures at several flood control dams in the
Muskingum River flood control system (Coffman and Franks 1982). Similarity of
the embankments and environments allowed a standard remedial action for several
of the dams at the downstream embankment toe. A 3- to 7-ft-thick pervious
blanket of appropriate length is placed over the soft seepage areas at the
downstream toe. This adds weight and provides a working platform for instal-
lation of relief wells at points of excessive seepage.. Another example of a
stability berm is given in the Addicks and Barker Dams example,
paragraph 12-7a.
12-7. Slurry Trench Cutoff. Two major technical considerations in the use of
slurry trenches as remedial seepage control measures are (a) the effect on
stability of the embankment due to excavation of the trench and the presence
of a vertical plane of relatively weak soil (in the case of a soil-bentonite
backfill) and (b) tying the slurry trench to other existing or proposed seep-
age control measures. If a competent upstream blanket exists, the trench may
be placed upstream of the embankment and tied to the blanket or may be placed
through the dam and any pervious substratum if stability requirements are met.
A cement-bentonite backfill may be placed in panels or a concrete wall may be
placed in separately excavated elements if an open trench and the relatively
weak soil-bentonite backfill are unacceptable because of stability risks. The
following experiences with slurry trenches provide general examples of this
cutoff type as a remedial measure.
12-6
EM 1110-2-1901
30 Sep 86
a. Addicks and Barker Dams, Houston, Tex. (U. S. Army Engineer District,
Galveston 1977a; U. S. Army Engineer District, Galveston 1977b; U. S. Army
Engineer District, Galveston 1983). Completed in the late 1940's, Addicks and
Barker Dams are rolled earth embankments providing flood control in the Hous-
ton, Texas, area, with respective maximum heights of 48.5 and 36.5 ft above
streambed. Neither normally impound water except in periods of rainfall. The
embankments contain some silts and sands, foundations have silt and sand
layers, and upstream borrow areas expose the foundation permeable layers. At
the time of construction, these conditions were not considered, significant
because of the large discharge capability and short detention time of the
reservoirs. Residential and commercial development of the downstream local
area caused several changes in operating conditions which increased detention
time and made the effect of seepage more critical. These included restriction
of discharge rates and construction of drainage channels on non-Federal land
within 200-300 ft downstream of the center line of the dams which expose the
pervious portion of the foundation. Erosion of the drainage channel slopes on
the side of the channel nearest the dam and boils in the channel bottom during
times of low reservoir impoundment indicated the potential for dangerous seep-
age conditions during high reservoir levels. Downstream piezometers also
indicated a quick response to changes in reservoir levels. This example
describes remedial actions at Addicks Dam; actions at Barker Dam were similar.
Several remedial measures were considered:
(1) Downstream drainage blanket and stability berm - rejected due to

requirement for additional right-of-way and Government responsibility for
maintenance of local interest's drainage ditch.
(2) Downstream drainage blanket, stability berm, and relief well system
and downstream slurry trench - (relief wells between embankment toe and slurry
trench) very positive control (blanket and berm control embankment seepage
while wells and slurry trench control underseepage), but very costly, long-
term well maintenance required, and all seepage forces would be directed at
the embankment toe.
(3) Same plan as (2) except slurry trench replaced with steel sheet pile
cutoff - same reasoning as (2) except sheet pile would greatly increase cost.
(4) Slurry trench cutoff through embankment and foundation - a very

positive, controlled cutoff for embankment and foundation; no maintenance; all
work on Government property; less costly and quicker than other alternatives.
For most of the remedial work, alternative (4) was chosen, though for selected
lengths of the embankment where they were the best alternative, alterna-
tive (1) was used and some relief wells were placed. With a maximum depth of
64 ft and width of 3-5 ft, the slurry trench penetrated 2-4 ft into a rela-
tively impervious clay underlying the pervious foundation materials. Fig-
ure 12-2 provides a general cross section of the design. The trench was
placed 10-20 ft upstream of the embankment center line with equipment working
from a platform established by degrading the upper portion of the embankment.
Cemented materials, present in some portions of the excavation, were broken by
dropping a 10-ton percussion tool on the cemented layers. Portions of the
trench collapsed but were successfully reexcavated. Additionally, small (3-in.
diameter) tunnels were encountered in the upstream side of the trench but were
12-7
EM 1110-2-1901
30 Sep 86
!i
SLURRY
PORTION OF EMBANKMENT THAT MAY BE REMOVED TRENCH EXISTING EMBANKMENT
DURING CONSTRUCTION OF SLURRY TRENCH AND
REPLACED AFTER COMPLETION OF SLURRY TRENCH
UPSTREAM DOWNSTREAM
LIMIT OF DEGRADING ROWl
---
_;.r;:::-'
3 - ...._ ....__2_, 1 EXISTING GROUND
1
...._ ...t. 4 - L--

- ...._ / E L 99 4 _/EXISTING BERM CONSTRUCTED FROM ADJACENT
- --
EL 94.3--........
--~,_,
~EXISTING GROUND
_..- _,.- _.-
---f./to>---- 3' MINIMUM TO 5' MAXIMUM SLURRY

TRENCH CUTOFF
1
4.5
DITCH EXCAVATION
L.:::..::..,_\: --:: ::::.._-=--= =-...----

.
ORIGINAL EMBANKMENT SLOPE
t-.. . .......
BOTTOM OF TRENCH CUTOFF VARIES
TYPICAL SECTION
STA 350+50 TO STA. 377+00
NO SCALE
12-8
Ci SLURRY TRENCH
TOP OF CLAY CAP
" LIMIT OF DEGRADING

CLAY OR SANDY CLAY, (CH OR CL)
-------------r'-orrr~~--------~~-
SLURRY TRENCH BACKFILL
PROTECTIVE CLAY CAP DETAIL

NO SCALE
Figure 12-2. Addicks Dams remedial slurry trench for embankment and foundation seepage
83
control (from U.S. Army Engineer District, Galveston )
EM 1110-2-1901
30 Sep 86
plugged with cloth. Backfill gradation is shown in table 12-1. For some por-
tions of the project, percent passing for the No. 200 sieve were 15-30 per-
cent. Backfill mixing and transport to the trench were conducted in several
ways. Some backfill was batched dry, placed in concrete trucks with slurry
added, then mixed and transported to the trench. The higher fines content
backfill in some cases proved too sticky to mix in trucks. Mixing was con-
ducted on the ground next to the trench but occasionally excess fines were
picked up from the working surface. A concrete mixing pad was used as an
alternative though wear from the mixing equipment destroyed the concrete.
Excess or unsatisfactory material was deposited in old borrow areas upstream of
the embankment to reduce underseepage. In one area, a slurry trench located at
the upstream toe of the embankment provided underseepage control while a down-
stream berm provided embankment stabilization. The berm of sandy clay had
permeability characteristics similar to the embankment and provided a 1V on 8H
slope. Several of the discharge conduits which suffered from seepage and pip-
ing were resealed, after cleaning, with ethafoam backer rods and a polyurethane
sealant. Where the sealant would not adhere to the concrete, joints were
talked with oakum soaked with a grouting compound. Well screens were placed in
weep holes to prevent loss of soil, and relief wells with submersible pumps
were installed. For certain portions of Barker Dam, use of an upstream clay
blanket and a downstream stability berm (1V on 8H) was more cost effective than
a slurry trench. There was intermittent surface exposure of pervious founda-
tion materials and a source of CH materials for the blanket was available
within the reservoir. Prior to placement of the blanket, ponded water and soft
surface materials were removed. The blanket was placed in 8-in. layers and
compacted with tamping rollers at natural moisture content.
Table 12-1. Backfill Mix for Slurry Trench, Addicks Dam (a)
Sieve Size or Number Percent Passing

(U. S. Standard) by Weight
3 in. 100
1-1/2 in. 95 to 100
3/4 in. 80 to 100
No. 4 55 to 100
No. 10 40 to 80
No. 40 18 to 45
No. 200 10 to 25
(a) From U.S. Army Engineer District, Galveston. 85
Though not yet severely tested, the control measures have performed satisfac-
torily based on the following observations:
12-9
EM 1110-2-1901
30 Sep 86
(a) Foundation downstream piezometers do not respond to reservoir levels

experienced so far.
(b) Phreatic surface has been raised upstream of the slurry trench.
(c) No embankment seepage, but there have been no significant pools.
(d) Settlement plates indicate no significant settlement of the slurry

trench. Though the restored embankment has cracked in the area of the
trenches, inadequate compaction of the embankment fill is considered the
cause.
b. Wolf Creek Dam, Ky. (Fetzer 1979). Constructed in the 1940's, Wolf
Creek Dam is a 200-ft-high combination earthfill and concrete dam founded on
limestone containing shale and solution cavities. During excavation of a
10-ft-wide cutoff trench, several interconnected solution cavities were dis-
covered in the limestone. These were backfilled for a short distance with
impervious material, and a 50-ft-deep single-line grout curtain was placed
beneath the bottom of the cutoff trench. In 1967, muddy flow was observed in
the tailrace, a small sinkhole developed near the downstream toe, and wet areas
existed near the downstream toe. In 1968, a larger sinkhole developed (13 ft
wide, 10 ft deep) and drilling revealed solution features running perpendicular
and parallel to the dam axis. It was concluded that reservoir water was
passing beneath the cutoff trench. Grout lines were placed along the dam axis
near the embankment-concrete contact and downstream of this area. During
1971-1972, an overall assessment of the seepage problem was made since the
remedial grouting had only addressed about 200 ft of the 4,000-ft embankment
portion of the dam. A diaphragm concrete cutoff wall was considered the best
solution because it could be installed without draining the reservoir, a very
costly operation due to reservoir use. Explorations , which included borings
spaced on 3.1-ft centers along the axis of the wall (parallel to the dam axis),
defined the depth and length of the wall. Depth was 10 ft below the lowest
indication of solution activity (maximum depth 278 ft) and length was 2,239 ft.
In 1974, a request for technical proposals resulted in seven proposals with two
acceptable. In the second stage, a bid invitation was issued and an award was
made for a wall in the area of the switchyard and 989 ft of the wall along the
dam axis. The award in 1975 was followed by a second competition and an award
in 1977 for the remaining 1,250 ft of the axis wall. The wall consists of
alternate cylindrical primary elements and connecting secondary elements
installed using bentonite slurry, figure 9-14. Primary elements are
2.17-ft-diam steel casings filled with tremied concrete (see table 9-8 for mix
proportions). Weak cement grout fills the volume between excavation walls and
the casing. A 25-ft-deep core hole was drilled beyond the bottom of each pri-
mary element to explore for cavities and was pressure-tested and grouted prior
to the placement of a closed-end primary casing. The primary element was
required to set for a minimum of 20 days before excavation of the secondary
element which is also filled with tremied concrete. Frequent piezometer
readings (as often as every 4 hours) were made during construction to determine
the hydraulic condition of the embankment and foundation and warn of any
potentially critical seepage conditions. Excavation and drilling were closely
monitored to observe any drill rod drops or mud losses. Sealers and reserve
mud, constantly on hand, provided for emergencies. Grout takes around
12-10
EM 1110-2-1901
30 Sep 86
the primary casings and volume of concrete used in the secondary elements were
closely monitored as was the embankment in general. Efficient management of a
large number of observations was necessary to determine the current condition
of the dam. The lack of major losses of slurry, grout, or concrete during
construction was probably due to the densely spaced borings and grouting done
during the earlier exploration program. Wall construction, completed in 1979,
took approximately 4 years and two construction contracts. Subsequent
piezometric levels indicate the wall is a successful seepage barrier.
c. Camanche Dike 2, California (Anton and Dayton 1972). One of several

earthfill dikes containing Camanche Reservoir, Dike 2, is a zoned earth embank-
ment about 70 ft high founded on alluvium containing an upper 20-ft strata of
clayey sand underlain by layered silty-to-fine uniform sand stratum. The
underlying sand stratum varies in permeability with its lower portion contain-
ing gravel. Original construction involved extending the core horizontally to
the upstream toe and discing and compacting the top of the alluvium to 1,000 ft
upstream of the dike axis. This was expected to provide acceptable underseep-
age conditions , while providing the option of tying an upstream cutoff through
the alluvium to the core if operating underseepage conditions were intolerable.
After reservoir filling, underseepage proved extensive and flowed over down-
stream property. Lowering of the reservoir reduced the seepage and revealed
holes in the compacted alluvium upstream of the embankment. Several seepage
control methods, including upstream impervious blanket, grout curtain, relief
wells, downstream drains, sheet piles, and others, were considered. Evaluation
of the options resulted in choosing an upstream slurry trench. This method
provided a positive cutoff, minimized piping potential, allowed retention of a
partial reservoir, and was the least expensive of positive cutoff methods.
Placed 50 ft upstream of the upstream berm toe, the 1,660-ft-long slurry
trench, 8 ft wide, extended through the alluvial materials to a maximum depth
of 95 ft. Backfill specifications required a 4-in. slump and a gradation as
shown in Table 12-2. An 8- to 11-ft-deep sandy clay blanket protected by a
1-ft-thick cobble and gravel cover connected the slurry trench to the horizon-
tal core extension. Excess slurry was blended into the top portion of the
blanket to decrease blanket permeability. Since slurry trench placement,
downstream piezometers reflect decreased influence of reservoir levels with
only very small seepage flows at high, prolonged reservoir levels. Two poten-
tial sources of seepage are the somewhat pervious bedrock formation which the
slurry trench is keyed into and the sandy clay blanket connection between the
slurry trench and extended core. Connection of the 8- to 11-ft-deep blanket to
the slurry trench after placement of the slurry was difficult and may allow
reservoir leakage into the alluvium. Placement of the blanket or a partial
thickness prior to slurry trench construction was recommended. This would
provide a platform for construction of the slurry trench and allow a more
secure attachment of the trench to the blanket. This procedure has been
standard practice on many subsequent slurry trench projects.
12-8. Relief Wells. Though Chapter 9 describes design and installation of

relief wells, additional factors must be considered when relief wells are used
for remedial seepage control. Relief wells can relieve excessive uplift and
potential piping when pervious layers are overlain by relatively impervious
strata by providing controlled release of relatively large volumes of water.
Relief wells, as compared with cutoffs, allow loss of reservoir water and
12-11
EM 1110-2-1901
30 Sep 86
Table 12-2. Backfill Mix for Slurry Trench, Camanche Dike 2 (a)
Sieve Size or Number Percent Passing

(U. S. Standard) by Weight
6 in. 100
3 in. 80 to 100
3/4 in. 60 to 100
No. 4 40 to 80
No. 30 20 to 60
No. 200 10 to 30
(a) Courtesy of American Society of Civil Engineers. 135
require proper handling of discharge flows and periodic maintenance. Flooding

and erosion from well discharges must be prevented. Wells may be installed
quickly with a minimum of downstream right-of-way and, in many cases, without
reducing reservoir levels. If high uplift is present, boring and installation
may be difficult requiring extra measures to keep the hole open and stable
until the screen and filter are installed.
12-9. Drainage of Downstream Slope. Seepage emerging on or at the toe of the

downstream slope will normally be controlled by one of the methods previously
mentioned, Expedient installation of filter materials and a toe drain can
help prevent piping of embankment and foundation materials and may increase
embankment stability, but will not normally reduce seepage quantities. If
seepage is confined to a small area or areas, horizontally drilled drains may
help control the problem (Royster 1977). Horizontal drains of slotted pipe
normally do not have a filter envelope and would generally be used for
"nuisance" seepage or as an expedient measure until a more permanent solution
could be installed.
12-12
EM 1110-2-1901
30 Sep 86
CHAPTER 13
MONITORING PERFORMANCE OF SEEPAGE CONTROL MEASURES
13-1. General Considerations.
a. Before seepage control measures are implemented, site characteriza-

tion by thorough exploration and testing is needed to determine if the seepage
control measures will serve the intended purpose. Knowledge of the in situ
site conditions along with the purpose of the seepage control measure and its
physical dimensions will help determine' the overall number and the placement
of monitoring devices. Several factors which can be monitored that can lead
to a conclusion regarding the safety of a dam are: (1) progressive increase
in the volume of seepage flow, (2) removal of solids by the seepage,
(3) increased uplift pressures or locally depressed gradients, and (4) soft or
wet areas on the downstream embankment.
b. Monitoring the performance of seepage control measures can lead to a

collective conclusion drawn from several measurements. The most common and
easiest monitoring is to rely on visual observations along with careful sur-
face inspections at predetermined intervals. Another type of monitoring which
should be completed before construction is the installation of piezometers,
observation wells, and drainage collection systems to determine a site depen-
dent pattern of behavior. Finally, the actual structure should be monitored
by the installation of a site specific network of piezometers, observation
wells, and drainage collection systems with flow measurements designed for the
anticipated seepage problems. A regular review of the data collected will
generally detect major changes between subsequent readings but equally as
important are the long-range trends manifested by steady changes or intermit-
tent surges.
c. If it is determined during the monitoring process that a possible

problem exists, an expanded instrumentation program may be needed. This could
include more piezometers, relief wells, etc., and a more extensive analysis of
the seepage water; i.e., both a physical and chemical analysis of the sediment
and water including temperature, salt content, and resistivity which could be
compared with samples from the embankment and possible seepage sources.
According to the complexity of the problem and/or the economics versus safety
involved a group of other studies could be added including, but not limited to:
resistivity and spontaneous potential of the embankment and foundation, dye
tracing, infrared (aerial or portable ground based), and seepage acoustic
emissions. In most cases, the scope of the monitoring program will be deter-
mined by the economics involved.
13-2. Piezometers for Seepage Pressures.
a. Foundation. To determine the performance of seepage control measures,

a pattern of behavior should be established prior to and during construction
where long-term trends can be related to design or seepage conditions. Piezom-
eters should generally be installed in all compressible foundation soils, the
number being dependent on the extent and thickness of the strata. If possible,
foundation piezometers should be installed-in the sections selected for embank-
ment piezometers and should extend beyond the upstream and downstream toes a
13-1
EM 1110-2-1901
30 Sep 86
distance equal to the expected migration of pore pressures. An effective

piezometer installation plan should convert preconstruction piezometers into
postconstruction seepage monitoring piezometers for an effective continuity of
data. Shown in figure 13-1 is an example of a piezometer installation with the
embankment resting on a compressibility foundation (Chapter 9) and an imper-
vious core cutoff through a sand and gravel layer. Effectiveness of the cutoff
and the presence of uplift pressures can be evaluated. Shown in figure 13-2 is
an embankment with an impervious core that intercepts a pervious soil and rests
on the top of rock (Chapter 11) and again the effectiveness of the cutoff
and/or the integrity of the rock can be evaluated. Shown in figure 13-3 is an
embankment founded on a thin impervious top stratum underlain by a deep zone of
pervious material (Chapter 9). In this case a cutoff was impractical and seep-
age pressures are simply monitored beneath the dam to determine time effects on
the control measures while the effectiveness of relief wells downstream is
determined with foundation piezometers. An effective monitoring system should
also include piezometers in the abutments to determine the effectiveness of
drains and/or of the embankment-abutment interface which could include grout
curtain cutoffs (Chapters 10 and 11). Also artesian flows may exist in the
abutments and need to be monitored. Shown in figure 13-4 is a piezometer
installation that is used to monitor a grout curtain cutoff and cut-slope
drains. Remedial seepage control measures, discussed in Chapter 12, might
require additional piezometers to monitor both the installation and the effect
of the new measures.
b. Embankment. As discussed in Chapter 8 three methods for seepage

control in embankments are: (1) flat slopes with or without drains,
(2) embankment zonation, and (3) vertical (or inclined) and horizontal drains.
An embankment with flat slopes (as defined in Chapter 8) constructed of imper-
vious material, and which has infrequent high reservoir levels, should have
only enough piezometers in the embankment to establish the phreatic surface.
A typical example is shown in figure 13-5. To monitor seepage control
measures in a zoned embankment (Chapter 8) the number and spacing of piezom-
eters depend not only on the height of the dam but also on the material prop-
erties of the zones. The core must be monitored to determine the phreatic
Figure 13-1. Example of foundation piezometer installation of Surry Mountain

Dam (from EM 1110-2-1908)
13-2
NOTE: ALTHOUGH PIEZOMETER
LEADS ARE SHOWN
LEGEND EXTENDING THROUGH
IMPERVIOUS ZONE, THIS
o EMBANKMENT PIEZOMETER IS NO LONGER PERMITTED.
• FOUNDATION PIEZOMETER
i PROJECT CONTAINS OTHER
TYPES OF INSTRUMENTATION
THAT ARE NOT SHOWN.
13-3
PERVIOUS FOUNDATION
SOIL·S
SCALE IN FEET
40 0 80
Figure 13-2. Piezometer installation for dams on pervious foundations, Deer Creek Dam (from
EM 1110-2-1908)
EM 1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
~ NOTE: ALL PIEZOMETERS ARt OPEN-SYSTEM
TYPE. PROJECT CONTAINS OTHER TYPES
0 U.AI!IANK'-IENT PIEZOMETER OF INSTRU'-IENTATION THAT ARE NOT
e FOUNDATION PIEZOMETER SHOWN.
f.-AXIS Otr DAM

RCLAriVELI' I
PERVIOUS MArERIAL
RCLArtVELY PERVIOUS
UNCOMPACrCD MArERIAL
13-4
roc DRAIN
e SAND AND GRAVELS

•
TERTIARY CLAYS AND GRAVELS
SCALE IN fEET
40 00
Figure 13-3. Piezometer installation with a thin impervious topstratum and relief well
systems, Enid Dam (from EM 1110-2-1908)
! SEE SHEETS 249 AND 251
FOil LOCATION)
;
8 & ~ I~ •
8
0
8
'
A-111
I
13-5
A·ZIID I
I
SA· S
LEGEND
D STRONG MOTION ACCELEROGR&PH

SHELTER
5 SURFACE MONUMENT
liD A TYPE PIEZOMETER (FOUNDATION)
TRIGGER AND POWER CABLE ALIGNMENT
FOR STRONG MOT ION AGGELEROGRAPHS
PLAN &
(SEE SHEET 253)
SURFACE MONUMENT , PIEZOMETER AND
STRONG MOTION ACCELEROGRAPH SHELTER LOCATlON~U-IZ
100 1) 100 zoo
EM 1110-2-1901
30 Sep 86
Figure 13-4. Piezometer installation for dam abutments having grout
curtains or cut-slope drains, Applegate Lake Dam (from U. S. Army
104
Engineer District, Portland )
EM 1110-2-1901
30 Sep 86
D\STANCl IN f'EET I="ROM Ct.NTI:._'<.LI~'E or OMvl
800 700 600 500 400 300 200 \00 0 \00 200 300 ~00 sor; 6C:i ?OO 800
,-.-----T - --~ --- T --- T T- r---·r-------r-- y-----T - ---~-- -r --r- - --,-~
SELECT IMPERVIOUS
FILL----
o?o
18 "RIPPAP ON 6 "8£0/)/A/r. -
S60
24 ~RIPRAP (}N 4 ,.8!00/NG
650
J €>40
"'
~ 630
ti- b20
w
....
;;e 610
~ GOO
13-6
j: 590
~
~ '580
w
570
se.o
550
3' HO/'!Z0/1/r/lt.
ORfil•'-46£ R!.•'lhfXT
[OUFrTOf' SYSTEM_/
FLOODPLAIN C:MBANKM£NT SECTION

STATI0"-1 IOCHC>") TO STATION 1~2+25
Figure 13-5. Estimated seepage for Aubrey Dam, Texas (from U. S. Army Engineer
81
District, Fort Worth )
EM 1110-2-1901
30 Sep 86
surface and in situ permeability. Using this information, cracking potential

can be estimated. On either side of the core there can be filters, transition
zones, random zones, outer shells, and blankets (Chapter 8). In general, the
different zones should increase in permeability outward and should have
piezometers in each zone. During rapid reservoir drawdown, the upstream
piezometers would detect excess pore pressure buildup while during conserva-
tion and high pools, all the piezometers would indicate the effectiveness of
the zones and check design values. A typical installation in a zoned embank-
ment is shown in figure 13-3. In a zoned embankment the piezometers located
near the upstream core face can be instrumental in determining the development
of cracking in the core (Vaughan et al. 1970). Depression of the piezometric
elevation in a localized area (cone of depression) can indicate velocity head
loss which would indicate leakage through the core. Remedial work for
Baldershead Dam (Vaughn et al 1970 and Lovenburg 1974) included placement of a
large number of piezometers near the upstream face of the core (figure 13-6)
which were successful in indicating piezometric depressions.
c. Drains. The purpose of vertical (or inclined) and horizontal drains

is to control seepage either through the embankment or beneath the dam (under-
seepage). A vertical (inclined or horizontal) drain in the embankment may be
used as a filter to prevent material from eroding from the core and/or as a
method of collecting seepage exiting from horizontally stratified soil layers.
Enough piezometers should be installed in the drain to determine if the seep-
age is coming through the embankment material or if it is underseepage, fig-
ure 13-7. If the horizontal drains intercept underseepage which in turn is
drained by lateral drains, piezometers should be placed on either side of the
laterals to determine their effectiveness, figure 13-8. Long-term trends
(pressure buildup or depression) detected in the drains not directly related to
the reservoir level could indicate either clogged drains (pressure buildup) due
to embankment or foundation material moving into the drains or piping and
erosion (pressure depression) due to material moving into pipe drains, high
permeability zones, or into fractured rock. Toe drains are effective in col-
lecting seepage and preventing saturated areas along the downstream toe.
Piezometers in or near toe drains would only be effective when a downstream
blanket has been added and uplift pressures need to be measured. Drainage
galleries and tunnels are used mostly in abutments in the United States to
intercept and control seepage in fractured rock. Drainage tunnels along
extensions of the dam's axis or in downstream abutment areas serve to collect
seepage. Piezometers located near the drainage tunnels would indicate the
effectiveness of the tunnel. Cut-slope drains can be used to intercept seep-
age and collect drainage along abutment slopes while piezometers placed both
upstream and downstream of the drains can determine effectiveness of the col-
lection system, figure 13-9.
d. Downstream Areas. Seepage can migrate beyond the embankment toe

particularly in clay shale or fissured formations. Geologic site characteri-
zation in most cases will determine the need for piezometers downstream of the
toe but if there is any doubt they should be installed in the questionable
formations 50 to 150 ft beyond the toe, figure 13-1. Piezometers should also
be installed downstream of the outlet works, spillway, and stilling basin if
they extend well beyond the toe and are not close to other piezometers.
13-7
EM 1110-2-1901
30 Sep 86
GROUT TREATMENT
DIAPHRAGM WALL I I
Ol 0 100 200 300 400 500 609 BOO 900m
~~----_.~~~~~-r~T-~-r~-r~~-r~~-r~~~~~~~~-r~-r-r-r-+--------~~
330 330
320 320
310 310
300 300
290 290
280 mO.D. m00. 280
340-
ICEY
610 620 630 640 650 660 670
13-8
330- 335 335

t Multi-unit
pie20111eter 330 330
& 320-
325 325
Perforat~d
I1 standpipe 320 320
315 315
~ <:,asogrgnde-type
poeZCMMter
310 310
Nowat.r-
290- WATER TESTS I I
I I I I I I I
water loss
280-
ADDITIONAL PIEZOMETERS INSTALLED IN 1~7Q
SECTION AT CH. 372
Figure 13-6. Layout of Baldershead Dam piezometer installation in upstream shale fill
212
(courtesy of British Geotechnical Society )
TYPICAL SECTION- FLOOD PLAIN
13-9
LEGEND SCALE OF FEET
• SUFFACE MONUMENT
bea:. Y •lo 100
I
+ POPE PRESSURE DEVICE IN SILT FOUNDATION
SUSQUEHANNA RIVER BASIN
..t. EMBANKMENT PORE PRESSURE DEVICE
TIOGA· H~MWOND LAKES
(!} PIEZOMETER IN FOUNDATION ROCK ABUTMENT INSTRUMENTATION

IS PIEZOMETER IN NATURAL GRAVEL AQUIFER Tl OGA DAM
TYPICAL SECTIONS
• PIEZOMETER IN BLANKET DRAIN
IAlTI .. Oitl Olllti!CT
Q PIEZOMETER IN FOUNDATION TILL (AT T.R.)
PLATE 19-5
Figure 13-7. Piezometer installation for dams with through seepage and underseepage (from
79
U. S. Army Engineer District, Baltimore )
EM
1110-2-1901
30 Sep 86
EM 1110-2-1901
30 Sep 86
OOWNSTREAM
HP 04lPAOJI
20' MAINTENANCE BERN
rRANSITION MATERIAL --~

j_ ~~LLLED ROCK""'
TYPICAL SECTION- VICINITY STA. 56+ 50
00
..,
SCALE OF FEET
IDO
13-10
LEGEND
• SURFACE MONUMENT
+ PORE PRESSUPE DEVICE IN SILT FOUNDATION SUSQUlttAirfNA ftiYP IASIN

TIOGA· KAWIIIOND LAII:[S
.t. EM :JA "KM[NT PORE PRESSURE DE VIC[
INSTRUMENTATION
0 PI[ZOME TER IN FOUNDATION ROCK ABUTMENT
HAMMOND DAM
S PIEZOMETER IN NATURAL GRAVEL AQUIFER TYPICAL SECTIONS
• P1EZOMET£R IN BLANKEr DRAIN
<> P1EZOM£TER IN I OlltWAIION TILL (AT T.RJ
Figure 13-8. Piezometer installation for dams with lateral drains (from
79
u. S. Army Engineer District, Baltimore )
•-••cr-t:•n
A TT"' I'IU:OifUIIIII'OUifDoUICNII
- TIUIU:II •110 IIOWIII c•kl •t.IIIIIIIIIT PLAN J:h
13-11
/
SUftfAt:l[ MONUMUT , "UOM[T[ft AND
DOuel.[ WEill fOft

f'tf'ELINI[S 'A' AM) 'W
DRAINAGE SYSTEM PLAN
SECTION@
EM 1110-2-1901
Figure 13-9. Drain and piezometer installation, Applegate Lake, Oregon (from
Portland 104 )
30 Sep 86
·
U• S • Army Eng1neer D1str1ct,
· ·
EM 1110-2-1901
30 Sep 86
e. Near Relief Wells. If a relief well system has been installed, pore
pressures should be checked in the vicinity to evaluate the efficiency of the
system. Piezometers should be located both upstream and downstream of the
relief wells and should intercept the stratum being drained, figure 13-3. If
there is a line of wells, piezometers should also be installed generally at
the midpoint between the wells or at the point expected to have the highest
pore pressures.
f. Spillways, Stilling Basins, and Outlet Works. Underseepage control

beneath the stilling basins of spillways and outlet structures founded on per-
vious foundations is generally provided by drainage blankets supplemented in
the case of stratified foundations by deep well systems (figures 13-10 and
13-11). Drainage blankets extend beneath the chute slabs, if necessary. As
shown in figure 13-10, piexometers should be installed to check the effective-
ness of the drainage blanket and relief wells, and to check pore pressure
beneath the outlet channel and against the stilling basin walls. Piezometers
are used to check pore pressures occurring below the relief well system (fig-
ure 13-10). Generally, a sheet pile wall, with a minimum penetration of
15 ft, is installed along the downstream toe of the stilling basin to control
piping and a piezometer installed downstream of the wall is needed to deter-
mine the effectiveness of the wall.
13-3. Flow Measurements.
a. Weirs. Seepage flow measurement is an important parameter of dam

performance. Most installations have used a relatively simple weir, measuring
the seepage over brass or stainless steel 90-deg V-notches like the collection
system shown in figures 13-12 and 13-13. A number of weirs can be installed
in drainage galleries to determine flows from different sources, i.e., left or
right abutment, dam underseepage, and total seepage. A certain amount of
sediments will settle out just upstream of the weir which is important if the
sediment load is not continuous and occurs between visual inspections. If the
seepage is exiting downstream of the dam and outside of a drainage collection
system, a weir pond can be formed in conjunction with the V-notch weir for the
specific purpose of determining long-term sediment content in the seepage.
b. Flumes. A flume is a short rigid-walled channel designed to con-

strict the flow and so give rise to critical velocity. A single measurement
of water level is sufficient to measure discharge at critical velocity. The
most commonly used flume for seepage measurements is the Parshall flume as
shown in figure 13-14. Empirical charts for flow discharge have been devel-
oped for specified flume dimensions (Bureau of Reclamation 1967). The flume
can be fabricated and placed in seepage flow that has been channelized. This
method is a relatively rapid and simple way to obtain precise flow
measurements.
c. Relief Wells. Relief wells, as the name implies, are widely used to
relieve pressures and control seepage through pervious strata beneath earth
dams, spillways, and outlet works. A thorough knowledge of the geologic con-
ditions and characteristics of the soils at the dam must be available to
design a system as part of initial construction or as remedial work. To be
effective, the well must flow but must not allow the loss of foundation
13-12
... "!\"~·· ··~··· ...... ,
f
I
If
•''
<?-
+
•
r
13-13
..,.,...,,J __ •f_,., ..
llt•nuttol .. 1-1 p.rr- .-.#
OUTLl'T STII:UCTUillll[)
e
.......
, . . _ . , Nl:n•• "lo.cr . .._.._..
PLAN AND PROfiLE
• Af.ZOIIU."'tTR
•
"..,.._,""•"""'•n••'"u..t w IIMI.I.Y
·EO--~ I "' ,._,_~1--
a..I.T..O•
1" •
Cll'fiC.I. 01 TMI. ...........T, ••att»....l
1111.,1.111 C-IIIIIS-
:::.~c, w•aa ••~o
.~.-:-_!.~· ..,.
~--..
;.;::_. ---- ~--'-
C[Ntr• II" reorn (
EM 1110-2-1901
Figure 13-10. Piezometer installation for dams with outlet structures (adapted from
30 Sep 86
115
U. S. Army Engineer District, Vicksburg )
30 Sep 86
EM 1110-2-1901
S"'ILLWAT
COMPLETED PLAN
>fO I :IOOC:C.t K ....'( t•IO•JOfl
..... ~ I I < 1 • •t I ,. • .,.. H~>•
C"+Cf M Tfll l"tllil&ID(NT, '*'SSISS_.f'l ••v(tl (Qt,H,ooSltOO<
.,.,(oU~..... S · Nllll.t•c•
Su-ftlO ~••OvtO.
~.~~~-- ~"i;..:.-~-:.. -;:..:-

..... ,_............
....... -
,. ~·-·
13-14
Figure 13-11. Piezometer installation for spillways (adapted from U. S. Army Engineer
115
District, Vicksburg )
EM 1110-2-1901
30 Sep 86
SLOPE VARIES
MAX 5 • 0.0100_
OliN. 5 • 0.0020
REGULATING OUTLET
CHANNEL
EL. 1762.:1
PIPELINE A A
ADJUST SLOPE TO CONFORII \

WITH TOP OF ROCK
MIN SLOPE • 0 0020
,----~~=-=T~-=---~
~--~==~~--~ ~~==~~~~~~_oEL.1772.~
1
I TYPE m
r---••o'---1 - •I
RIQ Ct<AHNEL
~~ .....-..---_-:;_-_-_-_-__
PIPELINE C 1773.~
PROFILES ALONG ft. OF DRAINAGE SYSTEM

NOT TO SCALE
_)__,.,..._____.
~~~...._R2_~EYoSED _l:~oDOS
---- - - - -~--- -
~
ANO Sci)P£.!_
A~
--y
7111A1t01 : R£VrS[D PLAN PA()fll.£S
~--- -~---- ;.~~---
-
'J. S, £NQ!t-IE~ OISTRICT, PORTI.ANO
ROGUE ~IVEA BASIN OREGON

-x;,s APPLEGATE RIVER
- EMBANKMENT DAM
APPLEGATE LAKE
EMBANKMENT
-
~
.fwt!.
DRAINAGE SYSTEM
PLAN AND PROFILE
~~, Dol'~"-:
::::t~~~..H'-.- --------
Tl hA-..OL
4~tl~••&.\\···ot(l0
~~If-. .
-•n >o AGE -1-1/23
Figure 13-12. Installation of drainage system with weirs (from

104
U. S. Army Engineer District, Portland )
13-15
30 Sep 86
EM 1110-2-1901
FURNISH AND INSTALL I" MARINE PLY
_ \ CMP
SELECT BACKFILL
PLAN - DOUBLE WEIR SECTION @

13-16
SCAlE 1/ 2" • 1' 0"
WEI A BOARD 13 REQUIRED)

SCA&..l. ,. • ··-o•
U. S. ARMY ENGINEER DISTRICT. PORTLAND
ROGUE RIVER BASt"- OREGON
__ ,,,__
APPL£G4.T£ Al'w'ER
} -SELECT APPLEGATE LAKE
f1U-t----'~r'-----"!l_:j BACK FILL ,. EMBANKMENT DAM [W8ANKW[frljf
I 1/.l•.x 1/4••,.••• st- Hit• • _,,
~151/Z •.,_..UA ,..,. DRAINAGE SYSTEM
MINIMUM EXCAVATION LINE ~~! ., •tt:lt . . . . .- •t
DETAILS n
(1/JIIS'flll . ., ,_~-Ill U.l
lttltlltll,..,,..,._,."'''.,
.
~'"" .,.. ., "•"'-'••'
-- ~_s
SECTION @ SECTION E-E

SCALf 112•• 1'0' JIQ SCAll
Figure 13-13. Typical weir installation (from U. S. Army Engineer District, Portland 104 )
EM 1110-2-1901
30 Sep 86
Figure 13-14. Modified Parshall flumes of 6- to 65-second-foot

24
capacity (after Bureau of Reclamation )
material which could lead to piping or erosion failure. A frequent use of

relief wells is near the downstream toe of a dam to control seepage through a
pervious zone, figure 13-13. A toe drain is used to collect the flow while
weirs in the drains or manometers on the wells can determine the quantity of
flow. Relief wells can be pumped to determine the effect on surrounding
ground water and the permeability of the soil near the well. In many cases
where dams are built on jointed or fractured rock, grout curtains are used as
seepage cutoffs. It has been found in some cases that grout curtains do not
appreciably affect the uplift pressures downstream of the curtain and that a
series of drainage wells is a more effective tool to reduce uplift pressures
when the volume of seepage loss is not a problem and when the high cost of
grouting is hard to justify (Casagrande 1961). Relief wells have been suc-
cessfully used when the pervious strata is too deep and wide-ranging to
effectively use any type of positive cutoff. Relief (drainage) wells are used
beneath spillway and outlet works slabs to relieve excess pressure in the
rock, the underslab drains, or a pervious strata. An example of a group of
wells designed to relieve pressure under an outlet works stilling basin is
shown in figure 13-10. Relief wells can be used in rock abutments to inter-
cept seepage and to control artesian flow.
d. Seepage Outlets. Monitoring seepage outlets downstream of the dam

is handled according to the present severity of the problem or to future
associative problems if the seepage worsens. A small wet zone near the toe
might require only routine visual examination as would a small trickle from a
rock abutment. If there is sufficient seepage to measure, an effort should be
made to estimate the volume with a container and a stopwatch and to note sedi-
ment content. Operation and maintenance personnel should be trained to
13-17
EM 1110-2-1901
30 Sep 86
observe the situation especially during high reservoir levels. All seepage
outlets should be monitored to determine long-term trends. Sand boils devel-
oping downstream require immediate attention. An estimate of the pore pres-
sure involved can be determined by sandbagging around the boil to measure the
height of water rise. If the seepage exiting at the toe or abutment is severe
and will require remedial work but the dam is not in imminent danger of fail-
ing, collection systems must be designed as a pattern develops. Temporary
control measures such as toe drains and surcharge berms might be installed
with weirs to establish the severity and the trends of the seepage for use in
design of remedial work. Any changes noted during monitoring, i.e., volume
change, sediment load change, etc., should be considered important. After the
remedial work has been designed it may include any number of the monitoring
systems discussed previously in this chapter. If the seepage is exiting from a
drain system that is not being monitored internally, visual monitoring should
continue at specified intervals and during heavy runoffs and high reservoir
levels.
13-4. Seepage Water Analysis.
a. Physical Analysis. Physical analysis of the seepage could include

information on suspended solids, temperature contours, and water resistivity
values. This information may be obtained from physical testing of samples or
by remote testing techniques.
(1) The amount of suspended solids in the seepage is an indication of

material movement and piping. Although there are obvious problems with muddy
or turbid flow, seepage which appears clear may often carry small amounts of
suspended solids that would be detected by occasional samples and analysis.
Sediment traps built in conjunction with manholes and weirs can be used to
indicate the amount of suspended sediment and to obtain samples for chemical
analysis.
(2) Several different methods of measuring temperature are designed to

help locate seepage areas not yet visible and to trace seepage from its origin
to its exit. One remote sensing method (U.S. Army Engineer District, Los
Angeles 1981) is an aerial survey which includes any or all of the following:
(a) color photography, (b) color infrared photography, (c) thermal infrared,
and (d) color oblique imagery. The basis for the study is that different
materials (wet or saturated versus dry) possess different heat absorption
rates; therefore, heat radiation rates will differ. Since the specific heat
of water is higher than soil or rock, a warm zone in a known seepage area is a
suspected seepage outlet. This method is intended for large areas but smaller
areas can be covered by thermal infrared using portable hand-held units (Leach
1982). A second method of thermal monitoring (U. S. Army District, LOS
Angeles 1981) is to physically place an array of gages in and adjacent to the
embankment and measure the diurnal temperature (temperature below the reach of
the surface but within the annual temperature zone). A temperature fluctua-
tion is interpreted as seepage related and becomes the basis for further study
in that zone. A vertical thermal contour can be made in present open system
piezometers or in remedial planned piezometers, again with temperature fluctu-
ations or inversions interpreted as seepage related (U. S. Army Engineer Dis-
trict, Los Angeles 1981 and Leach 1982). These data are interpreted or
13-18
EM 1110-2-1901
30 Sep 86
expanded by determining the temperature stratification of the reservoir, the

temperature of all known seepage sources, and the temperature at the seepage
exits.
(3) Seepage can be monitored by the physical detection of tracer ele-

ments such as dye and isotopes that have been introduced upstream of the seep-
age exit either in the reservoir or a piezometer close to the suspected seep-
age path. The dye selected should have a favorable absorption and decay rate
and should meet state water quality control requirements. Samples taken from
piezometers, drains, and downstream exit points are monitored using an instru-
ment capable of measuring parts per billion, e.g., a fluorometer for fluores-
cein dye. By recording time of arrival and concentrations, interpretations
can be made as to the source of the seepage and the permeability of the strata
along the path of seepage. Environmental isotopes are also traced by obtain-
ing samples which can be measured by a mass spectrometer for oxygen -18 and
deuterium and by the low level counting system for tritium.
(4) Another physical property of the seepage that can be measured is

its conductance or resistance. Resistance which would be defined in the field
by a resistivity survey is a measure of the ability to resist current flow
through the seepage, a factor that is altered by the introduction of salt com-
pounds, graphite, etc. A thorough sampling program from all possible sources
of seepage, all seepage exits, and all available piezometers can produce a
group of resistivity values that is an important tool in defining seepage
sources and possible paths. Using the geologic profile for the site, and by
comparing individual resistivity values or by comparison against a range of
known values (Telford et al. 1976), an interpretation as to the source of the
seepage and the strata through which it travels can be made.
b. Chemical Analysis. To monitor and interpret the chemical composi-

tion of seepage requires a thorough knowledge of the surrounding geology or
chemical analysis of samples in the different strata. If possible, a pre-
construction chemical analysis should be conducted on all water sources and on
any formation that might contribute minerals or salts or that might affect
acidity or alkalinity.
(1) One important chemical property would be the salt concentrations in

the seepage. In this case to determine correlations, reservoir and ground-
water concentrations are essential along with the mineral content of the area,
e.g., water flowing through limestone would generally increase in chloride
concentration. Minerals such as feldpoid sodalite and apatite (Turkish
National Committee 1976) or caliche in volcanic regions (U. S. Army Engineer
District, Los Angeles 1981) can release chloride ions into the water. Inter-
pretation of chloride concentrations and its long-term trends can help deter-
mine the relative length of seepage paths (shallow or deep seated), the extent
of the leaching of the formation (whether concentrations are constant,
increasing, or decreasing), and the source of the seepage (comparable concen-
trations). Interpretation is sight-dependent and any of the above or possibly
other conclusions may be reached.
(2) Another chemical property needed for interpretation is the mineral

content of the seepage. Mineral concentrations of calcium and magnesium
13-19
EM 1110-2-1901
30 Sep 86
bicarbonate containing dissolved carbon monoxide increase with length of flow

through a limestone formation while flow through some clay minerals may
increase concentrations of calcium and magnesium. Again, interpretation is
site dependent and requires a thorough knowledge of the existing conditions.
(3) Stiff diagrams, as shown in figure 13-15, are a graphical method of

presenting the anions and cations that are dissolved in the water (Hem 1970).
The Stiff method uses ions plotted in the same sequence to give an irregular
polygonal shape or pattern. In tracing the movement of seepage water, the
Stiff patterns are plotted on a map of the site for various locations down-
stream of the dam and the reservoir. The Stiff patterns may yield information
regarding the path(s) of seepage from the reservoir, prior land uses down-
stream of the dam (feedlot, septic field, etc.), and type of formation
(gypsum, dolomite, etc.).
13-5. Remote Sensing Methods.
a. Resistivity and Spontaneous Potential. As part of a total seepage

study, resistivity and spontaneous potential methods have been used success-
fully for seepage delineation in soil and rock (Cooper and Bieganousky 1978;
Cooper, Koester, and Tranklin 1982; and Koester et al. 1984). Resistivity
surveys used as part of a seepage study help identify possible zones of high
moisture as a function of depth and location. After the surveys are correlated
to previous borings or geologic information, new borings are placed in the
seepage flows. Spontaneous potential surveys are used to detect negative D. C.
voltage anomalies in the surface electrical field which have been found to
indicate zones of seepage flow (Cooper, Koester, and Tranklin 1982; and
Koester et al. 1984). Although flow is indicated, depth to flow can not be
determined for a given anomaly.
b. Photography. Methods using color, color infrared, aerial and ground

base thermal infrared, and color oblique photography were discussed previously
in paragraph 13-4.a.
c. Refraction Seismic Surveys. Seismic surveys can be used indirectly

in a seepage study by providing a bedrock profile that can be used as an aid
in determining the location and depths of observation wells, piezometers, and
relief wells. It is a quick and inexpensive method for obtaining subsurface
profiles.
d. Seepage Acoustic Vibrations. Acoustic emissions are the noises

generated whenever a material deforms or possibly by seepage whenever there is
turbulent flow against and around a casing (Koerner, Lord, and McCabe 1977).
The technique has been used to detect seepage by placing an accelerometer on a
waveguide that extends to the bottom of a borehole and recording the vibrations
present (Koerner, Lord, and McCabe 1977; and Leach 1982). Increases in emis-
sions activity are interpreted as seepage flow. A similar technique that
consists of lowering an acoustic microphone down into a reservoir has been
used to detect leakage on the upstream asphalt-covered face and in the reser-
voir itself (Coxon and Crook 1976). One disadvantage would be high background
noise levels.
13-20
EM 1110-2-1901
30 Sep 86
--r-
Ne+K~----t-----~~~~~~+-----+---~a
·C.~----t-----+-----+-----+-----+---~HCOa
Ma~--~~--~-----+-----+----~----~so.
F•~----t-----+-----+-----+-----+---~co,
12-6
15-l
30 u 20 15 10 5 0 10 25
CATIONS. IN MILUEQUIVAU:NTS ANIONS, IN MILLIEQUIVALENT$
PER UTER PER LITER
Figure 13-15. Stiff diagram used to graphically present

anions and cations in seepage water (from Hem 1970)
13-21
EM 1110-2-1901
30 Sep 86
CHAPTER 14
INSPECTION, MAINTENANCE, AND REHABILITATION OF
SEEPAGE CONTROL MEASURES
14-1. Introduction. Proper functioning of seepage control features requires

adequate maintenance, inspection, and, if necessary, rehabilitation. Some
seepage control methods such as relief wells are in the best condition they
will ever be the day they are installed and developed, while others such as
soil-bentonite slurry trench cutoffs may increase in effectiveness with time.
All seepage control features must function effectively for the life of the dam
or be rehabilitated or, if necessary, replaced. The effectiveness of some
seepage control methods, such as the toe trench drain, may be directly
observed. For other seepage control methods, such as cutoffs, performance
monitoring (see Chapter 13) is essentially the only means of determining the
degree of effectiveness.
14-2. Inspection. The procedure for periodic inspection and continuing evalu-
ation of dams is given in ER 1110-2-100. Procedures for reporting evidence of
distress in dams are given in ER 1110-2-101. Procedures in these two ER's are
often supplemented by Division Regulations as well. Details concerning the
monitoring performance of seepage control measures are given in Chapter 13.
The first general field inspection for new earth and rock-fill dams is carried
out immediately after topping out the embankment. The initial inspection of
concrete dams is accomplished prior to impoundment of reservoir water. The
second inspection for earth and rock-fill and concrete dams is made at a
reasonable stage of normal operating pool but no later than one year after
initial impoundment has begun. Subsequent inspections will be made at one-year
intervals for the next four years, at two-year intervals for the following
four years, and then may be extended to every five years if warranted. The
periodic inspections provide the opportunity for a group of specialists to
critically examine a project for existing and/or potential problems, to recom-
mend remedial action or changes in instrumentation , and to direct the attention
of the operating personnel toward the significant and critical features of a
project. However, the occasional inspection cannot take the place of daily
observations required to detect potentially dangerous problems at an early and
repairable stage. Table 14-1 outlines the inspection, instrumentation, mainte-
nance, and rehabilitation of seepage control facilities. Some seepage control
methods such as embankment zonation, cutoffs, and upstream impervious blankets
are not amenable to visual inspection. Other methods such as flat slopes
downstream of the dam and downstream seepage berms are most accessible and
should be inspected daily during periods of full reservoir pool to ascertain
that they are functioning properly in controlling seepage. Other seepage con-
trol facilities should be inspected on a regular schedule as shown in
table 14-1.
14-3. Maintenance. Timely performance of maintenance on seepage control

facilities is required for the facilities to perform satisfactorily. Some
seepage control methods such as flat slopes downstream of the dam and down-
stream seepage berms are accessible and require maintenance. Other seepage
control methods such as embankment zonation, cutoffs, and upstream impervious
blankets do not require maintenance. However, maintenance is required on the
instrumentation used to evaluate the degree of effectiveness of all seepage
14-1
EM 1110-2-1901
30 Sep 86
Seep••• Control P•c.UitY !JM of Haint•aance Method of R•h•bilitation
Control of S••e••• 'lbroup Eabankaent
Plat elope• Vithout Dally Wet •pot•. •louch ... Monthly Surf•c:e IIODUMI'I.U Pertilid.nl, .ow- btond dop41 thtckDan
drain• tna, •roaton ina, flll1na and lonath
Yari.. (b) (d)
PIIIOM:ter
Variu(b) (d)
Vertical and horizontal' Weekly Turbidity and dia- PUia.eter
~~train• charge rate
Control of Under•••p•••
(d)
Hortaontal drain Wukly Turbidity and dio- V.eriel(b) PiiiOMter
chat·a• rate
Cutoff
__ (d)
c..poeud baekt111
treneh
Slurry trench
Varie1 (b) (d)

SoU-be-ntonite Ptezoaeter lnatall •hart parallel
adjlcent tr•nch
(d)
Ce•ent-bentonite lnetall ehot:t parallel
adj1c1nt tr1nch
Vorho (b) (d)
Concrete w•ll PieiOHUt.'
Upetr••• iaparvioua Yarho(c) R•••rvoir Fillina by bars• duapinl

blanket eediMntation or follovins r•••rvoir
drevdovn
Downuna• aeepage bera Dolly See pas•, boil•, Monthly Fertilizin& • .av- Extend bn·• thlc:ka.ell
erosion tna, fUHna and lenath
Toe trench dr.ain w.. kly Tnbldity ond dh- Chock partodically<•l
c:hAr&• rau
bltef velle Vat hi (b) Flow nte and sand Ph.aoeeter Chock par1od1cally(f) RohabUitate ... u.
infUtrat1on
Con.crete galletietl Wtokly Turbidity and dh- Check parlodicolly <al

ch•rs• rau
Control of. Seep••• Throuah Abut•ent
Upstreu iapervtoua Var:lee (b) Reeervoir Filling by barge duapina

blanket eedillentation or followina reetrvoir
dravdovn
Weekly Turbidity end d11-

c:harl• ratl
Relief wlla Varh1 (b) Plow rate and sand Chock periodtcolly(f) Rehob!l1tate vella
tnfiltrat1on
Control of Stepaae Beneath Spillways and StUUna 8a11n1

(d)
Dra1naa• blanket Weekly Tqrbidity and dh- Variea (b) Piezouter
chars• rate
Var'he (b) Flow rate and 1.and Varies (b) Piezo.e ur Chack. ptriodically(f) Rehabilitate vella en
lnt'iltraticm unwatertna of
structuTe
(a) Schedule to be folloved upon obtainaant of full reservoir pool. The schedule durina initial fillin& will ba deterained durin& th.e initial
periodic inapection.
(b) Obearvat tou should be aad~ one week after ...xiatl• reservoir level and at three aubsequent fall ina raa&rvoir ataaea.
(c)V.eries dttpencUn& upon potential aed:t. . ntation probleM, 1..,act on project perfo~nc:e,. and: t..pact on 1tre. . ayat. . (aae EM 11 t0-2-4000).
(d}Rialna or f'aUtn& bead test abould be conducted on pier.oaeteu annually before the atora.,e aeaaon (Me EK lil0-2-1908, Part l).
(e)C.tcb baelu, unholea, ditcbel, drainage pipe, and weir• ahculd be cle&ned, •• • ainiau. in thtl! fall tn preparation for the vlnter
••••on, and •aa1n in the apring.
(f} Annually. before atoraae aeaaon, c:h&ck valva a, aaak.ata, vall au.rda. cover pl1tea,. flap aataa em the outl•t• and other appurtanaa:c&s (see
EM liiQ-1-1908, Part 1).
(&)Should be axaatned fo-r atra•• cract.. hula••· •hUta of ali&ftiMnt. ••ce•atve lealc.a&•• &net dabrt• cleaned fr,._ &utter• and weirs.
14-2
EM 1110-2-1901
30 Sep 86
control methods. The type of maintenance to be conducted on seepage control

facilities is given in table 14-1.
14-4. Rehabilitation. Inspection and maintenance of seepage control facili-

ties may indicate the need for rehabilitation. The type of rehabilitation to
be conducted on various seepage control facilities is shown in table 14-1. If
the seepage control facility cannot be satisfactorily rehabilitated, remedial
seepage control facilities should be installed (see Chapter 12). The majority
of rehabilitation of seepage control facilities is in connection with relief
wells. Often a relief well does not flow except during high reservoir levels.
The water in the well becomes stagnant, various chemicals precipitate, algae
grows, and the efficiency of the well deteriorates. This will be manifested
by a fall in relief well discharge accompanied by a rise in piezometric
levels. Rehabilitation of a relief well is in essence a redevelopment of the
well. In addition to mechanical methods such as water jetting, surging, com-
pressed air, and pumping, certain chemicals, detergents, and water softeners
can be used in the rehabilitation process.. Chemical tests on water samples
from the wells will indicate if and what chemicals are applicable. An exami-
nation of the well screen with a borehole TV camera is recommended, turbidity
of the water permitting, to determine the degree of deterloration and/or
clogging of the screen. For well screens constructed of metal (no wooden
screens) that will not be damaged by acid, incrustation from calcium carbonate
that has cemented gravel filter particles can be removed by treatment with
hydrochloric acid. For well screens constructed of metal, treatment with
chlorine can remove bacterial growths of slimes. Iron oxides may be removed
by treatment with polyphosphate and surging the well. Calcium hypochlorite
can be used with polyphosphates to kill iron bacteria (Johnson Division,
Universal Oil Products Co. 1972).
14-3
EM 1110-2-1901
30 Sep 86
APPENDIX A
REFERENCES
Government Publications
1. TM 5-818-5.
2. TM 5-545.
3. ER 1110-2-100.
4. ER 1110-2-101.
5. EM 1110-2-1901.
6. EM 1110-2-1902.
7. EM 1110-2-1905.
8. EM 1110-2-1906.
9. EM 1110-2-1908.
10. EM 1110-2-1911.
11. EM 1110-2-1913.
12. EM 1110-2-2300.
13. EM 1110-2-2400.
14. EM 1110-2-2501.
15. EM 1110-2-3501.
16. EM 1110-2-3506.
17. EM 1110-2-4000.
18. Albritton, J., Jackson, L., and Bangert, R., "Foundation Grouting Prac-
tices at Corps of Engineers Dams," TR GL-84-13, 1984, U. S. Army Engineer
Waterways Experiment Station, P. O. Box 631, Vicksburg, MS 39180-0631.
19. Al-Hussaini, M., and Townsend, F. C., "Investigation of Tensile Strength

of Compacted Soils," Miscellaneous Paper S-74-10, Jun 1974, U. S. Army
Engineer Waterways Experiment Station, P. O. Box 631, Vicksburg, MS
39180-0631.
20. Banks, D. C., "Three-Dimensional Electrical Analogy Seepage Model

Studies; Appendix A: Flow to Circular Well Arrays Centered Inside a Cir-
cular Source, Series G," Technical Report No. 3-619, Mar 1963, U. S. Army
A-1
EM 1110-2-1901
30 Sep 86

39180-0631.
21. Banks, D. C. "Three-Dimensional Electrical Analogy Seepage Model Studies;

Appendix B: Flow to a Single Well Centered Inside a Circular Source,
Series H," Technical Report No. 3-619, Mar 1965, U. S. Army Engineer
22. Bedinger, M. S., "Relationship Between Median Grain Size and Permeability
in the Arkansas River Valley, Arkansas," Professional Paper 424-C, 1961,
U. S. Department of the Interior, Geological Survey, 1717 H Street N.W.,
Washington, DC 20277.
23. Bureau of Reclamation, Groundwater Manual, Denver, Colo., 1977. Avail-

able from: Technical Information Center, U. S. Army Engineer/Waterways
Experiment Station, P. O. Box 631, Vicksburg, MS 39180-0631.
24. Bureau of Reclamation, Water Measurement Manual, 2nd ed., 1967, U. S.

Department of the Interior, P. O. Box 25007, Denver, CO 80225.
25. Cedergren, H. R., "Flow Net Studies for Several Dams," for U. S. Army
Engineer District, Sacramento, Mar 1975, 650 Capitol Mall, Sacramento, CA
95814.
26. Clay, F. M., A Century on the Mississippi, A History of the Memphis

District, U. S. Army Corps of Engineers 1876-1876, Jan 1976, U. S. Army
Engineer District, Clifford Davis Federal Building, Memphis, TN 38103.
27. Clough, G. W., "An Evaluation of the Technical Feasibility of a Slurry

Trench Cutoff for the Excavation for the Tensas-Cocodrie Pumping Plant,
and a Review of Slurry Trench Specifications," 1978, report prepared for
U. S. Army Engineer District, P. O. Box 60, Vicksburg, MS 39180.
28. Cobb, J. E., "Seepage Analysis Report - Stage 3, Specifications, Calamus

Dam, Nebraska," Technical Memorandum No. V-222-B-lg, Jan 1984, Bureau of
Reclamation, P. O. Box 25007, Denver, CO 80225.
29. Cooley, R. L., Harsh, J. F., and Lewis, D. C., "Principles of Ground-
Water Hydrology," Hydrologic Engineering Methods for Water Resources
Development, Vol 10, Apr 1972, U. S. Army Engineer Hydrologic Engineering
Center, 609 Second Street, Davis, CA 95616.
30. Cooper, S. S., and Bieganousky, W. A., "Geophysical Survey of Cavernous

Areas, Patoka, Indiana," Miscellaneous Paper S-78-1, Jan 1978, U. S. Army
39180-0631.
31. Cooper, S. S., Koester, J. P., and Franklin, A. G., "Geophysical Investi-
gation at Gathwright Dam," Miscellaneous Paper GL-82-2, Mar 1982, U. S.
Army Engineer Waterways Experiment Station, P. O. Box 631, Vicksburg, MS
39180-0631.
A-2
EM 1110-2-1901
30 Sep 86
32. Desai, C. S., "Seepage in Mississippi River Banks, Analysis of Transient

Seepage Using a Viscous-Flow Model and Finite Difference and Finite Ele-
ment Methods," Technical Report S-73-S, Report 1, May 1973, U. S. Army
39180-0631.
33. Desai, C. S., "Seepage in Mississippi River Banks; Analysis of Transient

Seepage Using Viscous Flow Model and Numerical Methods," Miscellaneous
Paper S-70-3, Report 1, Feb 1970, U. S. Army Engineer Waterways Experi-
ment Station, P. O. Box 631, Vicksburg, MS 39180-0631.
34. Desai, C. S., and Kuppusamy, T., "Development of Phreatic Surfaces in

Earth Embankments," report prepared for the Bureau of Reclamation, May
1980, Department of Civil Engineering, Virginia Polytechnic Institute,
Blacksburg, Va. Available from: Technical Information Center, U. S.
39180-0631.
35. Duncan, J. M., "Seepage Analysis for Columbia Lock and Dam," Miscellane-
ous Paper No. 3-503, Jun 1962, U. S. Army Engineer Waterways Experiment
Station, P. O. Box 631, Vicksburg, MS 39180-0631.
36. Duncan, J. M., "Three-Dimensional Electrical Analogy Seepage Model

Studies," Technical Report No. 3-619, Mar 1963, U. S. Army Engineer
37. Edris, E. V., Jr., and Vanadit-Ellis, W., "Geotechnical Computer Program
Survey," Miscellaneous Paper GL-82-1, Mar 1982, U. S. Army Engineer
38. Environmental Protection Agency, "Manual of Methods for Chemical Analysis

of Water and Wastes," EPA-625-6-76-003a, Jul 1976, 26 W. St. Clair
Street, Cincinnati, OH 45268.
39. Franke, P., "The Concrete-Pile Cutoff Wall According to the I.C.O.S. -
Vedner Patent," Translation No. 54-6, Sep 1954, U. S. Army Engineer
40. Gerwick, B. C., Jr., Holland, T. C., and Komendant, G. J., "Tremie Con-
crete for Bridge Piers and Other Massive Underwater Placements," Report
No. FHWA-RD-81-153, Sep 1981, U.S. Department of Transportation, Federal
Highway Administration, Washington, D. C. Available from: Technical
Information Center, U. S. Army Engineer Waterways Experiment Station,
P. O. Box 631, Vicksburg, MS 39180-0631.
41. Guertin, J. D., and McTigue, W. H., "Groundwater Control Systems for
Urban Tunneling," Groundwater Control in Tunneling, Vol 1, Report
No. FHWA-RD-81-073, Apr 1982, U. S. Department of Transportation, Federal
Highway Administration, Washington, D. C. Available from: Technical
A-3
EM 1110-2-1901
30 Sep 86
42. Hall, R. L., Tracy, F. T., and Radhakrishnan, N. "Documentation of Two-

and Three-Dimensional Seepage Problems Using the Finite Element Method,"
Miscellaneous Paper No. K-75-6, Jun 1975, U. S. Army Engineer Waterways
43. Hallford, C. R., "Fluid Trench Construction, Concrete Diaphragm Wall Con-
struction," Lecture for Construction of Earth and Rockfill Dams Course,
Mar 1983, U. S. Army Engineer Waterways Experiment Station, P. O.
Box 631, Vicksburg, MS 39180-0631.
44. Hanson, J. E., "Impervious Cutoff by Means of Slurry Trench Construction

Mill Site Dam and Reservoir, Ferron Creek, Ferron Utah," 1972, U. S.
Department of Agriculture, Soil Conservation Service, P. O. Box 11350,
Salt Lake City, UT 84147.
45. Harza Engineering Company, "Engineering Investigations and Design Studies

for Underseepage Control (Slurry Trench Cutoff) for Saylorville Dam, Des
Moines River, Iowa," Nov 1965, report prepared for U. S. Army Engineer
District Clock Tower Building, Rock Island, IL 61201.
46. Hem, J. D., "Study and Interpretation of the Chemical Characteristics of

Natural Water," Water-Supply Paper No. 1473, 1970, U. S. Department of
Interior Geological Survey, 1717 H Street, Washington, DC 20277.
47. Holland, T. C., and Turner, J. R., "Construction of Tremie Concrete Cut-
off Wall, Wolf Creek Dam, Kentucky," Miscellaneous Paper SL-80-10, Sep
1980, U. S. Army Engineer Waterways Experiment Station, P. O. Box 631,
Vicksburg, MS 39180-0631.
48. Jansen, R. B., Dams and Public Safety, 1980, U. S. Department of the
Interior, Bureau of Reclamation, Denver Federal Center, Denver, CO
80225.
49. Kealy, C. D., and Busch, R. A., "Determining Seepage Characteristics of

Mill-Tailings Dams by the Finite Element Method," Report of Investiga-
tions 7477, Jan 1971, U. S. Bureau of Mines, Spokane Mining Research
Center, Spokane, Wash. Available from: Technical Information Center,
U. S. Army Engineer Waterways Experiment Station, P. O. Box 631,
50. Koester, J. P., et al, "Geophysical Investigations in Support of Clear-

water Dam Comprehensive Seepage Analysis," Miscellaneous Paper GL-84-3,
May 1984, U. S. Army Engineer Waterways Experiment Station, P. O.
51. Leach, R. E., "Waterbury Dam Seepage Study," draft report prepared for
U. S. Army Engineer District, New York, 1982, U. S. Army Engineer Water-
ways Experiment Station, P. O. Box 631, Vicksburg, MS 39180-0631.
A-4
EM 1110-2-1901
30 Sep 86
52. Lohman, S. W., et al., "Definition of Selected Ground-Water Terms -

Revisions and Conceptual Refines," Water-Supply Paper 1988, 1972, U. S.
Department of the Interior, Geological Survey, 1717 H Street, Washington,
DC 20277.
53. Mansur, C. I., and Perret, W. R., "Efficacy of Partial Cutoffs for Con-
trolling Underseepage Beneath Levees," Technical Memorandum No. 3-267,
Jan 1949, U. S. Army Engineer Waterways Experiment Station, P. O.
54. Mantei, C. L., and Cobb, J. E., "Seepage Analysis, Progress Report -
Stage 2, Specifications, Calamus Dam, Nebraska," Technical Memorandum
No. V-222-B-1-F, 1981, Bureau of Reclamation, P. O. Box 25007, Denver, CO
80225.
55. Mantei, C. L., Esmiol, E. E., and Cobb, J. E., "Seepage Analysis, Calamus
Dam - Stage I, Pick-Sloan Missouri River Basin Project, Nebraska,"
Technical Memorandum No. V-224-A, Mar 1980, Bureau of Reclamation, P. O.
Box 25007, Denver, CO 80225.
56. McAnear, C. L., and Trahan, C. C., "Three-Dimensional Seepage Model

Study, Oakley Dam, Sangamon River, Illinois," Miscellaneous Paper S-72-3,
Jan 1972, U. S. Army Engineer Waterways Experiment Station, P. O.
57. Miller, S. P., "Geotechnical Containment Alternatives for Industrial

Waste Basin F Rocky Mountain Arsenal Denver, Colorado; A Quantitative
Evaluation," Technical Report GL-79-23, Sep 1979, U. S. Army Engineer
58. Montgomery, R. L., "Investigation of Relief Wells, Mississippi River

Levees, Alton to Gale, Illinois," Miscellaneous Paper S-72-21, Jun 1972,
59. Moser, J. H., and Huibregtse, K. R., "Handbook for Sampling and Sample
Preservation of Water and Wastewater," EPA-600-4-76-049, Sep 1976,
Environmental Protection Agency, 26 W. St. Clair Street, Cincinnati, OH
45268.
60. Nettles, E. H., and Calhoun, C. C., "Drainage Characteristics of Base

Course Materials, Laboratory Investigation," Technical Report No. 3-786,
Jul 1967, U. S. Army Engineer Waterways Experiment Station, P. O.
61. Nobari, E. S., Lee, K. L., and Duncan, J. M., "Hydraulic Fracturing in
Zoned Earth and Rockfill Dams," Contract Report CR S-73-2, Jan 1973,
62. Pace, M. E., et al., "Seepage Analysis of Confined Flow Problems by the
Method of Fragments," Instruction Report K-84-8, Sep 1984, U. S. Army
A-5
EM 1110-2-1901
30 Sep 86

39180-0631.
63. Perry, E. B., "Laboratory Tests on Granular Filters for Embankment Dams,"
Draft Technical Report, 1986, U. S. Army Engineer Waterways Experiment
64. Perry, E. B., "Piping in Earth Dams Constructed of Dispersive Clay;

Literature Review and Design of Laboratory Tests," Technical Report
No. S-75-15, Nov 1975, U. S. Army Engineer Waterways Experiment Station,
65. Perry, E. B., "Susceptibility of Dispersive Clay at Grenada Dam,

Mississippi, to Piping and Rainfall Erosion," Technical Report
No. GL-79-14, Sep 1979, U. S. Army Engineer Waterways Experiment Station,
66. Radhakrishnan, N., "Flow Nets and Computer Solutions to Seepage

Problems," Seminar on Construction Dewatering and Pressure Relief,
Jan-Feb 1978, U. S. Army Engineer Waterways Experiment Station, P. O.
67. Sherard, J. L., "Filters for Glacial Tills and Other Coarse Impervious
Soils," Memorandum IG, May 1984, Soil Conservation Service Soil Mechanics
Laboratory, Lincoln, Nebraska. Available from: Technical Information
Center, U. S. Army Engineer Waterways Experiment Station, P. O. Box 631,
68. Soil Conservation Service, "Ground Water," National Engineering Handbook,

Section 18, Jun 1978, U. S. Department of Agriculture, P. O. Box 2890,
Washington, DC 20013.
69. Spooner, P. A., et al., "Technical Handbook on Slurry Trench Construction

for Pollution Migration Control," EPA draft, Oct 1982, U. S. Environ-
mental Protection Agency, 26 W. St. Clair Street, Cincinnati, OH 45268.
70. Stanley Consultants, Inc., and Woodward-Clyde Consultants, "Lake Chicot

Pumping Plant, Inspection Guidelines for Slurry Trench Construction," Sep
1976, prepared for U. S. Army Engineer District, P. O. Box 60, Vicksburg,
MS 39180-0631.
71. Stanley Consultants, Inc., and Woodward-Clyde Consultants, "Lake Chicot

Pumping Plant, Letter Report, Slurry Trench," Apr 1977, U. S. Army
Engineer District, P. O. Box 60, Vicksburg, MS 39180-0631.
72. Talbot, J. R., and Nelson, R. E., "The Mechanics of Seepage Analysis,"
Soil Mechanics Note No. 7, Oct 1979, U. S. Department of Agriculture,
Soil Conservation Service, P. O. Box 2890, Washington, DC 20013.
73. Tracy, F. T., "An Interactive Graphics Finite Element Method Grid Genera-
tor for Two-Dimensional Problems," Miscellaneous Paper K-77-5, Aug 1977a,
A-6
EM 1110-2-1901
30 Sep 86

74. Tracy, F. T., "An Interactive Graphics Postprocessor for Finite Element
Method Results," Miscellaneous Paper K-77-4, Aug 1977b, U. S. Army Engi-
neer Waterways Experiment Station, P. O. Box 631, Vicksburg, MS
39180-0631.
75. Tracy, F. T., "A Plane and Axisymmetric Finite Element Program for
Steady-State and Transient Seepage Problems," Miscellaneous Paper K-73-4,
May 1973a, U. S. Army Engineer Waterways Experiment Station, P. O.
76. Tracy, F. T., "A Three-Dimensional Finite Element Program for Steady-State
and Transient Seepage Problems," Miscellaneous Paper K-73-3, May 1973b,
77. U. S. Army Corps of Engineers, "Foundation Pumping Tests," Civil Works

Engineer Letter 63-16, 4 Dec 1963, Washington, D. C. Available from:
Technical Information Center, U. S. Army Engineer Waterways Experiment
78. U. S. Army Engineer District, Baltimore, "Susquehanna River Basin, Kettle

Creek, Pennsylvania, Kettle Creek Reservoir, Embankments," Design
Memorandum No. 7, Oct 1958, P. O. Box 1715, Baltimore, MD 21203.
79. U. S. Army Engineer District, Baltimore, "Susquehanna River Basin, Tioga-

Hammond Lakes, Tioga River and Crooked Creek, Pennsylvania, Instrumenta-
tion and Inspection," Design Memorandum No. 19, Jun 1973, P. O. Box 1715,
Baltimore, MD 21203.
80. U. S. Army Engineer District, Ft. Worth, "Brazos River Basin, Aquilla
Creek, Texas, Aquilla Lake, Embankment, Spillway, and Outlet Works,"
Design Memorandum No. 7, May 1976a, P. O. Box 17300, Ft. Worth, TX 76102.
81. U. S. Army Engineer District, Ft. Worth, "Trinity River Basin, Elm Fork,
Trinity River, Aubrey Lake, Texas, Embankment and Spillway (Revised),"
Design Memorandum No. 5, Jun 1976b, P. O. Box 17300, Ft. Worth, TX 76102.
82. U. S. Army Engineer District, Ft. Worth, "Trinity River Basin, Mountain
Creek, Lakeview Lake, Texas, Embankment and Spillway," Design Memorandum
No. 9, Apr 1980, P. O. Box 17300, Ft. Worth, TX 76102.
83. U. S. Army Engineer District, Galveston, "Buffalo Bayou and Tributaries,

Addicks and Barker Dams, Texas, Emergency Seepage Control, Construction
Completion Report," Jan 1983, P. O. Box 1229, Galveston, TX 77553.

Addicks Dam, Texas," Inspection Report No. 3, 1977a, P. O. Box 1229,
Galveston, TX 77553.
A-7
EM 1110-2-1901
30 Sep 86

Texas, Addicks Dam, Letter Report for Emergency Seepage Control," May
1977b, P. O. Box 1229, Galveston, TX 77553.

Texas, Major Rehabilitation, Addicks and Barker Dams," Reconnaissance
Report, 1977c, P. O. Box 1229, Galveston, TX 77553.
87. U. S. Army Engineer District, Huntington, "Kokosing River Basin, North

Branch of Kokosing Dam and Reservoir," Design Memorandum No. 5, Vol 1, Feb
1969, 502 Eighth Street, Huntington, WV 25701.
88. U. S. Army Engineer District, Huntington, "Muskingum River Watershed,

Bolivar Dam, Sandy Creek/Ohio," Embankment Reanalysis Report, Vol 1, Sep
1979a, 502 Eighth Street, WV 25701.
89. U. S. Army Engineer District, Huntington, "Muskingum River Watershed,

Mohawk Dam, Walhonding River, Ohio," Embankment Reanalysis Report, Vol 1,
Dec 1979b, 502 Eighth Street, Huntington, WV 25701.
90. U. S. Army Engineer District, Kansas City, "Hillsdale Lake, Osage River
Basin, Big Bull Creek, Kansas," Plans for Hillsdale Dam, Apr 1978,
601 E. 12th Street, Kansas, City, MO 64106.
91. U. S. Army Engineer District, Kansas City, "Long Branch Lake, Chariton
River Basin, East Fork Little Chariton River, Missouri," Jul 1974,
601 East 12th Street, Kansas City, MO 64106.
92. U. S. Army Engineer District, Little Rock, "Laboratory Technical

Memorandum for Construction and Use of Flow Nets," Sep 1940,
P. O. Box 867, Little Rock, AR 72203.
93. U. S. Army Engineer District, Los Angeles, "Painted Rock Reservoir, Gila
River, Arizona, Seepage Control Measures," Oct 1981, P. O. Box 2711, Los
Angeles, CA 90053.
94. U. S. Army Engineer District, Louisville, "Taylorsville Lake, Ohio River

Basin, Salt River, Kentucky, Dam and Spillway," Design Memorandum No. 6,
Jan 1974, P. O. Box 59, Louisville, KY 60201.
95. U. S. Army Engineer District, Memphis, "Foundation Report, Design, Con-

struction, and Performance of the Impervious Cutoff at W. G. Huxtable
Pumping Plant, Marianna, Arkansas," Apr 1978, Clifford Davis Federal
Building, Memphis, TN 38103.
96. U. S. Army Engineer District, Mobile, "Carters Dam, Coosawattee River,

Georgia, Embankment Criteria and Performance Report," Mar 1976,
P. O. Box 2288, Mobile, AL 36628.
97. U. S. Army Engineer District, Mobile, "Okatibbee Dam, Okatibbee Creek,

Mississippi, Main Dam and Spillway," Design Memorandum No. 8, Apr 1965,
P. O. Box 2288, Mobile AL 36628.
A-8
EM 1110-2-1901
30 Sep 86
98. U. S. Army Engineer District, Nashville, "Wolf Creek Dam, Cumberland

River, Kentucky, Concrete Diaphragm Walls, Plans and Specifications,
Phase 1," Apr 1975, P. O. Box 1070, Nashville, TN 37202.
99. U. S. Army Engineer District, Omaha, "Embankment Criteria and Perfor-

mance Report, Missouri River, Fort Peck Lake, Montana," Vol I, Sep 1982,
215 N. 17th Street, Omaha, NE 68102.
100. U. S. Army Engineer District, Omaha, "Missouri River, Oahe Reservoir,

South Dakota, Relief Wells," Design Memorandum No. MO-162, Nov 1961,
215 N. 17th Street, Omaha, NE 68102.
101. U. S. Army Engineer District, Omaha, "Report on Underseepage Observa-

tions, Garrison Dam," May 1964, 215 N. 17th Street, Omaha, NE 68102.
102. U. S. Army Engineer District, Philadelphia, "Lehigh River Basin, Jordan

Creek, Pennsylvania, Trexler Lake, Embankment, Spillway and Outlet
Works," Design Memorandum No. 7, Nov 1974, US Custom House 2nd and
Chestnut Streets, Philadelphia, PA 19106.
103. U. S. Army Engineer District, Pittsburgh, "Upstream Concrete Cut-Off

Wall, Allegheny Reservoir Dam," Cut-Off Wall Construction, Part II, Jul
1965, 1000 Liberty Avenue, Pittsburgh, PA 15222.
104. U. S. Army Engineer District, Portland, "Applegate Lake, Applegate

River, Oregon, Instrumentation, Inspection, and Evaluation," Design
Memorandum No. 15, Apr 1978, P. O. Box 2946, Portland, OR 97208.
105. U. S. Army Engineer District, Rock Island, "Local Flood Protection Proj-
ect, Rockford - Stage I, Kent Creek, Winnebago County, Illinois,"
Specifications for Page Park Dam, Jun 1978a, Clock Tower Building, Rock
Island, IL 61201.
106. U. S. Army Engineer District, Rock Island, "Slurry Trench Cutoff, Dam-
Foundation Report, Saylorville Lake, Des Moines River, Iowa," Part VI,
Binder 3 of 5, Jan 1978b, Clock Tower Building, Rock Island, IL 61201.
107. U. S. Army Engineer District, Sacramento, "Buchanan Dam, H. V. Eastman

Lake, Chowchilla River, California, Embankment Criteria and Performance
Report," Apr 1977, 650 Capitol Mall, Sacramento, CA 95814.
108. U. S. Army Engineer District, Savannah, "Construction of Slurry Trench

Cutoff," May 1968, P. O. Box 889, Savannah, GA 31402.
109. U. S. Army Engineer District, Savannah, "Specifications for Rehabilita-

tion of Clemson Lower Division Dam, Hartwell Lake, Georgia and South
Carolina," Oct 1981, P. O. Box 889, Savannah, GA 31402.
110. U. S. Army Engineer District, Savannah, "West Point Dam, Chattahoochee

River, Alabama and Georgia, Embankment Criteria and Performance Report,"
Oct 1979, P. O. Box 889, Savannah, GA 31402.
A-9
EM 1110-2-1901
30 Sep 86
111. U. S. Army Engineer District, St. Louis, "Clarence Cannon Dam and
Reservoir, Upper Mississippi River Basin, Salt River, Missouri, Embank-
ment Design, Main Dam," Design Memorandum No. 12, Nov 1969, 210 Tucker
Boulevard, St. Louis, MO 63101.
112. U. S. Army Engineer District, St. Louis, "John H. Overton Lock and Dam,
Red River Waterway," Design Memorandum No. 17, Detail Design, Vol 2,
Appendices, Design Computations, Nov 1978, St. Louis, MO. (prepared for
U. S. Army Engineer District, New Orleans).
113. U. S. Army Engineer District, Tulsa, "Big Hill Lake, Big Hill Creek,
Kansas, Embankment, Outlet Works, and Spillway," Design Memorandum
No. 6, Oct 1975, P. O. Box 61, Tulsa, OK 74121-0061.
114. U. S. Army Engineer District, Tulsa, "El Dorado Lake, Walnut River,
Kansas, Embankment and Spillway," Design Memorandum No. 6, Jul 1974,
P. O. Box 61, Tulsa, OK 74121-0061.
115. U. S. Army Engineer District, Vicksburg, "Operation and Maintenance

Manual, Grenada Dam, Yalobusha River, Mississippi," 1977, P. O. Box 60,
Vicksburg MS 39180-0631.
116. U. S. Army Engineer District, Walla Walla, "Mills Creek Lake,

Washington, Storage Dam Rehabilitation," Design Memorandum No. 6, Feb
1980a, Building 602, City-County Airport, Walla Walla, WA 99362.
117. U. S. Army Engineer District, Walla Walla, "Storage Dam Rehabilitation,

Mill Creek Flood Control Project, Walla Walla, Washington, Plans and
Specifications," Sep 1980b, Building 602, City-County Airport, Walla
Walla, WA 99362.
118. U. S. Army Engineer Division, Ohio River, "Conference on Control of

Underseepage," Apr 1945, P. O. Box 1159, Cincinnati, OH 45201.
119. U. S. Army Engineer Waterways Experiment Station, "Analysis of

Piezometer and Relief Well Data, Arkabutla Dam," Technical Report
No. 3-479, Jun 1958, P. O. Box 631, Vicksburg/MS 39180-0631.
120. U. S. Army Engineer Waterways Experiment Station, "Investigation of

Underseepage and Its Control, Lower Mississippi River Levees," 2 Vols,
Technical Memorandum No. 3-424, Oct 1956a, P. O. Box 631, Vicksburg, MS
39180-0631.
121. U. S. Army Engineer Waterways Experiment Station, "Review of Soils

Design, Construction, and Prototype Analysis, Blakely Mountain Dam,
Arkansas," Technical Report No. 3-439, Oct 1956b, P. O. Box 631,
122. U. S. Army Engineer Waterways Experiment Station, "The Unified Soil

Classification System," Technical Memorandum No. 3-357, Apr 1960, P. O.
A-10
EM 1110-2-1901
30 Sep 86
123 U. S. Department of Agriculture, Soil Conservation Service, "Soil

Mechanics Note No. 5, Flow Net Construction and Use," Oct 1973, P. O.
Box 2890, Washington, DC 20013.
124 Warriner, J. B., and Banks, D. C., "Numerical Analysis of Partially

Penetrating Random Well Arrays," Technical Report S-77-5, 1977, U. S.
39180-0631.
125 Weeks, E. P., "Field Determination of Vertical Permeability to Air in

the Unsaturated Zone," U.S.G.S. Professional Paper 1051, 1978, U. S.
Geological Survey, National Center, 12201 Sunrise Valley Drive, Reston,
VA 22092.
126. Winter, C. D., "Design of Slurry Trenches on the Tennessee-Tombigbee

Waterway," paper presented at the CE Seminar on Design and Construction
of Slurry Trenches on Tennessee-Tombigbee Waterway, Tupelo, Miss., Aug
1978. Available from: Technical Information Center, U. S. Army Engi-
neer Waterways Experiment Station, P. O. Box 631, Vicksburg, MS
39180-0631.
127. Zangar, C. N., "Theory and Problems of Water Percolation," Engineering

Monograph No. 8, Apr 1953, Bureau of Reclamation, Denver, Colo. Avail-
able from: Technical Information Center, U. S. Army Engineer Waterways
128. Zeigler, T. W., "Determination of Rock Mass Permeability," Technical

Report S-76-2, Jan 1976, U. S. Army Engineer Waterways Experiment
Non-Government Publications
129. Ambraseys, N. N., "Cut-Off Efficiency of Grout Curtains and Slurry

Trenches," Grouts and Drilling Muds in Engineering Practice,
Butterworths, London, 1963, pp 43-46. Available from: Technical
130. American Petroleum Institute, "Oil-Well Drilling Fluid Materials,"

Specification No. 13A, 8th ed., Mar 1981, Dallas, Tex. Available from:
American Petroleum Institute, 211 North Ervay, Suite 1700, Dallas, TX
75201.
131. American Petroleum Institute, "Standard Procedure for Testing Drilling

Fluids," Recommended Practice No. 13B, 9th ed., May 1982, Dallas, Tex.
Available from: American Petroleum Institute, 211 North Ervay,
Suite 1700, Dallas, TX 75201.
132. Anonymous, "Fast Value Engineering Saves Over Budget Dam," Engineering
News Record, Vol 201, No. 19, 9 Nov 1978, pp 24-25. Available from:
A-11
EM 1110-2-1901
30 Sep 86
133. Anonymous, "Ground Water: Our Priceless Resource," Ground Water Age,
Vol 14, No. 8, Apr 1980, pp 33-36, 64-66. Available from: Technical
134. Anonymous, "World's Deepest Cutoff Wall Reaches 430 Ft," Engineering
News Record, Vol 188, No. 1, 6 Jan 1972, pp 26-28. Available from:
135. Anton, W. F., and Dayton, D. J., "Comanche Dike-2 Slurry Trench Seepage
Cutoff," Proceedings-of the Conference on Performance of Earth and
Earth-Supported Structures, Purdue University, Vol 1, Part 1, 1972,
pp 735-749. Available from: Technical Information Center, U. S. Army
39180-0631.
136. Arulanandan, K., and Perry, E. B., "Erosion in Relation to Filter Design
Criteria for Earth Dams," Journal of the Geotechnical Engineering
Division, American Society of Civil Engineers, Vol 109, No. 5, May 1983,
39180-0631.
137. Barber, E. S., and Sawyer, C. L., "Highway Subdrainage," Proceedings,

Highway Research Board, 1952, pp 643-666. Available from: Technical
138. Barron, R. A., "The Design of Earth Dams," Handbook of Dam Engineering,
A. R. Golze, ed., Van Nostrand, New York, 1977, pp 291-318. Available
from: Technical-Information Center, U. S. Army Engineer Waterways
139. Basak, P., and Madhav, M. R., "Upper and Lower Limit of Darcy's Law,"
Indian Geotechnical Journal, Vol 9, No. 2, Apr 1979, pp 134-153.
Available from: Technical Information Center, U. S. Army Engineer
140. Bear, J., Dynamics of Fluids in Porous Media, Elsevier, New York, 1972.
141. Blind, H., "Inspection Galleries in Earth and Rockfill Dams," Water
Power and Dam Construction, Vol 34, No. 4, Apr 1982, pp 25-31. Avail-
142. Bloom, E., Dynes, S., and Glossett, C., "Soil Bentonite Slurry Trench in
a PL-566 Rolled Earth Dam in Indiana," Paper No. 79-2579 presented at
American Society of Agricultural Engineers Meeting, New Orleans, La.,
A-12
EM 1110-2-1901
30 Sep 86
1979. Available from: American Society of Agricultural Engineers, P. O.

Box 229, St. Joseph, MI 49085.
143. Boer, S. A. De, and Molen, W. H. Van Der, "Electrical Models: Conduc-
tive Sheet Analogies," Drainage Principles and Applications, Interna-
tional Institute for Land Reclamation and Improvement, Wageningen,
Netherlands, Publication 16, Vol 1, 1972, pp 201-221. Available from:
144. Bourdeaux, G. and Imaizumi, H., "Dispersive Clay at Sobradinho Dam,"

Dispersive Clays, Related Piping, and Erosion in Geotechnical Projects,
Special Technical Publication 623, May 1977, pp 13-24. American Society
for Testing and Materials, Philadelphia, Penn, Available from: Techni-
cal Information Center, U. S. Army Engineer Waterways Experiment
145. Brand, E. W., and Armstrong, R., "Sand Model Studies of Seepage Through
Earth Dams," Proceedings, Symposium on Earth and Rockfill Dams, Indiana,
Vol 1, 1968, pp 188-196. Available from: Technical Information Center,
146. Bros, B., and Orzeszyna, H., "The Influence of Particle Angularity and
Surface Roughness on Engineering Properties of Sand," Design Parameters
in Geotechnical Engineering, British Geotechnical Society, London, 1979,
39180-0631.
147. Burmister, D. M., "Principles of Permeability Testing of Soils,"

Symposium on Permeability of Soils, ASTM Special Technical Publication
No. 163, American Society for Testing and Materials, Apr 1955, pp 3-26.
148. Burmister, D. M., "The Importance and Practical Use of Relative Density
in Soil Mechanics," Proceedings, American Society for Testing and
Materials, Vol 48, 1948, pp 1249-1268. Available from: Technical
149. Cary, A. S., Walter, B. H., and Horstad, H. T., "Permeability of Mud
Mountain Dam Core Material," Transactions, American Society of Civil
Engineers, Vol 108, 1943, pp 719-737. Available from: Technical
150. Casagrande, A., "Control of Seepage Through Foundations and Abutments of

Dams," Geotechnique, Vol 11, No. 3, 1961, pp 161-182. Available from:
A-13
EM 1110-2-1901
30 Sep 86
151. Casagrande, A., "Seepage Through Dams," New England Waterworks Associa-
tion, Vol LI, No. 2, Jun 1937, Dedham, Mass. Available from: Technical
152. Case International Company, "Case Slurry Wall Notebook," 1982, Roselle,
Ill. Available from: Case International Company, P. O. Box 40,
Roselle, IL 60172.
153. Cedergren, H. R., "Drainage and Dewatering," Foundation Engineering

Handbook, H. F. Winterkorn and H.-Y. Fang, eds., Van Nostrand Reinhold,
New York, 1975, pp 221-243. Available from: Technical Information
154. Cedergren, H. R., "Seepage Control in Earth Dams," Embankment-Dam

Engineering, R. C. Hirschfield and S. J. Poulos, eds., Wiley, New York,
1973, pp 21-45. Available from: Technical Information Center, U. S.
39180-0631.
155. Cedergren, H. R., Seepage, Drainage and Flow Nets, 2nd ed., Wiley, New
York, 1977. Available from: Technical Information Center, U. S. Army
Engineer Waterways Experiment Station, P. O. Box 631, Vicksburg,
MS 39180-0631.
156. Christian, J. T., "Flow Nets by the Finite Element Method," Ground
Water, Vol 18, No. 2, Mar-Apr 1980a, pp 178-181. Available from:
157. Christian, J. T., "Flow Nets from Finite Element Data," International
Journal for Numerical and Analytical Methods in Geomechanics, Vol 4,
Apr-Jun 1980b, pp 191-196. Available from: Technical Information
158. Chugaev, R. R., "Seepage Through Dams," Advances in Hydroscience,

V. T. Chow, ed., Academic Press, New York, Vol 7, 1971, pp 283-325.
159. Coffman, J. A., Jr., and Franks, L. W., "Rehabilitating the Muskingum
River System," Transactions of the Fourteenth International Congress on
Dams, Q52 R.50, Rio de Janeiro, Brazil, 1982, pp 827-845. Available
from: Technical Information Center, U. S. Army Engineer Waterways
160. Couch, F. B., and Ressi deCervia, A. L., "Seepage Cutoff Wall Installed
Through Dam is Construction First," Civil Engineering, American Society
of Civil Engineers, Vol 49, No. 1, Jan 1979, pp 62-66. Available from:
A-14
EM 1110-2-1901
30 Sep 86

161. Coxon, R. E., and Crook, D. E., "Some Simple Approaches to Leakage
Detection in Dams," Transactions of the Twelfth International Congress
on Large Dams, Vol II, 1976, Mexico City, pp 527-540. Available from:
162. Dank, A. I., The Influence of Sand on the Engineering Properties of

Cement Bentonite Slurry Used for Cut-off Walls, M.S. Thesis, Aug 1981,
Northwestern University, Evaston, Ill. Available from: Technical
163. D'Appolonia, D. J., "Soil-Bentonite Slurry Trench Cutoffs," Journal of

the Geotechnical Engineering Division, Vol 106, No. GT4, Apr 1980,
American Society of Civil Engineers, New York. Available from: Techni-
cal Information Center, U. S. Army Engineer Waterways Experiment Station,
164. Davis, S. N., and Dewiest, R. J. M., Hydrogeology, Wiley, New York,
neer Waterways Experiment Station, P. O. Box 631, Vicksburg,
MS 39180-0631.
165. Denny, K. J., Grain Size Distribution and Its Effect on the Permeability
of Unconsolidated Sands, M.S. Thesis, 1965, University of Texas, Austin,
Tex. Available from: Technical Information Center, U. S. Army Engineer
166. Desai, C. S., "Flow Through Porous Media," Numerical Methods in

Geotechnical Engineering, C. S. Desai, and J. T. Christian, eds.,
McGraw-Hill, New York, 1977, pp 458-505. Available from: Technical
167. Desai, C. S., and Abel, J. F., Introduction to the Finite Element
Method, Van Nostrand Reinhold, New York, 1972. Available from:
168. Duguid, D. R., et al., "The Slurry Trench Cut-Off for the Duncan Dam,"
Canadian Geotechnical Journal, Vol 8, No. 1, Feb 1971, pp 94-108.
169. Dvinoff, A. H., Response of the Phreatic Surface in Earthen Dams to

Headwater Fluctuations, Ph. D. Thesis, 1970, Purdue University, West
Lafayette, Ind. Available from: Technical Information Center, U. S.
39180-0631.
A-15
EM 1110-2-1901
30 Sep 86
170. Eisenstein, Z., "Computer Analyses in Earth Dam Engineering," Computers

in Soil Mechanics: Present and Future, R. L. Schiffman, ed., MAA
Publishing Co., Taipei, 1979, pp 79-122. Available from: Technical
171. Fenn, C. D., The Effects of Montmorillonite and Kaolinite Dispersions on

the Permeability of a Porous Media, M.S. Thesis, 1966, Mississippi State
University, State College, Miss. Available from: Technical Information
172. Fetter, C. W., Jr. Applied Hydrology, Charles E. Merrill, Columbus,

Ohio, 1980. Available from: Technical Information Center, U. S. Army
39180-0631.
173. Fetzer, C. A., "Wolf Creek Dam - Remedial Work, Engineering Concepts
Actions, and Results," Transactions of the Thirteenth International
Congress on Large Dams, Vol 2, 1979, New Delhi, India, pp 57-82.
174. Forcheimer, P. H., Hydraulik, Teubner, Berlin, 1914. Available from:

175. Freeze, R. A., and Cherry, J. A., Groundwater, Prentice-Hall, New York,
1979. Available from: Technical Information Center, U. S. Army
39180-0631.
176. Fuquay, G. A,, "Foundation Cutoff Wall for Allegheny Reservoir Dam,"
Closure to discussion, Journal of the Soil Mechanics and Foundations
Division, pp 1363-1366, American Society of Civil Engineers, New York,
Vol 94, No. SM6, Nov 1968. Available from: Technical Information
177. Gebhart, L. R. "Foundation Seepage Control Options for Existing Dams,"

Inspection, Maintenance and Rehabilitation of Old Dams, American Society
of Civil Engineers, New York, 1974, pp 660-676. Available from:
178. Greer, D. M., Moorhouse, D. C., and Millet, R. A., "Sheet Pile Interlock
Seepage Study," Woodward-Moorhouse & Associates, Inc., New York, 1969.
Available from: Woodward-Clyde & Associates, Inc., Two Pennsylvania
Plaza, New York, NY 10001.
179. Hanna, T. H., "Diaphragm Walls," Developments in Soil Mechanics,

C. R. Scott, ed., Applied Science, London, 1978, pp 213-249. Available
A-16
EM 1110-2-1901
30 Sep 86

180. Harr, M. E., Groundwater and Seepage, McGraw-Hill Book Company, New
York, 1962. Available from: Technical Information Center, U. S. Army
39180-0631.
181. Harr, M. E., Mechanics of Particulate Media, McGraw-Hill, New York,

1977. Available from: Technical Information Center, U. S. Army
39180-0631.
182. ICOS, "The ICOS Company in Underground Works," Vol 3, 1968, Milano,
Italy. Available from: ICOS Corporation of America, Four West 58th
Street, New York, NY 10019.
183. Isaacs, L. T., "Adjustment of Phreatic Line in Seepage Analysis by Finite

Element Method," Research Report No. CE2, Mar 1979, Department of Civil
Engineering, University of Queensland, Australia. Available from:
184. Isaacs, L. T., "Location of Seepage Free Surface in Finite Element

Analyses," Transactions, Institution of Engineers, Australia, Vol CE22,
No. 1, Feb 1980, pp 9-16. Available from: Technical Information
185. Jacob, C. E., "Flow of Groundwater," Engineering Hydraulics, H. Rouse,

ed., Wiley New York, 1950, p 346. Available from: Technical Information
186. Jansen, R. B., "Engineering of Earth and Rockfill Dams in California,"

presented at Recent Developments in the Design and Construction of Earth
and Rockfill Dams, Mar 1968, University of California, Berkeley. Avail-
187. Jefferis, S. A., "Bentonite-Cement Slurries for Hydraulic Cut-Offs,"

Proceedings, Tenth International Conference on Soil Mechanics and
Foundation Engineering, Stockholm, Sweden, Vol 1, 1981, pp 435-440.
188. Jenkins, J. D., and Bankofier, D. E., "Hills Creek Dam Seepage Correc-
tion," Performance of Earth and Earth-Supported Structures, American
Soceity of Civil Engineers, New York, 1972, pp 723-733. Available from:
A-17
EM 1110-2-1901
30 Sep 86
189. Johnson Division, Universal Oil Products Co., Ground Water and Wells,
St. Paul, Minn., 1972. Available from: Technical Information Center, U.
S. Army Engineer Waterways Experiment Station, P. O. Box 631, Vicksburg,
MS 39180-0631.
190. Jones, J. C., "Deep Cut-Offs in Pervious Alluvium Combining Slurry

Trenches and Grouting," Proceedings of the 9th International Congress on
Large Dams, Istanbul, Turkey, Vol 1, 1967, pp 509-524. Available from:
191. Jones, L. B., "Reservoir Seepage and Ground Water Control, McNary
Reservoirs," paper presented at the American Society of Civil Engineer
Convention, Phoenix, Ariz., 1961. Available from: Shannon & Wilson,
Inc., 1105 North 38th Street, Seattle, WA 98103.
192. Kaufman, R. I., "Effect of Top Stratum Permeability on Design of Seepage

Control Measures," In Situ Measurement of Soil Properties, North
Carolina State University, Vol II, 1976, pp 39-45. Available from:
193. Klohn, E. J., "Seepage Control for Tailings Dams," Mine Drainage,
G. O. Argall, Jr., ed., Miller Freeman Publications, San Francisco, 1979,
39180-0631.
194. Knabach, M. L., and Dingle, S. O., "Slurry Trench Cutoff in Dam
Construction," Paper No. 74-2014 presented at American Society of
Agricultural Engineers Meeting, Stillwater, Okla., 1974. Available
from: American Society of Agricultural Engineers, P. O. Box 229, St.
Joseph, MI 49085.
195. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic Emission
(Microseismic) Monitoring of Earth Dams," The Evaluation of Dam Safety,
American Society of Civil Engineers, New York, 1977, pp 274-291.
196. Kramer, H., "Deep Cutoff Trench of Puddled Clay for Earth Dam and Levee
Protection," Engineering News Record, Vol 136, No. 26, 27 Jun 1946,
39180-0631.
197. Krumbein, W. C., and Pettijohn, F. J., Manual of Sedimentary

Petrography, Appleton-Century Crofts, New York, 1938. Available from:
A-18
EM 1110-2-1901
30 Sep 86
198. Krumbein, W. C . , and Sloss, L L ., Stratigraphy and Sedimentation,

W. H. Freeman, San Francisco, 1951. Available from: Technical Informa-
tion Center, U. S. Army Engineer Waterways Experiment Station, P. O.
199. Kruseman, G. P., and DeRidder, N. A., "Analysis and Evaluation of Pump-
ing Test Data," Bulletin 11, 1970, International Institute for Land
Reclamation and Improvement, Wageningen, Netherlands. Available from:
200. Lambe, T. W., Soil Testing for Engineers, Wiley, New York, 1951. Avail-
201. Lambe, T. W., and Whitman, R. V., Soil Mechanics, Wiley, New York, 1969.
202. Lane, K. S., and Wohlt, P. E., "Performance of Sheet Piling and Blankets
for Sealing Missouri River Reservoirs," Transactions of the Seventh
International Congress on Large Dams, Vol 4, 1961, pp 25, Rome, Italy.
203. La Russo, R. S., "Wanapum Development - Slurry Trench and Grouted Cut-
Off," Grouts and Drilling Muds in Engineering Practice, 1963,
pp 196-201, Butterworths, London. Available from: Technical Informa-
204. Lefebvre, G. C. L., Pare, J. J., and Tournier, J.-P., "Effectiveness of

Seepage Control Elements for Embankments on Semipervious Foundations,"
Canadian Geotechnical Journal, Vol 18, No. 4, Nov 1981, pp 572-576.
205. Leonards, G. A., ed., Foundation Engineering, McGraw-Hill, New York,

MS 39180-0631.
206. Leps, T. M., "Flow Through Rockfill," Embankment-Dam Engineering,

R. C. Hirschfield and S. J. Poulous, eds., Wiley, New York, 1973,
MS 39180-0631.
207. Lewis, D. C., Kriz, G. J., and Burgy, R. H., "Tracer Dilution Sampling
Techniques to Determine Hydraulic Conductivity of Fractured Rock," Water
Resource Research, Vol 2, No. 3, 1966. Available from: Technical
A-19
EM 1110-2-1901
30 Sep 86

208. Ley, J. E. "Foundations of Existing Dams - Seepage Control," Inspec-

tion, Maintenance and Rehabilitation of Old Dams, American Society of
Civil Engineers, New York, 1974, pp 584-608. Available from: Technical
209. Logani, K. L., and Lhez, M. H. H., "Dispersive Soils Used in the
Construction of the Ullum Dam in Argentina," Proceedings of the Sixth
Panamerican Conference on Soil Mechanics and Foundation Engineering,
Vol 111, Dec 1979, Lima, Peru, pp 394-411. Available from: Technical
210. London, A. G., "The Computation of Permeability from Simple Soil Tests,"
Geotechnique, Vol 3, No. 4, 1952, pp 165-183. Available from: Techni-
211. Louis, c., "A Study of Groundwater Flow in Jointed Rock and Its Influence
on the Stability of Rock Masses," Rock Mechanics Research Report No. 10,
1969, Imperial College, London, England. Available from: Technical
212. Lovenbury, H. T., "The Detection of Leakage Through the Core of an

Existing Dam," Field Instrumentation in Geotechnical Engineering, Wiley,
39180-0631.
213. Lowe, J., III. "Foundation Design - Tarbela Dam," The Fourth Nabor
Carrillo Lecture, Mexican Society for Soil Mechanics, Mexico, 1978.
214. Lynch, E. J., Formation Evaluation, Harper and Row, New York, 1962.
215. Maini, Y. N. T., In Situ Hydraulic Parameters in Jointed Rock: Their

Measurement and Interpretation, Ph. D. Dissertation, 1971, Imperial
College, London, England. Available from: Technical Information Center,
216. Mansur, C. I,, and Dietrich, R. J., "Pumping Test to Determine

Permeability Ratio," Journal of the Soil Mechanics and Foundations
Division, Vol 91, No. SM4, Jul 1965, American Society of Civil Engi-
neers, New York, pp 151-183. Available from: Technical Information
A-20
EM 1110-2-1901
30 Sep 86

217. Mantei, C. L., and Harris, D. W., "Finite Element Seepage Analysis on
Reclamation Dams," Preprint 3691, paper presented at American Society of
Civil Engineers Convention and Exposition, Atlanta, Ga., Oct 1979.
Available from:. Technical Information Center, U. S. Army Engineer
218. Marsal, R. J., and Resendiz, D., "Effectiveness of Cutoffs in Earth

Foundations and Abutments of Dams," Proceedings of the Fourth
Pan-American Conference on Soil Mechanics and Foundation Engineering,
Vol 1, Jun 1971, San Juan, Puerto Rico, pp 237-312. Available from:
Technical Information Center,, U. S. Army Engineer Waterways Experiment
219. Masch, F. D., and Denny, K. J., "Grain Size Distribution and Its Effect
on the Permeability of Unconsolidated Sands," Water Resources Research,
Vol 2, No. 4, 1966, pp 665-677. Available from: Technical Information
220. McDaniel, T. N., and Decker, R. S., "Dispersive Soil Problem at Los
Esteros Dam," Journal of the Geotechnical Engineering Division, Vol 105,
No. GT9, Sep 1979, pp 1017-1030, American Society of Civil Engineers,
New York. Available from: Technical Information Center, U. S. Army
MS 39180-0631.
221. Meier, J. G., and Rettberg, W. Q., "Report on Cement-Bentonite Slurry

Trench Cutoff Wall: Filden Tailings Project," Tailings Disposal Today,
Vol 2, pp 341-365 , 1978, Miller Freeman Publications, San Francisco.
222. Miller, E. A., and Salzman, G. S., "Value Engineering Saves Dam Proj-
ect," Civil Engineering, American Society of Civil Engineers, Vol 50,
No. 8, Aug 1980, pp 51-55. Available from: Technical Information
223. Millet, R. A., and Perez, J.-Y., "Current USA Practice: Slurry Wall
Specifications," Journal of the Geotechnical Engineering Division,
American Society of Civil Engineers, Vol 107, No. GT8, Aug 1980,
39180-0631.
224. Milligan, V., "Field Measurement of Permeability in Soil and Rock," In

Situ Measurement of Soil Properties, American Society of Civil Engi-
neers, New York, Vol 2, 1976, pp 3-36. Available from: Technical
A-21
EM 1110-2-1901
30 Sep 86

225. Missbach, A., "Listy Cukrovar," Vol 55, 1937, p 293, Prague,
Czechoslovakia. Available from: Technical Information Center, U. S.
39180-0631.
226. Mitchell, J. K., Fundamentals of Soil Behavior, Wiley, New York, 1976.
227. Mitchell, J. K., Guzikowski, F., and Villet, W. C. B., "The Measurement
of Soil Properties In-Situ," Report No. LBL-6363, Mar 1978, Department
of Civil Engineering, University of California, Berkeley, Calif. Avail-
228. Morgenstern, N., and Amir-Tahmasseb, I., "The Stability of a Slurry

Trench in Cohesionless Soils," Geotechnique, Vol 15, No. 1, Dec 1965,
Engineer Waterways Experiment Station, P. Q. Box 631, Vicksburg,
MS 39180-0631.
229. Muskat, M., The Flow of Homogeneous Fluids Through Porous Media,
J. W. Edwards, Ann Arbor, Mich., 1946. Available from: Technical Infor-
mation Center, U. S. Army Engineer Waterways Experiment Station, P. O.
230. Nash, J. K. T. L., "Slurry Trench Walls, Pile Walls, Trench Bracing,"
Proceedings, Sixth European Conference on Soil Mechanics and Foundation
Engineering, Vienna, Austria, Vol 2.1, 1976, pp 27-32. Available from:
231. Nash, J. K. T. L., and Jones, G. K., "The Support of Trenches Using
Fluid Mud," Grouts and Drilling Muds in Engineering Practice, 1963,
pp 177-180, Butterworths, London. Available from: Technical Informa-
232. Neuman, S. P. "Analysis of Pumping Test Data from Anisotropic Uncon-

fined Aquifers Considering Delayed Gravity Response," Water Resource
Research, Vol 11, 1975, pp 329-342. Available from: Technical Informa-
233. Olson, R. E., and Daniel, D. E., "Field and Laboratory Measurement of
the Permeability of Saturated and Partially Saturated Fine-Grained
Soils," paper presented at the American Society for Testing and
Materials Symposium on Permeability and Groundwater Contaminant
Transport, Philadelphia, Pa., Jun 1979. Available from: Technical
A-22
EM 1110-2-1901
30 Sep 86

234. Parker, D. G., and Thornton, S. I., "Permeability of Fly Ash and Fly Ash
Stabilized Soils," Dec 1976, Department of Civil Engineering, University
of Arkansas, Fayetteville, Ark. Available from: Technical Information
235. Pavlovsky, N. M., Collected Works, 1956, Akad. Nauk USSR, Leningrad
(cited in Harr 1977). Available from: Technical Information Center,
236. Pavlovsky, N. M., "Seepage Through Earth Dams," Instit. Gidrotekhniki i

Melioratsii, 1931, Leningrad, USSR, translated by U. S. Army Corps of
Engineers. Available from: Technical Information Center, U. S. Army
MS 39180-0631.
237. Perloff, W. H., and Baron, W., Soil Mechanics, Ronald Press, New York,
MS 39180-0631.
238. Pettyjohn, W. A,., et al., "A Ground-Water Quality Atlas of the United
States," May 1979, National Demonstration Water Project, Washington,
D. C. Available from: Technical Information Center, U. S. Army Engi-
MS 39180-0631.
239. Polubarinova-Kochina, P. Ya., "Concerning Seepage in Heterogeneous

(Two-Layered) Media," Inzhenernii Sbornik, Vol 1, No. 2, 1941 (cited in
Harr 1977). Available from: Technical Information Center, U. S. Army
MS 39180-0631.
240. Prickett, T. A., "Ground-Water Computer Models - State of the Art,"

Ground Water, Vol 17, No. 2, Mar-Apr 1979, pp 167-173. Available from:
241. Prickett, T. A., "Modeling Techniques for Groundwater Evaluation,"

Advances in Hydroscience, Vol 10, 1975, V. T. Chow, ed., Academic Press,
New York. Available from: Technical Information Center, U. S. Army
MS 39180-0631.
242. Prickett, T. A., and Lonnquist, C. G., "Selected Digital Computer

Techniques for Groundwater Resource Evaluation," Bulletin 55, 1971,
Illinois State Water Survey, Urbana, Ill. Available from: Technical
A-23
EM 1110-2-1901
30 Sep 86

243. Quirk, J. P., and Schofield, R. J., "The Effect of Electrolyte

Concentration on Soil Permeability," Journal of Soil Science, Vol 6,
No. 2, 1955, pp 163-178. Available from: Technical Information Center,
244. Radukic, V., "Observed Behavior of Embankment Dams at Some Typical

Interfaces," Proceedings, Thirteenth International Congress on Large
Dams, India, Vol 1, 1979, pp 1005-1922. Available from: Technical
245. Rouse, H., and Ince, S., History of Hydraulics, State University of
Iowa, Iowa City, 1957. Available from: Technical Information Center,
246. Royster, D. L. "Some Observations on the Use of Horizontal Drains in

the Correction and Prevention of Landslides," Tennessee Department of
Transportation, Nashville, Tenn. 1977. Available from: Technical
247. Rushton, K. R., and Redshaw, S. C., Seepage and Groundwater Flow, Wiley,
New York, 1979. Available from: Technical Information Center, U. S.
Army Engineer Waterways Experiment Station, P. O. Box 631, Vicksburg,
MS 39180-0631.
248. Ryan, C. R., "Slurry Cut-Off Walls, Design and Construction," paper
presented at Slurry Wall Construction, Design, Techniques, and
Procedures Course, Chicago, Ill., Apr 1976. Available from: GO-CON,
Inc., P. O. Box 17380, Pittsburgh, PA 15235.
249. Ryan, C. R., "Slurry Cut-Off Walls, Design Parameters and Final
Properties," paper presented at Slurry Wall Construction, Design,
Techniques, and Procedures Course, Miami, Fla., Feb 1977. Available
from: GO-CON, Inc., P. O. Box 17380, Pittsburgh, PA 15235.
250. Schwartz, P. H., Analysis and Performance of Hydraulic Sand-Fill Levees,

Ph. D Thesis, University of Iowa City, Iowa, May 1976. Available from:
251. Scott, R. F., Principles of Soil Mechanics, Addison-Wesley, Reading,

Mass., 1963. Available from: Technical Information Center, U. S. Army
MS 39180-0631.
A-24
EM 1110-2-1901
30 Sep 86
252. Sherard, J. L., A Study of the Influence of the Earthquake Hazard on the
Design of Embankment Dams, Jul 1966, Woodward, Clyde, Sherard & Asso-
ciates, Oakland, Calif. Available from: Technical Information Center,
253. Sherard, J. L., "Earthquake Considerations in Earth Dam Design," Journal

of the Geotechnical Engineering Division, Vol 93, No. SM4, Jul 1967,
American Society of Civil Engineers, New York. Available from: Techni-
254. Sherard, J. L., "Sinkholes in Dams of Coarse, Broadly Graded Soils,"

Transactions of the 13th International Congress on Large Dams, New
Delhi, India, Vol 11, 1979, pp 25-35. Available from: Technical
255. Sherard, J. L., "Some Considerations Concerning Underseepage Control for

Earth Dams," Proceedings of the Symposium on Earth and Rockfill Dams,
India, Vol 1, 1968, pp 204-214. Available from: Technical Information
256. Sherard, J. L., et al., Earth and Earth-Rock Dams, Wiley, New York,
MS 39180-0631.
257. Singh, B., and Sharma, H. D., Earth and Rockfill Dams, Sarita Prakashan,
India, 1976. Available from: Technical Information Center, U. S. Army
MS 39180-0631.
258. Snow, D. T., A Parallel Plate Model of Fractured Permeability Media,

Ph. D Dissertation, 1965, University of California, Berkeley, Calif.
259. Soletanche, "Laguna Dam, Mexico, 1970," Paris, France. Available from:
RECOSOL, Rosslyn Center, 1700 North Moore Street, Suite 2200, Arlington,
VA 22209.
260. Soletanche, "Los Reyes Dam, Mexico, 1972," Paris, France. Available
from: RECOSOL, Rosslyn Center, 1700 North Moore Street, Suite 2200,
Arlington, VA 22209.
261. Soletanche, "Nexapa Dam, Mexico, 1971-1972," Paris, France. Available

from: RECOSOL, Rosslyn Center, 1700 North Moore Street, Suite 2200,
Arlington, VA 22209.
A-25
EM 1110-2-1901
30 Sep 86
262. Soletanche, "Tenango Dam, Mexico, 1971," Paris, France. Available from:
VA 22209.
263. Soletanche, "The Razaza Dam, Iraq, 1969," Paris, France. Available from:
VA 22209.
264. Sowers, G. F., Earth and Rockfill Dam Engineering, Asia Publishing
House, New York, 1962. Available from: Technical Information Center,
265. Sowers, G. F., "Earth Dam Failures," Lectures of the Seminar, Failures
of Large Dams, Reasons and Remedial Measures, W. Wittke, ed., Publica-
tion No. 4, 1977, Institute for Foundation Engineering, Soil Mechanics,
Rock Mechanics, and Waterways Construction, Aachen, Germany, pp 178-226.
266. Stephenson, D., Rockfill Hydraulic Engineering, Elsevier, New York,

1979. Available from: Technical Information-Center, U. S. Army Engi-
MS 39180-0631.
267. Strohm, W. E., Nettles, E. H., and Calhoun, C. C., "Study of Drainage
Characteristics of Base Course Materials," Symposium on Subsurface
Drainage, Highway Research Record 203, 1967, pp 8-28, Highway Research
Board. Washington. D. C. Available from: Technical Information Center,
268. Taylor, D. W., Fundamentals of Soil Mechanics, Wiley, New York, 1948.
Available from: Technical Information Center, U. S. Army Engineer Water-
269. Taylor, H., and Chow, Y. M., "Design Monitoring, and Maintaining
Drainage System of a High Earthfill Dam," Transactions, Twelfth
International Congress on Large Dams, Mexico, Vol 2, 1976, pp 147-167.
270. Telford, W. F., et al., Applied Geophysics, Cambridge University Press,

New York, 1976. Available from: Technical Information Center, U. S.
MS 39180-0631.
271. Telling, R. M., Menzies, B. K., and Coulthard, J. M., "A Design Method
for Assessing the Effectiveness of Partially Penetrating Cut-Off Walls,"
Ground Engineering, Vol 11, No. 8, Nov 1978, pp 48-51. Available from:
A-26
EM 1110-2-1901
30 Sep 86
272. Telling, R. M., Menzies, B. K., and Simons, N. E., "Cut-Off Efficiency,
Performance and Design," Ground Engineering, Vol 11, No. 1, Jan 1978a,
MS 39180-0631.
273. Telling, R. M., Menzies, B. K., and Simons, N. E., "The Effectiveness of
Jointed Cut-Off Walls Beneath Dams on Pervious Soil Foundations," Ground
Engineering, Vol 11, No. 4, May 1978b, pp 27-37. Available from:
274. Terzaghi, K., Theoretical Soil Mechanics, Wiley, New York, 1943. Avail-
275. Terzaghi, K., and Peck, R. B., Soil Mechanics in Engineering Practice,
2d ed., Wiley, New York, 1967. Available from: Technical Information
276. Theis, C. V. "The Relation Between the Lowering of the Piezometric

Surface and the Rate and Duration of Discharge of a Well Using
Groundwater Storage," Transactions of American Geophysics Union, Vol 16,
MS 39180-0631.
277. Thomas, H. H., The Engineering of Large Dams, Vol II, Wiley, 1976.
278; Todd, D. D., and Bear, J., "River Seepage Investigation," Water Resources
Center Contribution No. 20, Sep 1959, Hydraulic Laboratory, University of
California, Berkeley. Available from: Technical Information Center,
279. Todd, D. K., Groundwater Hydrology, 2nd ed., Wiley, New York, 1980.
280. Turkish National Committee, "Investigation of Leakages at Keban Dam,"

Transactions of the Twelfth International Congress on Large Dams, Vol II,
1976, Mexico City, pp 459-483. Available from: Technical Information
281. Twelker, N. H., "Analysis of Seepage in Pervious Abutments of Dams,"

Proceedings of the Fourth International Conference on Soil Mechanics and
Foundation Engineering, London, England, Vol 2, 1957, pp 389-393.
A-27
EM 1110-2-1901
30 Sep 86
Available from: Technical Information Center, U. S. Army Engineer Water-

282. Vaughan, P. R., "Design of Filters for the Protection of Cracked Dam
Cores Against Internal Erosion," Preprint 3420, Oct 1978, American
Society of Civil Engineers, New York. Available from: Technical Infor-
mation Center, U. S. Army Engineer Waterways Experiment Station, P. O.
283. Vaughan, P. R., et al., "Cracking and Erosion of the Rolled Clay Core of
Balderhead Dam and the Remedial Works Adopted for Its Repair," Transac-
tions of the Tenth International Congress on Large Dams, Vol 1, 1970,
Montreal, Canada, pp 73-93. Available from: Technical Information
284. Veder, C., "Closing Address," Diaphragm Walls and Anchorages, 1975,
pp 221-225, Institution of Civil Engineers, London. Available from:
285. Veder, C., "Excavation of Trenches in the Presence of Bentonite

Suspension for the Construction of Impermeable and Load-Bearing
Diaphragms," Grouts and Drilling Muds in Engineering Practice,
Butterworths, London, 1963, pp 181-188. Available from: Technical
286. Vennard, J. K., Elementary Fluid Mechanics, 4th ed., Wiley, New York,
MS 39180-0631.
287. Walton, W. C., "Selected Analytical Methods for Well and Aquifer Evalua-
tion," Bulletin 49, 1962, Illinois State Water Survey, Urbana, Ill.
288. Wei, C.-Y., and Shieh, W. Y. J., "Transient Seepage Analysis of Guri
Dam," Journal, Technical Councils, American Society of Civil Engineers,
Vol 105, No. TCl, Apr 1979, pp 135-147. Available from: Technical
289. Wilkins, J. K., "Flow of Water Through Rockfill and Its Application to
the Design of Dams," Proceedings, Second Australia-New Zealand
Conference on Soil Mechanics and Foundation Engineering, New Zealand,
MS 39180-0631.
A-28
EM 1110-2-1901
30 Sep 86
290. Wilson, C. R., and Witherspoon, P. A., "An Investigation of Laminar Flow
in Fractured Porous Rocks," 1970, Department of Civil Engineering,
University of California, Berkeley, Calif. Available from: Technical
291. Wilson, S. D., and Marsal, R. J., "Current Trends in Design and Con-
struction of Embankment Dams," 1979, American Society of Civil Engi-
neers. Available from: Technical Information Center, U. S. Army Engi-
MS 39180-0631.
292. Winter, G., and Nilson, A. H., Design of Concrete Structures, 9th ed.,
McGraw-Hill, New York, 1979. Available from: Technical Information
293. Xanthakos, P. P., Slurry Walls, McGraw-Hill, New York, 1979. Available
294. Yong, R. N., and Warkentin, B. P., Introduction to Soil Behavior,

Macmillan, New York, 1966. Available from: Technical Information
Center, U. S. Army Engineer Waterways Experiment Station, P. O.. Box 631,
295. Younger, J. S., and Lim, C. I., "An Investigation into the Flow Behavior
Through Compacted Saturated Fine-Grained Soils with Regard to Fines
Content and over a Range of Applied Hydraulic Gradients," Fundamentals
of Transport Phenomena in Porous Media, International Association for
Hydraulic Research, New York, 1972, pp 312-326. Available from:
A-29
EM 1110-2-1901
30 Sep 86
APPENDIX B
APPROXIMATE METHODS FOR ANALYSIS OF FLOW PROBLEMS
B-1. Introduction. As previously mentioned in Chapter 4, various methods are

available, in addition to flow nets, for solving idealization of seepage prob-
lems. As shown in figure 4-3, these methods include electrical analogy,
hydraulic or sand tank models, viscous flow models, method of fragments, finite
difference method, and finite element method. Prior to conducting an analysis,
the problem to be studied must be defined in terms of:
a. Aquifer and embankment dimensions.
b. Coefficients of permeability of the embankment and foundation soils.
c. Horizontal to vertical permeability ratios.
d. Boundary conditions (impermeable and symmetrical).
e. Exits and entrances (fixed potential areas).
f. Head versus time relationships for unsteady flow.
Sensitivity studies may be run to establish the effect of parameters not known
accurately.
B-2. Electrical Analogy.
a. General. Processes which involve movement of current due to differ-

ences in energy potential operate on the same principles as movement of con-
fined ground water as shown in table B-l. Therefore, to obtain the pattern of
equipotential lines or flow lines (see figure 4-4), the flow domain is trans-
ferred by an electrical conductor of similar geometric form as first proposed
by Pavlovsky in 1918 (Harr 1962). Electrical analogies may involve two-
dimensional conducting paper models or three-dimensional tanks containing
aqueous solution.
b. Two-Dimensional Models. When field conditions can be approximated

by a two-dimensional plan or section, teledeltos conducting paper models may
be used to obtain a flow net. Two-dimensional teledeltos models are simple to
use and can accommodate various geometries. However, it is difficult to simu-
late varying permeabilities and they are generally restricted to steady state
confined aquifers (Bear 1972, Boer and Molen 1972).
c. Three-Dimensional Models. The use of electrical analogy models is

described by various authors (Zangar 1953, Todd and Bear 1959, and Duncan
1963). The three-dimensional electrical analogy model at the U. S. Army
Engineer Waterways Experiment Station (WES) (see figure B-l) is a plexiglass
tank filled with dilute copper sulfate solution and having a calibrated
elevated carrier assembly for the accurate placement of a point electrode probe
B-1
EM 1110-2-1901
30 Sep 86
Table B-1. Analogy Between Darcy's Law and Ohm's Law (a)
Darcy's Law Ohm's Law
Q = rate of flow or water I = current (rate of flow of electricity)
K = coefficient of permeability K' = conductivity coefficient
A = cross-sectional area A ' = cross-sectional area
H = head producing flow V = voltage producing current
L = length of path of percolation L' = length of path of current
(a) From Bureau of Reclamation. 127
anywhere in the fluid. Extensive use of the WES model has been made to:
(1) Determine uplift values and seepage quantities for use in the
design of Columbia Lock and Dam, Louisiana (Duncan 1962).
(2) Determine the uplift values and seepage quantities for fully and
partially penetrating well arrays from line and circular sources (Duncan 1963,
Banks 1963, and Banks 1965).
(3) Determine uplift pressures beneath the spillway, piezometric heads

at the downstream toe of the dam, and total seepage quantities for use in the
design of Oakley Dam, Illinois (McAnear and Trahan 1972).
B-3. Sand Tank Model. The sand tank model (hydraulic model), as shown in
figure B-2, consists of a rigid, watertight container with a transparent front,
filled with sand, deaired water,(1) and measuring devices. The geometry of the
sand tank corresponds to that of the prototype. The sand may be placed under
water to provide a homogeneous condition, or layers of different sand sizes may
be used to study anisotropy. If the flow is unconfined and the same material
is used for model and prototype, the capillary rise must be compensated for in
the model. When a steady-state flow is reached, dye can be introduced at
various points along the upstream boundary close to the front wall to form
traces of the streamlines. Piezometers are used to measure the pressure heads
at various locations (Bear 1972 and Harr 1962). A sand tank model was employed
to investigate the effect of length of horizontal drain on the through seepage
flow nets and quantities for a homogeneous and isotropic sand embankment
(1)
For prolonged tests, disinfectants such as Formol should be added to the
water to prevent bacterial growth that causes clogging (Bear 1972).
B-2
B-3
EM 1110-2-1901
30 Sep 86
36
Figure B-1. Three-dimensional electrical analogy apparatus (after Duncan )
EM 1110-2-1901
30 Sep 86
PARTIALLY PENETRATING WELL

WITH A CIRCULAR SOURCE
SEEPAGE FROM CANALS
TRACER DYE
SEEPAGE IN EMBANKMENTS
Figure B-2. Hydraulic or sand tank model
(prepared by WES)
B-4
EM 1110-2-1901
30 Sep 86
(Brand and Armstrong 1968). Sand tank models are also used extensively in
petroleum engineering, ground-water quality, and pollution research (Bear 1972
and Prickett 1979).
B-4. Viscous Flow Models. The viscous flow model, also called the Hele-Shaw
or parallel plate model, is based on the similarity between the differential
equations governing saturated flow in a porous medium and those describing the
flow of a viscous liquid in the narrow space between two parallel plates. The
viscous flow model contains the shape of the structure to be Investigated and
once a steady-state flow is obtained, colored dyes can be injected along the
upstream edge and patterns of streamlines can be observed. A camera (movie or
still) is normally used to record the results of experiments. Inhomogeneous
hydraulic conductivity, such as would exist in a zoned earth dam, can be simu-
lated by varying the width of the interspace between the parallel plates, as
shown in figure B-3. The viscous flow model experiments should be conducted in
a temperature-controlled room because viscosity plays an important role in
analog scaling. If this is not feasible, the temperature should be measured at
all inflow and outflow points during the test and scales must be recomputed
according to the varying average temperature of the liquid in the model (Bear
1972 and Harr 1962). A viscous flow model was constructed at WES to simulate
seepage conditions induced in streambanks by sudden drawdown of the river level.
The results from the model study compared favorably with field observations,
finite difference, and finite element methods (Desai 1970 and Desai 1973).
B-5. Method of Fragments.
a. General. The method of fragments is an approximate analytical

method for the computation of flows and pressure heads for any ground-water
system. The underlying assumption of this procedure developed by Pavlovsky in
1935 (Pavlovsky 1956 and Harr 1962) is that equipotential lines at various
critical locations in the flow region can be approximated by straight vertical
lines. These equipotential lines divide the flow region into parts or frag-
ments. Other assumptions inherent in the method of fragments procedure are
(a) Darcy's law is valid, (b) steady-state flow exists, and (c) the soil medium
is approximated as a single homogeneous and isotropic layer or at series of
such layers. The transformation of anisotropic soil to an equivalent isotropic
soil is described in Section 4.7 of this manual.
b. Basic Concepts. The quantity of flow through a single fragment is

computed as:
(B-1)
where
k - coefficient of permeability
h i = head loss through the fragment
= dimensionless form factor, = Ne/Nf
B-5
EM 1110-2-1901
30 Sep 86
a. Homogeneous earth dam
b. Zoned earth dam
140
Figure B-3. Viscous flow model (courtesy of Bear )
B-6
EM 1110-2-1901
30 Sep 86
Because the fragment boundaries consist of equipotential lines, the flow

through each fragment must be equal to the total flow through the system. Thus
(B-2)
Since summation of the head loss in each fragment is equal to the total head
loss, the total quantity of flow can be expressed as
(B-3)
where h is the total head loss through the section . Along the same line, the
head loss in each fragment can be calculated from
(B-4)
The head loss along any impermeable boundary of a fragment is assumed to

change linearly. Thus the head loss within fragment i up to point A is
equal to the head loss in the fragment times the ratio of the length of the
boundary to point A to the total length of boundary. The basic concept of the
method of fragment procedure is to break the flow region into parts for which
the form factor is shown in figure B-4 (Harr 1977). This manual will describe
how to calculate the factors for each type of fragment (Harr 1962 and Harr
1977).
c. Fragment Types. There are currently nine different fragment types.

Of these, the first six are for confined flow while the last three are for
unconfined flow.
(1) Type I. This fragment type represents a region of parallel

horizontal flow between impervious boundaries. For this internal type
fragment, shown in figure B-5a, the flow per unit width is equal to
(B-5)
B-7
EM 1110-2-1901
30 Sep 86
Thus from Equation B-l, the form factor is
(B-6)
An elemental Type I section shown in figure B-5b illustrates that
(B-7)
This elemental section will be used to derive the form factors for fragment
types IV, V, and VI.
(2) Type II. This fragment type represents a vertical impervious

boundary embedded a length S into a pervious layer of thickness T . This
fragment can represent either an entrance condition (figure B-6a) or an exit
condition (figure B-6b). The form factor is obtained from the plot in
figure B-4 where the scale of is given as one-half the reciprocal of
Q/kh or
(B-8)
The form factor could also be expressed as the ratio of the elliptic integral
of the first kind with modulus m over the elliptic integral of the complemen-
tary modulus, m' . For this fragment type, the modulus value is a function of
the ratio S/T . The graph in figure. B-7 was obtained by solving the elliptic
integrals for various combinations of S/T . For the type II fragments, the
ratio of b/T equals 0 .
(3) Type III. This type of fragment represents an impervious layer of

length b , a vertical boundary of depth S , in a pervious layer of thickness
T . Either of the sections shown in figure B-8 can represent this fragment
type. The form factor is obtained directly from figure B-7 with b/T other
than zero. For this case, the elliptic integral modulus is a function of both
b/T and S/T .
(4) Type IV. This type is an internal fragment with boundary length
b , embedment length S , in a pervious layer of thickness T . Figure B-9a
illustrates the two possible configurations. Pavlovsky divided the flow region
into active and passive parts based on the results of electrical analogue tests
as shown in figure B-9b by line AB . An angle of 45 deg was assumed for the
line dividing the two parts of the fragment. This resulted in two cases,
depending on the relation between b and S . For the case where b < S , the
B-8
Fragmcnl Form f<Klor, 11• t h is head Fragment Form fm.:lor. 11t ( h is head
Illustration llluslruliun
1rpc lns'i lhrou~h fru~mcnl) luss lhrou~h fntKmenll
·~"~- -- - - -- - -----·---
L s 2s:
ct• ... ,_ ~ = 2 In (I + :~ I
II v L2!2s:
~=21n (I+ ~)+L_~ 21
L 2:s' + s":
II ~= t<th),Fig.5-13 ~ = In I (I + .!:a
> (1 +~.:I I
II
I. - ( s' + s"l
+-------
T
VI
L ::s s' + s":
~ = In I (I + a~:)(I + II~:) I
where L+(s'-s"l
Ill ~ = I (~h), Fig. 5-13 ,'1' = - - - - 2 -
h" =!_,=(J'_-__!:'1
B-9
VII
b::ss: b y
«<•=ln(l+;;-)
IV VIII
1l>=ln(i+~)
+ b-
- ---'
T
IX
EM 1110-2-1901
30 Sep 86
Figure B-4. Summary of fragment types and form factors (courtesy of McGraw-Hill
181
Book Company )
EM 1110-2-1901
30 Sep 86
Figure B-5. Type I fragment (courtesy of McGraw-Hill

181
Book Company )
Figure B-6. Type II fragment (courtesy of McGraw-Hill

181
Book Company )
B-10
EM 1110-2-1901
30 Sep 86
1.5
1.4 l
j
1.3
1.2 \ I u
=-------_:~~ "- - - ~
·...Oo
i
1.1 \ I I - - - - - oo·a··
I -- - - - - - - - - 0 P.O. ..wa·O;t:E""
o
1.0
'r\ I I
I
I ,;;,,;gy:;..J§_2"fH··£C
~/,1'/ffiW/~$&##/m~r
0.9
\I I
\t I
-1~ 0.8
II
0.7
r--... ~ ~~ I
ol~
0.6
~~ I I
.-.!!so ~.
I
...... I
0.5
0. 7;-~~ '-
1 I
I
0.4
0.3
l.OO-
1.25
r---r---~
I ~ I
:::::::::::±-::~ ~
..:,~
I
1.50
0.2 -...~
~
I
........
~:
0.1
~
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
_s
T
h2 TTb
(f m = cos Ti
17"$
tan 2 t + tan 2 2T
TT S
Q 1 4
for m s 0.3, kh = -;r In m
09...Q_ _ _ ,.
for m 2 ~ . . kh - 1 2
21n(--...':!!.)
16
Figure B-7. Quantity of discharge for symmetrically

placed pilings (courtesy of McGraw-Hill Book
181
Company )
B-11
EM 1110-2-1901
30 Sep 86
(a) (b)
Figure B-8. Type III fragment (courtesy of McGraw-Hill
Book Company 181 )
active zone is composed of elements of type I fragments of width dx illus-

trated in figure B-9c. The form factor is the integral of dx over y from
0 to b which results in a form factor of
(B-9)
If b > S, then the fragment can be divided into two fragments as shown in
figure B-9e. The first is a type IV with b > S and the second is a type I
fragment with L equal to b - S . Thus the form factor is the sum of the
form factors which would be
(B-10)
(5) Type V. This fragment type has two vertical boundaries of equal
embedment S in a pervious layer of thickness T . As shown in figure B-10,
the form factor for this fragment is twice that for the type IV fragment.
B-12
EM 1110-2-1901
30 Sep 86
Figure B-9. Type IV fragment (courtesy of McGraw-Hill Book

181
Company )
B-13
EM 1110-2-1901
30 Sep 86
Figure B-10. Type V fragment (courtesy of McGraw-Hill

181)
Book Company
Since there were two cases of type IV fragment, there are two cases for the
type V fragment. The two cases are for L < 2s and L > 2S . For the first
case, the form factor is
(B-11)
For the second case which consists of a type I fragment within two type IV
fragments, the form factor is
(B-12)
(6) Type VI. This fragment type, illustrated in figure B-11, is the
same as the type V fragment except that the embedment lengths are different.
Using the same approximations as in fragment type IV, there are two cases for
the form factor. For the first case where L > (S' + S"), the form factor is
(B-13)
B-14
EM 1110-2-1901
30 Sep 86
Figure B-11. Type VI fragment (courtesy of McGraw-Hill

181
Book Company )
For the second case where L < (S' + S"), the form factor is
(B-14)
where
(7) Type VII. This fragment represents the condition of unconfined

flow. This flow is characterized by having one boundary of the flow domain as
a free surface (line AB in figure B-12). This free surface separates the
saturated region from that region where no flow occurs. From Darcy's law and
Dupuit's assumptions, the hydraulic gradient is (h1 - h2)/L and the
cross-sectional area is (h1 + h2)/2 , thus the flow is
(B-15)
B-15
EM 1110-2-1901
30 Sep 86
Figure B-12. Type VII fragment (courtesy of McGraw-Hill

181
Book Company )
From this, the form factor is
(B-16)
(8) Type VIII. This fragment type represents an upstream slope

entrance condition on an earth dam of height hd and is illustrated in
figure B-13. It was assumed that the curve streamlines (cd) could be
approximated by horizontal flow channels of length ed (Pavlovsky 1956 and
Harr 1977). With this assumption, the hydraulic gradient in each channel is
(B-17)
B-16
EM 1110-2-1901
30 Sep 86
Figure B-13. Type VIII fragment (courtesy of McGraw-Hill

181
Book Company )
Integrating the ratio dy/(hd - y) from 0 to h generates the expression

for the quantity of flow as
(B-18)
(9) Type IX. This fragment type, shown in figure B-14, represents the
exit condition where the surface of seepage exists. The surface of seepage
(DE) is not an equipotential line or a streamline. Pavlovsky assumed that the
flow is horizontal. For the portion of the slope between D and E , the
flow is the coefficient of permeability times the integral of dy over
cot The flow for E to F is the permeability times the integral of
a2dy over the cot (a2 + h2 - y) . When the integration is performed, the
expression for the flow is
(B-19)
B-17
EM 1110-2-1901
30 Sep 86
Figure B-14. Type IX fragment (courtesy of McGraw-Hill

181
Book Company )
d. Exit Gradient. The method of fragments procedure can be used to

determine the exit gradient discussed in paragraph 4.9 of this manual. For
this procedure, the last fragment (downstream) needs to be either a type II or
a type III fragment. The exit gradient is defined as (Harr 1962)
(B-20)
where
h = head loss in the last fragment

m
K = complete elliptic integral of the first kind with modulus m
T = depth of flow region
As defined before, the modulus m is a function of both b/T and S/T and
is defined as
(B-21)
B-18
EM 1110-2-1901
30 Sep 86
Instead of calculating the various values, for type II fragments figure B-15
can be used with S/T to obtain the fraction for IES/hm . By substituting
the appropriate values, the exit gradient is calculated.
e. Example 1. The method of fragment procedure for confined flow will

be illustrated in the following example obtained for John H. Overton Lock and
Dam (U. S. Army Engineer District, St. Louis 1978). This problem will analyze
the steady state flow conditions for a two-dimensional idealization of the
lock structure. The quantity of flow and head along the bottom of the lock
will be determined. For illustrative purposes, the exit gradient procedure
will be included. The dimensions of the structure, shown in figure B-16, are
those used in the analysis after the cross section has been transformed to
account for soil anisotropy. The original analysis contained two soil layers,
but for illustrative purposes the soil will be modeled as one layer. The first
step is to determine the form factors for each region. The first region is a
type II fragment with S = 19 ft and T = 89 ft . Using figure B-7 with
S/T = 0.21 , the fraction for Q/kh = 0.78 , thus = 0.641 . Region 2 is a
type I fragment with L = 456 ft and a = 70 ft . From figure B-4, the form
factor for the type I fragment is equal to L/a , thus = 6.514 . The third
region is a type II fragment with S = 9 ft and T = 79 ft . Using figure B-7
with S/T = 0.114 , the fraction for Q/kh = 1.01 , thus = 0.495 . The
summation of the form factors is 7.650. The quantity of flow is calculated
-4
from equation B-3. Using transformed permeability of 400 x 10 cm/sec and a
total head of 18 ft, the quantity of flow is calculated to be 266.8 ft3/day/
foot of lock width. The head loss in each fragment is calculated from equa-
tion B-4. The following table lists the head loss for each fragment in this
problem:
Region hi
1 0.641 1.51
2 6.514 15.33
3 0.495 1.16
= 7.650 = 18.00
The head along the bottom decreases from 16.49 ft at the upstream end to 1.16
at the downstream end. Using the assumption of a linear distribution of the
head loss within a fragment, the head at any point along the bottom of the
lock could be calculated as
head at pt a = 16.49 ft - distance to pt A for upstream of lock

15.33 ft (B-22)
total length of lock
B-19
EM 1110-2-1901
30 Sep 86
Figure B-15. Relationship for determining the exit gradient as a function

of the head loss in the fragment and geometry for a Type II fragment
181
B-20
EM 1110-2-1901
30 Sep 86
Figure B-16. Transformed section of John H. Overton Lock simplified to one

112
soil layer (from U. S. Army Engineer District, St. Louis )
The exit gradient is calculated for region 3 which is a type II fragment.

Using S/T = 0.114 with figure B-15, the fraction for IES/hm is found to be
0.63. With a head loss of 1.16 ft in this fragment, the exit gradient is
calculated to be 0.082.
f. Example 2. This example will illustrate the method of fragment pro-

cedure for unconfined flow problems. The example is obtained from John H.
Overton Lock and Dam (U. S. Army Engineer District, St. Louis 1978). The
problem is to locate the free surface in the closure dam and to determine the
quantity of flow through the dam under steady state conditions. The dimen-
sions of the structure shown in figure B-17 are after the material has been
transformed to account for soil anisotropy. This sample problem assumes an
impervious boundary at the base of the dam. To account for some flow under
the dam, the impervious boundary could be lowered. By lowering the boundary
to the lowest possible point, bounds for the problem would be established.
There are three fragment types in this earthen embankment. Region 1 is a
B-21
EM 1110-2-1901
30 Sep 86
Figure B-17. Transformed section of John H. Overton closure dam (from

112
U. S. Army Engineer District, St. Louis )
type VIII fragment while region 2 is a type VII fragment and region 3 is a
type IX fragment. To calculate the flow through region 1, equation B-18 is
used with h = 32 ft; h d = 37 ft, and = 19.9 deg (cot = 2.76).
i
Substituting into the equation produces
(B-23)
For region 2, the quantity of flow is calculated from equation B-15.

Substituting into this equation produces
(B-24)
For region 3, equation B-19 defines the quantity of flow. By substituting

cot = 2 and H2 = 14 ft produces
(B-25)
B-22
EM 1110-2-1901
30 Sep 86
From the embankment geometry, L can be defined as
(B-26)
By substituting into equation B-26, there are four equations with four
unknowns, h , a2 , Q/k , and L . There are several methods to solve these
four equations (Harr 1977). For the case where h2 = 0 , a reduction of
two equations and two unknowns occurs. For this example, equation B-23 will
be combined with equation B-24 and equation B-26 will be substituted for L .
This produces
(B-27)
Also equation B-23 can be combined with equation B-25, producing
(B-28)
Equations B-27 and B-28 have reduced the equations and unknowns by two. Thus
with two equations and two unknowns, a trial and error graphical process can
be used. The results of this process are shown in figure B-18 and indicate
that h = 28.9 ft and a2 = 0.9 ft., Substituting into equations B-23 and B-25
generates a Q/k value of 1.71 which results in an estimated flow of
242.4 ft3/day/ft of dam. Knowing h and a2 , the location of the phreatic
surface can be estimated.
g. Flow in Layered Systems. One of the limitations of the method of

fragments is that the flow layer is assumed to be homogeneous and isotropic.
An approximate procedure to determine flow characteristics of a layered system
was proposed by Polubarinova-Kochina (1941). Harr (1977) extended this method
as follows. The coefficients of permeability for the two layers are related by
a dimensionless parameter by the expression
(B-29)
where
k 1 = coefficient of permeability of the upper layer
k 2 = coefficient of permeability of the lower layer
B-23
EM 1110-2-1901
30 Sep 86
Equation B-27
h a2
28.8 -0.4
29.0 2.1
30.0 8.6
31.0 13.6
Equation B-28
h a
2
29.9 0.6
27.8 1.2
25.6 1.7
Figure B-18. Graphical solution-of equations B-27 and B-28

112
(from U. S. Army Engineer District, St. Louis )
B-24
EM 1110-2-1901
30 Sep 86
The ratio of permeabilities can vary from 0 to infinity. Over this range,
ranges from 0 to 1/2. The basis of this method is to determine the flow and
head losses for three certain special cases of and then interpolate
between these values. The three special cases are as follows:
(1) = 0 . For to be equal to 0, k2 must equal 0. Therefore

the problem is reduced to a one-layer problem with a flow region thickness
equal to the upper layer.
(2) = 1/4 . For to be equal to 1/4, k2 must equal k1 .

Therefore, the problem is reduced to a one-layer problem with a flow region
the thickness of the upper and lower layers.
(3) = 1/2 . For to be equal to 1/2, k2 must be infinite. This

case represents the infinite flow where there is no resistance to flow in the
lower layer. Since Q/k1h = the inverse of this ratio is equal to zero.
This procedure can be expanded to a three-layer system by the use of two

values. The first value would be for the top two layers, while the second
would be for the bottom two layers.
h. Example 3. This example will illustrate the method of fragment pro-

cedure for confined flow in a two-layer system. The example is obtained from
John H. Overton Lock and Dam (U. S. Army Engineer District, St. Louis 1978).
This problem will analyze the steady-state flow conditions for the dam and
stilling basin. The quantity of flow and head along the bottom of the struc-
ture will be determined. For illustrative purposes, the exit gradient proce-
dure will be included. The effect of various parameters like the length of
sheetpile cutoff can be studied using this procedure. The dimensions of the
structure, shown in figure B-19, are those used in the analysis after the cross
section has been transformed to account for soil anisotropy. There are three
fragments for this problem and three cases to be evaluated. For the first
case, 0 , all the flow is assumed to occur in the clay layer. Region 1 is
a type II fragment with S = 25 ft and T = 35 ft. Using figure B-7 with
S/T = 0.71 , the fraction for Q/kh = 0.36 , thus = 1.38 . The second
region is a type V fragment with S = 13 ft, T = 23 ft, and L = 73.5 ft.
Since L > 25 , equation B-13 is used to calculate the form factor. For the
above values, the form factor is 3.73. Region 3 is a type II fragment where
S = 24 ft and T = 34 ft. The form factor, using figure B-7, is calculated to
be 1.36. Using equation B-1, the ratio Q/k1 is 2.78 and k1/Q is 0.36 .
For the second case, = 1/4 , the flow is assumed to be equal in both layers.
The form factors are recalculated using the same fragment types. The value of
Q/k1 is 6.50 which results in a k1/Q value of 0.15. For the last case,
= 1/2 , all flow is assumed to be in the lower sand layer. Only vertical
flow occurs in the top or clay layer. For this case, Q/k1 is infinite which
results in a k1/Q of 0 . A plot of k1/Q versus is shown in fig-
ure B-20a. For this problem k2 is 200 times k1 , therefore equals 0.48.
B-25
EM 1110-2-1901
30 Sep 86
Figure B-19. Transformed section of John B. Overton Dam and stilling

basin for one case of sheet pile lengths (from U. S. Army Engineer
112
District, St. Louis )
By interpolation for = 0.48 , k1/Q equals 0.01 which results in a flow Q

3
of 56.7 ft /day/ft of dam. To determine the head along the bottom of the
structure, the head at points A and B in figure B-19 must be determined. Using
the procedure described in example 1, equation B-22, the following head loss
and total head values are calculated.
Point A Point B
Head Loss Total Head Head Loss Total Head
ft ft ft ft
0 5.2 52.8 12.8 45.2
1/4 6.2 51.9 12.0 46.0
For the case where = 1/2 , the head anywhere along the bottom of the struc-
ture is equal to half the total head loss, or for this case 9 ft. Thus the
total head on points A and B is equal to 49 ft. Figure B-20b is the plot of
the total head versus and shows, for an of 0.48, the total head at
point A is 49.4 ft while the total head at point B is 48.5 ft. The exit gra-
dient for each case is calculated by the procedure described in example 1.
For the case of 0 , the fraction IES/hm is 0.55 which with S = 24 ft
B-26
EM 1110-2-1901
30 Sep 86
04
0.3
0
'~-0.2
0.1
0.5
E
IAI
€
1- 51 IC)
LL.
ci'
<(
50
~
...J 49
<(
1-
~ 48
47
46
45
0.5
€
18)
Figure B-20. value plots for John H. Overton Dam and stilling
basin (from U. S. Army Engineer District, St. Louis 112 )
B-27
EM 1110-2-1901
30 Sep 86
and hm = 3.8 ft produces IE = 0.087 . For the = 1/4 case, the fraction
I E S / h m is 0.615 which with S = 24 ft and hm = 5.0 ft produces
IE = 0.128 . For the case where = 1/2 , the head loss is the total head and
the distance is the thickness of the top layer. Using the equation
(B-30)
the exit gradient is 0.265. The exit gradient versus plot is shown in
figure B-20c. For an value of 0.48, the exit gradient is 0.245.
i. Uses and Limitations. The method of fragment procedure should be

used as a design tool where various factors are changed to evaluate their
effect or as an analytical tool when quick approximate results are needed.
When numerous factors are varied, the construction of flow nets becomes very
tedious and time consuming. The method of fragment procedure will generate
reasonable results for problems where the assumptions are not greatly violated.
There are several points the user needs to be aware of when using this proce-
dure. The flow region must be generalized so that it consists of horizontal
and vertical boundaries. The procedure models the actual flow paths within
the flow region, thus if there is any doubt as to the direction, a rough flow
net should be drawn. This becomes important when a small portion of a struc-
ture is modeled with several fragments because the flow could be modeled in
unnatural paths. The accuracy of the results is dependent upon how well the
fragment boundary actually represents vertical equipotential lines. The
greater the deviation, the greater the degree of error. However, for many
practical problems reasonable results are generated. Comparison of the method
of fragment results with finite element solutions for a one-layer system
showed that the quantity of flow values for the fragment procedure are within
8 percent of the finite element results, while the uplift values are within
38 percent of the finite element results. The Computer-Aided Structural Engi-
neering (CASE) project has developed a computer program for the method of
fragments procedure with a user manual describing the program (Pace et al.
1984).
B-6. Finite Difference Method.
a. Method of Solution. As previously mentioned in Chapter 4, the

finite difference method solves the Laplace equations by approximating them
with a set of linear algebraic expressions. The approximation is mathematical
rather than physical. The early methods of solving finite difference expres-
sions for Laplace's equation were based upon hand calculations by the relaxa-
tion method(1). However, more recently a wide range of finite difference
solutions suited to the digital computer have been developed. A description of
(1) For example, see Appendix A of EM 1110-2-2501.
B-28
EM 1110-2-1901
30 Sep 86
available methods used to solve finite difference problems including example

applications, case studies, and computer program listings is available (Rushton
and Redshaw 1979).
b. Advantages. Use of iterative techniques such as successive over-

relaxation which converge to the correct solution allows solution of uncon-
fined and transient flow problems, For simple problems, the finite difference
method is usually more economical than the finite element method (Rushton and
Redshaw 1979).
c. Disadvantages. The finite difference method is not suited to com-

plex geometry, including sloping layers and pockets of materials of varying
permeability. Irregular grids are difficult to input. Therefore, zones where
seepage gradients or velocities are high cannot be accurately modeled (Rushton
and Redshaw 1979).
d. Applications. The finite difference method was used at WES to simu-

late seepage conditions in streambanks induced by sudden drawdown of the river
level. As mentioned previously, this study included a viscous flow model,
field observations, and application of the finite element and finite differ-
ence methods. The results of the study indicated that the finite difference
method provided satisfactory and economical solutions for transient unconfined
fluid flow in porous, anisotropic, and nonhomogeneous media (Desai 1970 and
Desai 1973). The finite difference method was used to predict the location of
the phreatic surface within a zoned embankment with arbitrary fluctuations of
the reservoir (Dvinoff 1970). Generalized digital computer programs have been
developed which use the finite difference method to simulate one-, two-, and
three-dimensional nonsteady flow problems in heterogeneous aquifers under
water table and artesian conditions (Prickett and Lonnquist 1971 and Desai
1977). The finite difference method has been used to predict unsteady flow in
gravity wells. Good agreement was found between computed results and
laboratory test results obtained using a sand tank model (Desai 1977).
B-7. Finite Element Method.
a. Method of Solution. As previously mentioned in Chapter 4, the

finite element method is conceptually a physical rather than a mathematical
approximation. The flow region is subdivided into a number of elements and
permeabilities are specified for each element. Boundary conditions are speci-
fied in terms of heads and flow rates and a system of equations is solved to
compute gradients and velocities in each element (Desai and Abel 1972 and Desai
1977). Two- and three-dimensional finite element seepage computer programs for
both confined and unconfined flow problems have been developed at WES. Steady-
state and transient problems (that can be treated as a series of steady-state
problems) can be solved (Tracy 1973a; Tracy 1973b; and Hall, Tracy, and
Radhakrishnan 1975). An interactive graphics preprocessor is available to
generate the finite element grid (Tracy 1977a). It is possible to compute the
stream function and potential and plot contours of these values to obtain the
flow net (Christian 1980 and Christian 1980). Details concerning the selection
of spatial and time meshes, computer time required, convergence, and stability
are available (Desai 1977). Also, an interactive graphics postprocessor is
available to assist in the analyses of the finite element results (Tracy
B-29
EM 1110-2-1901
31 Jul 86
1977b). A listing of finite element seepage computer programs used within the
Corps is available (Edris and Vanadit-Ellis 1982).
b. Advantages. The finite element method is well suited to complex

geometry, including sloping layers and pockets of materials of varying perme-
ability. By varying the size of the elements, zones where seepage gradient or
velocity is high can be accurately modeled.
c. Disadvantages. The finite element method is usually more costly than

the finite difference method for simple problems (Rushton and Redshaw 1979).
d. Applications. The finite element method has been used in several

cases to provide solutions to seepage problems.
(1) WES studies. As discussed previously, the finite element method was
used at WES to simulate seepage conditions in streambanks induced by sudden
drawdown of the river level. This study included a viscous flow model, field
observations, and application of the finite difference and finite element
methods. The results of the study indicated that the finite element method
provided satisfactory solutions for transient unconfined fluid flow in porous,
anisotropic, and nonhomogeneous media (Desai 1970 and Desai 1973).
(2) Location of phreatic surface. The finite element method has been
used to determine the location of the phreatic surface in earth dams (Isaacs
1979; Isaacs 1980; Wei and Shieh 1979; and Desai and Kuppusamy 1980). The
finite element method was used to locate the phreatic surface within tailings
pond embankments and to define the subsurface flow of water from the pond.
Results predicted using the finite element model were confirmed with
measurements in a laboratory model and in the field (Kealy and Busch 1971).
(3) W.A.C. Bennett Dam. The finite element method was used to assess
the potential seepage flows and uplift pressures in the foundation rock for
W.A.C. Bennett Dam in British Columbia, Canada (see figure B-21). The finite
element analysis (see figure B-22) was carried out assuming the following
conditions:
(a) With an effective grout curtain.
(b) Without an effective grout curtain.
(c) With a drainage system.
(d) Without a drainage system.
(e) With various rock permeabilities.
The results of the finite element analysis, shown in figure B-23, indicate the
greatest reduction in seepage flow and hydrostatic pressure could be accom-
plished by an effective grout curtain and downstream-drainage system (Taylor
and Chow 1976).
B-30
EM 1110-2-1901
30 Sep 86
Figure B-21. Cross section of W.A.C. Bennett Dam, British Columbia,

269
Canada (courtesy of International Commission on Large Dams )
(4) Corps of Engineers levees. The finite element method was used by
the U. S. Army Engineer District, Rock Island, to study hydraulic sand fill
levees along the Mississippi River (Schwartz 1976). Finite element and
gradient plane (1) analyses were used in conjunction with data from a full scale
test levee to establish the material properties of the sand levees and to
determine the exit point of the free seepage surface, the quantity of through
seepage, and the exit gradients along the free discharge face. A parameter
study was performed and dimensionless design charts were developed.
(5) Reservoir loading conditions on zoned embankments. The use of the

finite element method to study the effect of initial filling of the reservoir,
steady seepage, and rapid drawdown of the reservoir on zoned embankments has
been given (Eisenstein 1979).
(6) Bureau of Reclamation dams. The Bureau of Reclamation has utilized

two- and three-dimensional finite element methods, electrical. analogy, and
mathematical methods to analyze seepage flow through a dam embankment and
(1) The gradient plane method is a graphical solution by means of the hodo-
graph (see description by Casagrande 1937).
B-31
EM 1110-2-1901
30 Sep 86
Figure B-22. Finite element study results (courtesy of International

269)
Commission on Large Dams
Figure B-23. Uplift pressure under various conditions (courtesy of

International Commission on Large Dams 269 )
B-32
EM 111O-2-1901
30 Sep 86
foundation (Mantei and Harris 1979). Narrows Dam, Colorado, on the South
Platte River, was analyzed for seepage at the feasibility stage. Because of a
pervious foundation, the planners called for a positive vertical cutoff by
constructing a slurry wall down to the underlying shale. However, near the
right abutment the shale drops away to depths too great for economical slurry
wall construction. A three-dimensional finite element model (see figure B-24)
was used to determine the vertical exit gradients at the downstream toe of the
dam. The finite element method was used to study the effect of a toe drain,
partial depth slurry trench, partially and fully penetrating relief wells (see
figure B-25). Calamus Dam, Nebraska, on the Calamus River, was also analyzed
by the Bureau of Reclamation for seepage at the feasibility stage (Mantei and
Harris 1979; Mantei, Esmiol, and Cobb 1980; Mantei and Cobb 1981; and Cobb
1984). Calamus Dam has a setting very similar to Narrows Dam in the sense that
it is an earth dam on a pervious foundation. However, the underlying shale at
Calamus Dam is at such a great depth that it cannot be used as the base for a
cutoff wall as it was for Narrows Dam. Early thinking on the project involved
the use of a slurry trench cutoff down to a pervious sandstone fromation. A
three-dimensional finite element model (see figure B-26) was used to determine
the effects of an embankment toe drain, slurry trench under upstream blanket,
and/or relief wells at the downstream toe of the dam on the seepage rates and
hydraulic gradients in the dam foundation. Time and expense in operating the
large three-dimensional finite element models made it necessary that priority
be given to studying the various design alternatives using the best estimate
of permeability for each foundation material rather than conducting sensitiv-
ity studies to establish the effect of varying the permeability (see para-
graph B-l). The three-dimensional finite element models were five elements
deep, with the bottom layer of elements representing the sandstone, the next
layer sand and gravel, recent alluvium, interbedded fine sand, and dune sand.
A detailed three-dimensional finite element model was made for the outlet
works area that defined more of the design details, such as the filter blanket
under the stilling basin and channel and water table elevation controls, to
study the effectiveness of relief wells around the stilling basin.
(7) Corps of Engineer dams. The finite element method was used by the
U. S. Army Engineer District, Huntington, (1) in a reanalysis of the underseep-
age at Bolivar Dam, Ohio, completed in 1938 on Sandy Creek (U. S. Army Engineer
District, Huntington 1977a). A two-dimensional finite element model (see fig-
ure B-27) was used to determine the effects of an embankment toe drain, up-
stream impervious blanket, and proposed relief wells on seepage quantities,
exit gradients, and uplift pressures. A sensitivity study was conducted using
the two-dimensional finite element model to determine the influence of various
pool and rock surface levels, the permeability ratio of foundation soils, the
existence of a downstream gravel layer, and the effective source of seepage
entry upon underseepage. Typical test results for one set of boundary condi-
tions and permeability values are given in figure B-28. Additional applica-
tions of the two-dimensional finite element method to conduct sensitivity
analysis to assess the effect of permeability anisotropy and various seepage
control measures was given by Lefebvre and coworkers (Lefebvre, Part, and
(1) Work was performed by Soil Testing Services, Inc., Northbrook, Illinois.
B-33
EM 1110-2-1901
30 Sep 86
~xis of Dam
I
f§j ~J~I~1~~~~~~~ff-J *I
SECTION
...-Assume oarallel flow line no flow

/ ,' I
'i !
,,:
j
I I I
: ; .: I
; I:
! l I :1 0
0
I ; j il N
(
0
:
~!
v.,
I
~l ~st\.
I
I I'
G
... I I ~;
-;;
i0 '
I
' I
I
.............
....End o kev rend
and cuto f
\ I\ 1\ g ~
Cl
\ \I ~oiiiiiiiF
-
u \
0
...0 ': ~~
e
•
~
', 0
...
I»
c: ...... I Ui
0 .... ,.. I 0.
:l
0 -o0
• I ~
0
v ..c:
I
·o...
Toe of dcim-........ 0
0
... - ./
I
C\1
.,...
Cl >
~
i I II ...
~
.,.
E I Q)
~
I .,.,E
:l
I
I
II 11
~
r
PLAN
Figure B-24. Contours of exit gradient from three-dimensional finite

element model study of Narrows Dam, Colorado (courtesy of American
217
B-34
EM 1110-2-1901
30 Sep 86
~~----~------~------~------~----~
~~----~-------+-------r------~----~
t-
z ~~----~-------+-------r------~----~
LLJ
0
<l: 0
o--- ~ .... -. -~--.
a::
(!)
t:
X
LLJ
_J <.0
~ 0~1----tr--------+-------~----~------~
(..)
t- ~
a:: 6~~--~-------+------~------~----~
~ ~---L- -
0
6~----~------~------~------~----~
-300·0 0·0 300·0 600·0 900-<J 1200-Q
DISTANCE BEYOND THE. END OF SLURRY TRENCH-FEET
LEGEND
D 3-0 MODEL WITHOUT REUEF WELLS
0 2-o MODEL W\THOUT RELIEF WELLS
~ QUASI 3-0 MODEL WtTH RELIEF WELLS
AT 100' SPACING
Figure B-25. Vertical exit gradients from three-dimensional finite

element model study of Narrows Dam, Colorado (courtesy of American
217
B-35
EM 1110-2-1901
30 Sep 86
Figure B-26. Grid for three-dimensional finite element model

study of Calamus Dam, Nebraska (courtesy of American Society
217
of Civil Engineers )
B-36
llf. . .TION
I
•• ,
I
tJ
~
~::_.-;,_
I ~
- - c=: -- ----- - - -- -- - ----
_100
' I !
-- e---t-=1 . I
I
I
I-
-- ...
_
+=-+-
L ____
i
I
--t------
--
_•oo
B-37
I. ll.f.VIlTIONI lllnlll TO -Nt K.A Uvtl DATW
SCAL£ tFtET)
Figure B-27. Grid for two-dimensional finite element model ·study of Bolivar Dam, Ohio
EM 1110-2-1901
88
30 Sep 86
30 Sep 86
EM 1110-2-1901
MODEl PREDICTIONS
-----l
I f (;:rilrotfTRlt }..~'-LL$ !'l 1
1
1,/JI\Nlll'i (f SC(f-'AG(. Am-t~'r..A..
[Uit-i.I ~"a~:.r:AUr'l.'j 1...{,- C!UJ'C"'
1 9
~ ~-~:-:t~,t.fl":(~2~~:r"'t:l! ~ I~ r.. '' (.4~ !It AI~~ I 1 !til f{'J lt.-o I
i~-.:._--: --:;-i-:-~-~J::Ji~---=i_:Ol_~fl
t 1 _ _ _ ~ _ _L _
-==-t;:::-t~
J _ _ t __ ~ :~ _~~ ~ ~:-
_ _jt-- __1:: _J
J
1 1fCLIJOl, fii[L.f .. ll~ rL(N • . , nr~-
~- .
,~ ~
~
!""
~
.!'0
B-38
.!!!!!.
'!"____ 'I""
NOIIIIOIIITAL DltT4f'Ct
~~
UI'AT. . . "'Ill
to . . . . . . UVIL DllfUM
-
....,.·-flle
__ __.... ,.... ,...
"'--AIUT"I' •m ••• , •• a.a, . . .,..
-.uwtD ---~
. .. fill. . . . ..,..
. . . . . . IITOTOlii'MI
......,..,... . ,. . . . . . . . . . ."' . . . Dill

...
SCALE Cf[£T)
Figure B-28. Results from two-dimensional finite element model study of Bolivar Dam, Ohio
88
(from U. s. Army Engineer District, Huntington )
EM 1110-2-1901
30 Sep 86
Tournier 1981). The finite element method was used by the U. S. Army Engineer
District, Huntington, (1) in a reanalysis of the underseepage at Mohawk Dam,
Ohio, completed in 1937 on the Walhonding River (U. S. Army Engineer District,
Huntington 1979b). A three-dimensional finite element model (see figure B-29)
was used to study the cause of unusually high relief well flows. Typical test
results for one set of boundary conditions are given in figure B-30.
(1) Work was performed by Soil Testing Services, Inc., Northbrook, Illinois.
B-39
EM 1110-2-1901
30 Sep 86
- IL ...
1\. •••
l '/
...'"
Ito
,'
}-LI,--~+------~---~~~=~~:::_~,~~~---~-~------~---~---
I
I ' ' ' •
.'
.
/1'
I '
I I ' ..,....... j
l
,/ ' . t
. ; 3i 0 g 8 8
! !
0
~ ~ .• .: .• ! 3
. .. ;
:,; . ! .
0 0
~ ~.; .• ;
g 0
•• !
0
i : ; 0 !l.l'l~ ...,.. . .ll't!.
• •
..
~ ~ ~ ~ ~ •I ,; ,;
.
0
"
I
- ·".
I I I __ ___, CDOC . . . Afl
"'I
I I I I I I I
'
I
t
B-40
-·-
1-- ----·- --~t.c::!!.:........l!~~-............ ...J!:.-~~ ... •t;'?!IIT' pw ....-roa.__

IU..c;( l,.'•..c.t
(!) ' 01'•.·-· ..
-.· 1\f.
'
).:.' I
..... --··•c. wa~~

.... ........
!!U!:!!>
....
nwptt
... !!
LJIIW!I
!!ol. (
"!!!!I.!!!I.'Y
!!.. !! ""!!!!Ig_
......
£2!!Sl
,
<D
..........
\! ................. ,~\..
.
(i) ..............~
. ....
(i)
w ......
JCU'CUittaOUh...l"
~-~ tcl'I.UI
•• '""'
.... .___•••u•·SCAU
..., .. ..c-·•••
.
.
..
.......:, :J) •:..tU •'-·
""'
. ·-
. _ , ,_ _ ~k~
... -.a.u
,..
...""
f) ~,.._
(fl .......,~, ......
""'"'" 0 ....,. .......r "
_,.VICIUS ••
....
@ ..._.. OCDOI 0-
JfiUIIIOI''i(loOUI~
" ......... ..,.,.

(£).,.....,...~
0001 o• u
E'~::L~
Figure B-29. Grid for three-dimensional finite element model study of Mohawk Dam, Ohio
89
OIIT~UO lt&ll
I l l fiiiiT
SEC 88 HORIZONTAL PLANE THROUGH TOP OF FOUNDATION OUTWASH

""'OJ(CTtO fiii~IIAIIIICl. U"1Gl>4 • 1~00' IAS£0 ON Pt[ZOW(T£111-
Olr.>f:ltVU•OIIIS IN llllMC" ltl'' .,.
: ::~ ~ I I
I
I
zo4tn•l
-i
I
II
~
I
I
I
I
IliUM
- I
I 1;
I
1/ ~
~
~
I
'
' ~
I
i
!II
- .
Cf ..
I
I I I I I
I
i
I I
SU,...GI1\
(Ni'Hi
I
:= ~
i
~ •.,..c,.. j
;- i
:
i ·I
111
:
I
I
I
I
: I: I I: I . ~~I I=
I•
I• i! ~' 1- • i• .•• " ..
!j =~ jl if j if ~ -
al f-
! I!
I
i
i' i I
;
' '
i~l ~('
j~ ~~~
! I I I i
I I ; r~· :, I II I I
I
I I
i
i ·~ ~·
~- ! : I
!
I'
i
!
I
!
I
i
B-41
SEC CC HORLZONTAL PLANE THROUGH BOTTOM OF FOUNDATION OUTWASH
1T \ \
I
I
i
I t: ~I
I It~~~ II : •
-
ta4 trw
- .
.,. c, ..
.
I
I
I
.
I
I
I
i
I
I
I
I
I
I
·~·I
MI ~
. I""'-~··
I
I
i
'
Il
.=
i I
.
d L! <~ H J
:- r i" r- g 'i
1 2
\
~~:1i\
1'\ - II
i
I! I
I.
!.. ". : \
~--
1. . . " .
~--
I
. 1
j
Ir- ..
m
: I
!~ l I 'l/ I ~~~~/:
:
';·~ ~
·~·I
I I
-
t •
I l l " .. I
! I i I
I i
IM\[biCTtD !iTUD' STAT( "'ttttCTUI fTt:A01' ST.Ifl

N:.ZOWT£1t LIVUS r [~if' w(:,.:., I)ISC14Ait0( X:JW
•·•o
1"·20
·-·
•·t
TOTIL WE.L.l DISC:>IMG( •
llV C:I'"M
7i"i"Cni"
EM 1110-2-1901
110 5
1100
30 Sep 86
Figure B-30. Results from three-dimensional finite element model study of Mohawk Dam, Ohio
89
EM 1110-2-1901
30 Sep 86
APPENDIX C
ANALYSIS OF PRESSURE INJECTION TESTS (Ziegler 1976)
C-1. Water Pressure Tests
a. Water pressure tests are conducted by pumping water into a borehole

at a constant pressure and measuring the flow rate. Water enters the rock mass
along the entire length of borehole or along a test section (i.e., borehole
interval) sealed off by one or more packers as shown in figure C-1. The test
is rapid and simple to conduct and by conducting tests within intervals along
the entire length of borehole, a permeability profile can be obtained.
b. In most pressure tests the water injection pressure is limited to a

value which is not expected to produce an increase in the fracture width. An
increase in the fracture width will cause erroneously high flow rates result-
ing in higher permeabilities than actually exist. A common criterion is to
limit the water injection pressure to 1 psi/ft of borehole depth above the
water table and 0.57 psi/ft of borehole depth below the water table. This
criterion results in a maximum injection pressure less than the effective over-
burden pressure if the overburden has a unit weight greater than 144 lb/ft3.
c. The coefficient of permeability, based upon laminar flow, is com-

puted for a vertical test section with the following assumptions:
(1) Medium is homogeneous and isotropic.
(2) Laminar flow governed by Darcy's law.
(3) Radial flow from a cylindrical and vertical borehole test section
length, Radial flow implies that the equipotential surfaces form
cylinders symmetrical about the axis of the borehole test section.
(4) No inertia effects.
(5) Boundary conditions (fig. C-2).
(a) At r = ro . H = Ho .
(b) At r=R , H=0.
Where
r = radial distance from the test section (L)
r o = radius of borehole (L)
H = excess pressure head (L)
H o = excess pressure head at the center of the test section (L) (either
measured within the test section, or computed allowing for fric-
tional head losses within the drill pipe)
C-l
EM 1110-2-1901
30 Sep 86
R = radius of influence: radial distance from the test section
{WATER FLO*)
corresponding to a 100 percent loss in excess head, Ho (L)
GROUND SURFACE
~ z
---
lLJ
...J
~j:: PACKER
lJ
lLJ
""I-
V,
lLJ
1-
/
"-- \ "'""
t
\.:)'
z
'"-I---PACKER ) ..
lLJ
...J
z0
i:
\
lJ
lLJ
""
1-
""lLJ1-
CLOSED
/ ) ~
a. DOUBLE PACKER SETUP b. SINGLE PACKER SETUP
Figure C-1. Typical water pressure test setups (prepared by WES)
C-2
EM 1110-2-1901
30 Sep 86
Figure C-2. Homogeneous isotropic material--radial flow

(prepared by WES)
d. Darcy's equation may be written
where
v = flow velocity (L/T)
k = laminar equivalent permeability (L/T)

e
dh/dr = hydraulic gradient in the radial flow system (L/L)
e. The volume flow rate from the borehole cavity is
where
Q = volume flow rate (L3/T)
C-3
EM 1110-2-1901
30 Sep 86
2
A = area of an equipotential surface (L )
The area A in the radial system is
Thus
which can be written
and integrated
(C-1)
The negative sign can be dropped by choosing flow away from the borehole as
positive, thus
and the permeability is
C-4
EM 1110-2-1901
30 Sep 86
(C-2)
f. Before equation C-2 can be applied, however, the radius of influ-

ence, R , must be determined or estimated. A rough estimate of R is nor-
mally adequate since large variations in R generally produce only small
variations in the computed permeability. In equation C-2 the permeability is
directly related to the term 1n (R/r0) . The rate of change of 1n (R/ro) ,
and thus permeability, decreases rapidly as R increases. This is illus-
trated by considering the analysis of a pressure test conducted in an NX-size
borehole. The increase in permeability, ke (based on equation C-2), as R
is increased from 1 ft to 106 ft is shown below:
Equivalent Permeability,
Radius of Influence,
k e , ft/sec
R , ft
1
ke
10 2.1 k e
100 3.2 ke
1,000 4.3 k e
10,000 5.4 k e
100,000 6.5 ke
1,000,000 7.5 k e
Thus as R is increased from 1 ft to 106 ft, the computed permeability

increases by less than one order of magnitude. To eliminate the arbitrary
choice of R as a source of error in the permeability calculation, it has
been suggested that during a pressure test, pressure in the surrounding mass
be observed in a nearby borehole. By changing the upper limits of integration
in equation C-1 to correspond to an arbitrary distance r , within the
boundaries ro to R (fig. C-2) the following equations are developed:
Integration yields
C-5
EM 1110-2-1901
30 Sep 86
where H= excess pressure head at an arbitrary distance, rl , from the test

1
section (L) which can be written as
(C-3)
g. The coefficient of permeability, based upon nonlinear or turbulent

flow, is computed for a vertical test section in a homogeneous, isotropic
medium with the following assumptions:
(1) Medium is homogeneous and isotropic.
(2) Nonlinear or turbulent flow governed by the Missbach law
(3) Radial flow from cylindrical and vertical borehole test section of
length,
(a) At r = ro , H = Ho .
(b) At r=R, H=0.
h. The Missbach law may be written
where
m = degree of nonlinearity (generally between 1 and 2)

m
k' = turbulent equivalent permeability (L/T)
e
Volume flow rate for radial flow from the cavity is
C-6
EM 1110-2-1901
30 Sep 86
and it follows that:
Drop the negative sign by choosing flow away from the borehole as positive,
yielding
(C-4)
Equation C-4 can be rewritten as
(C-5)
C-7
EM 1110-2-1901
30 Sep 86
where
The logarithms of terms in equation C-5 yield the equation of a straight line:
log Ho = log E1 + m log Q
A plot of log Ho versus log Q will be a straight line with an arithmetic

slope equal to the degree of nonlinearity m .
i. The coefficient of permeability for laminar flow in fissures is

computed for a vertical test section with the following assumptions:
(1) Vertical borehole test section of length is intersected by an

arbitrary number of horizontal fissures.
(2) Fissures are constant aperture openings between smooth parallel

plates.
(3) Radial flow occurs within each fissure and is governed by Darcy's
law. No flow occurs in material between fissures.
(4) Each fissure has the same equivalent parallel plate aperture, e .
(a) At r = ro , H=Ho .
(b) At r = R , H=0.
j. The derivation of equations C-6 and C-7 proceeds in the same fashion
as that given above for laminar flow through a homogeneous isotropic medium
(equations C-2 and C-3). For the fissured medium the test section length,
is replaced by the quantity (ne) where n is the number of fissures
intersecting the test section and e is the equivalent parallel plate
aperture of each fissure. The resulting expression for the flow rate Q is
C-8
EM 1110-2-1901
30 Sep 86
Figure C-3. Fissured-medium--radial flow (prepared by WES)
and the permeability of each fissure is
(C-6)
where kj = laminar fissure permeability (L/T). The general expression for

excess pressure head H1 at an arbitrary distance r1 between the boundaries
ro and R is given by
which can be written as
(C-7)
C-9
EM 1110-2-1901
30 Sep 86
From the theory of viscous flow between smooth parallel plates
(C-8)
where
Equation C-8 is substituted into equation C-6 to solve for the equivalent
parallel plate aperture, e :
k. The coefficient of permeability for nonlinear or turbulent flow in

fissures is computed for a vertical test section with the following
assumptions:
(1) Vertical borehole test section of length intersected by an

arbitrary number of horizontal fissures.
(2) Fissures are constant aperture openings between smooth parallel

plates.
(3) Radial flow occurs within each fissure and is governed by the
Missbach law (vm = k'i). No flows occur in material between fissures.
j
(4) Each fissure has the same equivalent parallel plate aperture, e .
(5) No inertial effects.
(a) At r = ro , H = Ho .
(b) At r=R, H=0.
1. The derivation of equation C-9 proceeds in the same fashion as that

given above for nonlinear flow through a homogeneous isotropic medium (equa-
tion C-4). For the fissured medium, the test section length, , is replaced
C-10
EM 1110-2-1901
30 Sep 86
by the quantity ne where n is the number of fissures intersecting the test

section and e is the equivalent parallel plate aperture of each fissure.
The resulting equation for the turbulent fissure permeability is
(C-9)
where k' = turbulent fissure permeability (L/T)m . Equation C-9 can be

j
rewritten as
(C-10)
where
The logarithms of terms in equation C-10 yielded the equation of a straight

line:
A plot of log Ho versus log Q will be a straight line with an arithmetic

slope equal to the degree of nonlinearity, m .
C-2. Air Pressure Tests.
a. Air pressure tests are conducted similarly to water pressure tests,

with the essential difference being the replacement of water with air. The
use of air, however, requires that permeability equations be modified for
application to a compressible fluid and a conversion must be made from air to
water permeability. To compute the permeability, flow is assumed to be
laminar and governed by Darcy's law. The material tested is assumed to be a
homogeneous isotropic porous medium. Darcy's law can be written
v = ki
C-11
EM 1110-2-1901
30 Sep 86
where
v = flow velocity (L/T)
i = hydraulic gradient (L/L)
b. The coefficient of permeability, k , is dependent on properties of

the medium and fluid or gas. Since the test involves the flow of air and
permeability relating to the flow of water is desired, it is convenient to
solve for the intrinsic permeability, k , which is characteristic of the
medium alone. The coefficient of permeability and intrinsic permeability are
interrelated (Davis and Dewiest 1966, and Muskat 1946). For the flow of water
(C-11)
where

e
= unit weight of water (F/L3)
= dynamic viscosity of water
c. The conversion of intrinsic permeability measured by an air test to

the water permeability can be in error due to differences in gas and fluid
flow phenomena i.e., the Klinkenberg Effect (Weeks 1978), or alteration of the
material's physical properties caused by a chemical reaction with the fluid or
gas (Davis and Dewiest 1966, and Lynch 1962). In sediments rich in clay
minerals, water permeability calculated from air measurements may be overesti-
mated by a factor of 100 (Davis and Dewiest 1966). This overestimation is
caused by a hydration of clays during waterflow which does not occur during
airflow. Test results should be applied with caution as more experience and
studies are needed to determine the limitations of equation C-11.
d. Muskat (1946, p. 679) observed that the flow of incompressible

liquids is easily modified for application to the flow of a compressible gas.
Steady-state solutions for analyzing constant water pressure test results are
generally written in the form
C-12
EM 1110-2-1901
30 Sep 86
where
Q = volume flow rate (L3/T)
= length of test section (L)
H o = excess pressure head at the center of the test section (L)
The term [C] is dependent on assumed flow and boundary conditions. The follow-
ing equation is commonly used in analyzing water pressure test results.
(C-12)
where
R = radius of influence (L)
ro = radius of borehole (L)
After solving for Q and converting to intrinsic permeability, k ,

equation C-12 becomes
(C-13)
where Po = excess pressure at the center of test section (F/L2) the

equation (C-13) is based-on the assumption that:
(1) The medium is a homogeneous and isotropic porous continuum.
(2) The flow emitting from a cylindrical section of borehole is

laminar, and governed by Darcy's equation.
(3) The flow pattern is ellipsoidal and symmetrical about the axis of
the borehole test section.
e. Equation C-13 can be modified for analyzing air pressure test

results by making similar assumptions and considering the fluid to be
compressible. Before modification, equation C-13 must be written in terms of
absolute pressures. The pressure, Po , equals the absolute pressure in the
test section, minus atmospheric pressure, Equation C-13 becomes
C-13
EM 1110-2-1901
30 Sep 86
(C-14)
f. Equation C-14 is modified to apply to the flow of gas by replacing

the pressure term by the expression
(C-15)
where is a constant at any point in the flow and defined as
(C-16)
where
= mass density (F-T2/L)/L3

2
g = acceleration due to gravity (L/T )
= absolute pressure (F/L2)
= unit weight of gas (F/L3)
and the exponent, M , determines the thermodynamic nature of the expansion of

a gas as it moves from high- to low-pressure regions. The resulting equation
(equation C-18 below) is in terms of weight flow rate rather than volume flow
rate. Since gases are highly compressible, the volume flow rate will vary
with pressure and temperature along the flow path. However, the weight flow
rate can be assumed to remain constant. Weight flow rate, QWF , and volume
flow rate, Q , at any point in the flow are related by
(C-17)
Substitution of expression C-15 into equation C-14 (replace µw with µa)

yields the following expression for the weight flow rate of air:
C-14
EM 1110-2-1901
30 Sep 86
(C-18)
where
Q W F = weight flow rate (F/T)
µa = dynamic viscosity of air (F T/L2)
g. For convenience, the value of M is assumed equal to 1 which cor-

responds to isothermal expansion of an ideal gas as it moves from the borehole
through the medium (for adiabatic expansion, M < 1). Since equation C-16 is
valid at any point in the flow and M + 1 , can be replaced by
where and are the unit weight of air and absolute pressure in the
test section, respectively. By substituting M = 1 and in
equation C-18, the intrinsic permeability, k , can be expressed as
(C-19)
h. Parameters used in equation C-19 are measured or computed from the

test data. The test section length, , and radius of test section, ro ,
are determined from the test setup. Standard atmospheric pressure,
2
equals 14.7 psi or 2,120 lbF/ft . The absolute pressure in the test section,
is equal to the gage pressure measured in the test section, PO , plus
atmospheric pressure, The dynamic viscosity of air, µa , depends only
on the temperature, and its variation is small over a large range of tempera-
-5
tures. For example, air viscosity increases from approximately 0.035 x 10
-5
to 0.045 x 10 lbF-sec/ft2 as temperature increases from 0 to 250° F. In
general, the dynamic viscosity of air can be assumed to equal 0.38 x
10-5 lbF-sec/ft2 which is the viscosity of air at 68° F.
i. The weight flow rate from the test section, QWF , is constant along
the flow path and can be determined at the manifold by applying equation C-17:
C-15
EM 1110-2-1901
30 Sep 86
where
Q W F = weight flow rate (F/T)
Qm = volume flow rate at the manifold (L3/T)
= unit weight of air at the manifold (F/L3)
The value of is related to the pressure and temperature by the equation

of state (Vennard 1965):
where
= unit weight of air at the manifold (lbF/ft3)
absolute pressure at the manifold (lbF/ft2)
R g = engineering gas constant for air (53.3 ft-lbF/lbF-degrees Rankine)
= absolute temperature at the manifold (degrees Rankine)
j. The unit weight of air in the test section, , is also computed

from the equation of state:
where = absolute temperature in test section (degrees Rankine). All the

parameters needed to compute the intrinsic permeability, k , by equation C-19
have now been determined. The equivalent permeability applicable to the flow
of water, ke , is computed by substitution of k into equation C-11 yielding
the following general equation:
C-16
EM 1110-2-1901
30 Sep 86
where

e
Qm = volume flow rate at the manifold (L3/T)
= length of the test section (L)
= unit weight of water (F/L3)
µa = dynamic viscosity of air
µ w = dynamic viscosity of water
= absolute temperature in test section (degrees Rankins)
= absolute pressure at the manifold (F/L2)
= absolute temperature at the manifold (degrees Rankins)
= absolute pressure in test section (F/L2)
= atmospheric pressure (F/L2)
R = radius of influence (L)
ro = radius of the borehole (L)
C-3. Pressure Holding Tests.
a. The pressure holding test, sometimes called a pressure duration or

pressure drop test, is conducted by pressurizing the test section to a known
value, then stopping the water supply and observing the rate of pressure
decay. A pressure holding test is usually conducted in conjunction with a
water pressure test. For the conduct of the test, a pressure transducer
mounted in the test section is used to provide a continuous record of pressure
versus time.
b. Tests in fissured rock can be analyzed through application of the

parallel plate analogy. The laminar fissure permeability, kj , is related to
the pressure drop data by the expression
C-17
EM 1110-2-1901
30 Sep 86
(C-20)
where
ro = radius of borehole (L)
= excess pressure head at the center of the test section at the

initiation of a pressure drop test (L)
= excess pressure head at the center of the test section at the
completion of a pressure drop test (L)
R = radius of influence (estimated) (L)
n = number of fissures intersecting the test section
t d r o p = duration of the pressure drop test (T)
c. Equation C-20 was derived by Maini (1971) based on the following

assumptions.
(1) Radial flow occurs from a vertical test section and is governed by
Darcy's law.
(2) Fissure system intersecting the test section is represented by hori-

zontal parallel plates of equal aperture and spacing.
(3) Test zone is saturated. Maini (1971) suggests that before conduct-
ing tests in zones above the ground-water table, water be pumped into the bore-
hole test section to saturate the fissure system in the immediate vicinity of
the borehole.
d. The equivalent parallel aperture, e , in equation C-20 is unknown;

however, by parallel plate analogy
(C-21)
which when substituted into equation C-20 yields
(C-22)
C-18
EM 1110-2-1901
30 Sep 86
The aperture, e , is computed by equation C-22 and substituted into equa-

tion C-21 to obtain the permeability kj . If the test zone is modeled as
a homogeneous isotropic porous medium and assumptions (1) and (3) above are
valid, an expression for the laminar equivalent permeability, ke , is
obtained by replacing the quantity ne in equation C-20 with the length of the
test section, , to yield
(C-23)
e. The pressure drop test is a suitable supplement to the water pres-

sure test and possesses certain advantages. The pressure drop test is likely
to require a significantly smaller volume of water than that needed in a
constant pressure test. The savings in water can be important when conducting
tests in regions with a limited water supply. The measurement of pressure in
the test section by an electric transducer allows the pressure drop test to be
conducted with low initial pressures regardless of the depth of the test
section, since the water level in the flow pipe only needs to be above the top
of the test section to initiate a test. The use of low initial pressures has
the additional advantage of reducing friction pressure losses during the con-
duct of a test.
C-19
EM 1110-2-1901
30 Apr 93
Change 1
APPENDIX D
FILTER DESIGN
D-1. General. The objective of filters and drains used as seepage control
measures for embankments is to efficiently control the movement of water
within and about the embankment. In order to meet this objective, filters and
drains must, for the project life and with minimum maintenance, retain the
protected materials, allow relatively free movement of water, and have suffi-
cient discharge capacity. EoLdesign,. these. three. necessLties are termed,
respectively, piping or stability requirement, permeability requirement, and
discharge capacity. This appendix will explain how these requirements are met
for cohesionless and cohesive materials, and provide general construction
guidance for installation of filters and drains. The terms filters and drains
are sometimes used interchangeably. Some definitions classify filters and
drains by function. In this case, filters must retain the protected soil and
have a permeability greater than the protected soil but do not need to have a
particular flow or drainage capacity since flow will be perpendicular to the
interface between the protected soil and filter. Drains, however, while meet-
ing the requirements of filters, must have an adequate discharge capacity
since drains collect seepage and conduct it to a discharge point or area. In
practice, the critical element is not definition, but recognition, by the
designer, when a drain must collect and conduct water. In this case the drain
must be properly designed for the expected flows. Where ic is not possible to
meet the criteria of this appendix, che design must be cauciously done and
based on carefully controlled laboratory filter tests (Perry 1985).
D-2. Stability. Filters and drains< 1 l allow seepage to move out of a pro-
tected soil more quickly than the seepage moves within the protected soil.
Thus, the filter material must be more open and have a larger grain size than
the protected soil. Seepage from the finer soil to the filter can cause move-
ment of the finer soil particles from the protected soil into and through the
filter. This movement will endanger the embankment. <2 > Destruction of the
protected soil structure may occur due to the loss of material. Also, clog-
ging of the filter may occur causing loss of the filter's ability to remove
(1)
In paragraphs D-2 and D-3 the criteria apply to drains and filters; for
brevity, only the word filter will be used.
(2)
In practice, it is normal for a small amount of protected soil to move
into the filter upon initiation of seepage. This action should quickly
stop and may not be observed when seepage first occurs. This is one
reason that initial operation of embankment seepage control measures
should be closely observed by qualified personnel.
D-1
EM 1110-2-1901
30 Apr 93
Change 1
water from the protected soil. Criteria developed by many years of experience
are used to design filters and drains which will prevent the movement of pro-
tected soil into the filter. This criterion, called piping or stability cri-
terion, is based on the grain·size relationship between the protected soil and
the filter. In the following, the small character "d" is used to represent
the grain size for the protected (or base) material and the large character
"D" the grain size for the filter material.
Determine filter gradation limits using the following steps:( 3 l
1. Determine the gradation curve (grain-size distribution) of the base soil

material. Use enough samples to define the range of grain sizes for the base
soil or soils and design the filter gradation based on the base soil that
requires the smallest D15 size.
2. Proceed to step 4 if the base soil contains no gravel (material larger

than No. 4 sieve).
3. Prepare adjusted gradation curves for base soils with particles larger
than the No. 4 (4.75 mm) sieve.
a. Obtain a correction factor by dividing 100 by the percent passing

the No. 4 (4.75 mm) sieve.
b. Multiply the percentage passing each sieve size of the base soil
smaller than No. 4 (4.75 mm) by the correction factor from step 3a.
c. Plot these adjusted percentages to obtain a new gradation curve.
d. Use the adjusted curve to determine the percent passing the No. 200
(0.075 mm) sieve in step 4.
4. Place the base soil in a category based on the percent passing the No.
200 (0.075 mm) sieve in accordance with Table D-1.
5. Determine the maximum D15 size for the filter in accordance with
Table D-2. Note that the maximum D15 is not required to be smaller than
0.20 nun.
6. To ensure sufficient permeability, set the m1n1mum D15 greater than or

equal to 3 to 5 x maximum d 15 of the base soil before regrading but no less
than 0.1 mm.
( 3)
Guide for Determining the Gradation of Sand and Gravel Filters, Soil
Mechanics Note No. 1, U.S. Department of Agriculture Soil Conservation
Services, Engineering Division, Jan 1986.
D-2
EM 1110-2-1901
30 Apr 93
Change 1
Table D-1. Categories of Base Soil Materials
Percent Finer Than the No. 200

Category (0.075 mm) Sieve
1 >85
2 40-85
3 15-39
4 <15
7. Set the maximum particle size at 3 in. (75 mm) and the maximum passing the
No. 200 (0.075 mm) sieve at 5 percent. The portion of the filter material
passing the No. 40 (0.425 mm) sieve must have a plasticity index (PI) of zero
when tested in accordance with EM 1110-2-1906, "Laboratory Soils Testing."
8. Design the filter limits within the maximum and minimum values determined
in steps 5, 6, and 7. Standard gradations may be used if desired. Plot the
limit values, and connect all the minimum and maximum points with straight
lines. To minimize segregation and related effects, filters should have rela-
tively uniform grain-size distribution curves. without "gap grading"--sharp
breaks in curvature indicating absence of certain particle sizes. This may
require setting limits that reduce the broadness of filters within the maximum
and minimum values determined. Sand filters with D90 less than about 20 mm
generally do not need limitations on filter broadness to prevent segregation.
For coarser filters and gravel zones that serve both as filters and drains,
the ratio D90 /D 10 should decrease rapidly with increasing D10 size. The limits
in Table D-3 are suggested for preventing segregation during construction of
these coarser filters.
D-3
I I
'EM 1110-2-1901
30 Apr 93
Change 1
Table D-2. Criteria for Filters
Base Base Soil Description, and

Soil Percent Finer Than No. 200
Category (0.075 mm) Sieve (a) Filter Criteria (b)
1 Fine silts and clays; (c) D 15 :,; 9 X da 5

more than 85% finer
2 Sands, silts, clays, and D15 :,; 0. 7 mm

silty and clayey sands;
40 to 85% finer
3 Silty and clayey sands (d), (e) D15 <

-
r40-A)
L4o rs
and gravels;
15 to 39% finer [ ( 4 X d 85 ) - 0 . 7 mm ] + 0 . 7 mm
4 Sands and gravels; less (f) D15 :,; 4 to 5 x d 85

than 15% finer
(a) Category designation for soil containing particles larger than 4.75 mm is
determined from a gradation curve of the base soil which has been adjusted
to 100% passing the No. 4 (4.75 mm) sieve,
(b) Filters are to have a maximum particle size of 3 in. (75 mm) and a maximum
of 5% passing the No. 200 (0.075 mm) sieve with the plasticity index (PI)
of the fines equal to zero. PI is determined on the material passing the
No. 40 (0.425 mm) sieve in accordance with EM 1110-2-1906, "Laboratory
Soils Testing." To ensure sufficient permeability, filters are to have a
D15 size equal to or greater than 4 X d 15 but no smaller than 0.1 mm.
(c) When 9 X d 85 is less than 0_2 mm, use 0.2 mm_
(d) A - percent passing the No. 200 (0.075 mm) sieve after any regrading.
(e) When 4 X d 85 is less than 0.7 mm, use 0.7 mm-
(f) In category 4, the D15 :,; 4 X d 85 criterion should be used in the case of
filters beneath riprap subject to wave action and drains which may be
subject to violent surging and/or vibration.
D-4
EM 1110-2-1901
30 Apr 93
Change 1
Table D-3. D10 and D90 Limits for Preventing Segregation
Minimwn D10 Maximum D90

(mm) (mm)
<0.5 20
0.5 1.0 25
1.0 2.0 30
2.0 5.0 40
5.0 10 50
10 50 60
D-3. PermeabilitY. The requirement that seepage move more quickly through
the filter than through the protected soil (called the permeability criterion)
is again met by a grain-size relationship criterion based on experience:
Permeability
15 percent size of filter material

" 3 to 5 (D-1)
15 percent of the protected soil
Permeability of a granular soil is roughly proportional to the square of the

10- to 15-percent size material. Thus, the permeability criterion ensures
that filter materials have approximately 9 to 25 or more times the permeabil-
ity of the protected soil. Generally, a permeability ratio of at least 5 is
preferred; however, in the case of a wide band of uniform base material
gradations, a permeability ratio as low as 3 may be used with respect to the
maximum 15-percent size of the base material. There may be situations, par-
ticularly where the filter is not part of a drain, where the permeability of
the filter is not important. In those situations, this criterion may be
ignored.
D-4. Applicability. The previously given filter criteria in Table D-2 and
Equation D-1 are applicable for all soils (cohesionless or cohesive soils)
including dispersive soils.( 4 l However, laboratory filter tests for disper-
sive soils, very fine silt, and very fine cohesive soils with high plastic
limits are recommended.
( 4)
Sherard, J. L., L. P. Dunnigan, "Filters and Leakage Control in Embank-
ment Dams," Proceeding of the Symposium on Seepage and Leakage from Dams
and Impoundments, ASCE National Convention, Denver, Colorado, 1985.
D-5
EM 1110-2-1901
30 Apr 93
Change 1
D-5. Perforated Pipe.C 5 l The following criteria are applicable for prevent-
ing infiltration of filter material into perforated pipe, screens, etc.:
Minimum 50 percent size of filter material (D- 2)

;, l. 0
hole diameter or slot width
In many instances a filter material meeting the criteria given by Table D-2
and Equation D-1 relative to the material being drained is too fine to meet
the criteria given by Equation D-2. In these instances, multilayered or
"graded" filters are required. In a graded filter each layer meets the
requirements given by Table D-2 and Equation D-1 with respect to the pervious
layer with the final layer in which a collector pipe is bedded also meeting
the requirements given by Equation D-2. Graded filter systems may also be
needed when transitioning from fine to coarse materials in a zoned embankment
or where coarse material is required for improving the water-carrying capacity
of the system.
D-6. Gap-Graded Base. The preceding criteria cannot, in most instances, be

applied directly to protect severely gap- or skip-graded soils. In a gap-
graded soil such as that shown in Figure D-1, the coarse material simply
floats in the matrix of fines. Consequently, the scattered coarse particles
will not deter the migration of fines as they do in a well-graded material.
For such gap-graded soils, the filter should be designed to protect the fine i.
matrix rather than the total range of particle sizes. This is illustrated in
Figure D-1. The 85-percent size of the total sample is 5.2 mm. Considering
only the matrix material, the 85-percent size would be 0.1 mm resulting in a
much finer filter material being required. This procedure may also be fol-
lowed in some instances where the material being drained has a very wide range
of particle sizes (e.g., materials graded from coarse gravels to significant
percentages of silt or clay). For major structures such a design should be
checked with filter tests.
D-7. Gap-Graded Filter. A gap-graded filter material must never be specified

or allowed since it will consist of either the coarse particles floating in
the finer material or the fine material having no stability within the voids
produced by the coarse material. In the former case the material may not be
permeable enough to provide adequate drainage. The latter case is particu-
larly dangerous since piping of the protected material can easily occur
through the relatively large, loosely filled voids provided by the coarse
material.
(5) EM 1110-2-2300 states, "Collector pipe should not be placed within the
embankment, except at the downstream toe, because of the danger of
either breakage or separation of joints, resulting from fill placement
and compaction operations, or settlement, which might result in either
clogging and/or piping."
D-6
u .4 , _ ...I ,.._
I I
r\,
'"'f"'
. •
1
... .... .. . -." ,. ..
I I
\l t. fi_,.PAM
••
•t~ NWI:!US
1• UIO
•
_,..
-- -~--- ·+ -- -
..
f- ., -- 1- -.
1-- r-- o., o., - "

• .... ~.;. M~TfOIAL
.. - . -I- f- - ~- .
"""
- TOTAL MATf~l.~l 7 ~ - r---..--- •
- -- -
I - ..... -- ···- -
.
. -- - 1- ~
\.
- \- - - - 1-
-I ~
I' ~· .. ~
-- - .. \ I
I - ·-· - -- - ··- -I- 1- -- '' - -- - -- e
fi
.
-" ~-
. -- - -·- 1 - . ---- - . - -- I -
-
.. ·- -- ~-
1-r-1- -- ~-
. , __ •
-I _,. I
..
-
-
.......I
.
...
•
I - I
UWM ••
san
•!""
..
.._LIN(fla
~
I ''!"
-. -
. ...
I
1- --·
...
• • Dill Q.Af
- 1- -·
....
-- ...
Figure D-1. Analysis of gap-graded material (from EM 1110-2-1913)

EM 1110-2-1901
30 Apr 93
Change 1
•
D-8. Broadly Graded Base. One of the more common soils used for embankment
dams is a broadly graded material with particle sizes ranging from clay sizes
to coarse gravels and including all intermediate sizes. These soils may be of
glaGial, alluvial-colluvial, or weathered rock origin. As noted by Sherard,
since the 85-percent size of the soil is commonly on the order of 20 to 30 mm,
a direct application of the stability criteria D15 jd 05 :5 4 to 5 would allow
very coarse uniform gravel without sand sizes as a downstream filter, which
would not be satisfactory (Sherard 1979). The typical broadly graded soils
fall in Soil Category 2 in Table D-2 and require a sand or gravelly filter
with D15 :5 0. 7 mm.
D-9. Example of Graded Filter Design for Drain. Seldom, if ever, is a single
gradation curve representative of a given material. A material is generally
represented by a gradation band which encompasses all the individual gradation
curves. Likewise, the required gradation for the filter material is also
given as a band. The design of a graded filter which shows the application of
the filter criteria where the gradations are represented by bands is illus-
trated in Figure D-2. A typical two-layer filter for protecting an impervious
core of a dam is illustrated. The impervious core is a fat clay (CH) with a
trace of sand which falls in Category 1 soil in Table D-2. The criterion
D15 :5 9 x d 85 is applied and a "point a" is established in Figure D-2. Filter
material graded within a band such as that shown for Filter A in Figure D-2 is .-
acceptable based on the stability criteria. The fine limit of the band was (
arbitrarily drawn, and in this example, is intended to represent the gradation
of a readily available material. A check is then made to ensure that the
15-percent size of the fine limit of the filter material band (point b) is
equal to or greater than 3 to 5 times the 15-percent size of the coarse limit
of the drained material band (point c). Filter A has a minimum D10 size and a
maximum D90 size such that, based on Table D-3, segregation during placement
can be prevented. Filter A meets both the stability and permeability require-
ments and is a suitable filter material for protecting the impervious core
material. The second filter, Filter B, usually is needed to transition from a
fine filter (Filter A) to coarse materials in a zoned embankment dam.
Filter B must meet the criteria given by Table D-2 with respect to Filter A.
For stability, the 15-percent size of the coarse limit of the gradation band
for the second filter (point d) cannot be greater than 4 to 5 times the
85-percent size of the fine limit of the gradation band for Filter·A
(point e). For permeability, the 15-percent size of the fine limit (point f)
must be at least 3 to 5 times greater than the 15-percent size of the coarse
limit for Filter A (point a). With points d and f established, the fine and
coarse limits for Filter B may be established by drawing curves through the
points approximately parallel to the respective limits for Filter A. A check
is then made to see that the ratio of maximum D90 /minimum D10 size of Filter B
is approximately in the range as indicated in Table D-3. A well-graded filter
which usually would not meet the requirements in Table D-3 may be used if
segregation can be controlled during placement. Figure D-2 is intended to
show only the principles of filter design. The design of thickness of a
D-8
U.S. STANDARD SIEVE OPENING IN INCHES U.S. STANDARD SIEVE NUMBERS HYDROMETER
111512 10 I I 4 3 2 1¥1 1 ¥. ~ ~ 3 • I I 10 U 11 20 3D •o 50 70 100 140 2:00
100
I' \ ' .' ' '' . D
.. ' \e
"-..
D. (das) min
10
10 11 1\ 1\ 1\ 20
\ ' 1\ \
70 '\ \
~
.,.
\ "" ~~ 30
\
_1 ~
\ \ ':::<
?
..
'\ ~0.~
~ 1\ '\.~c
t:1
; 50
fP
\ '-o,..,
' 1\ \
"' "' \ \ ''-
\ a.;; 9 (das) min '
30
\ \ ,'\ ~ 70
\ I ;;.3 to 5a l\ ~
,.
' I
al '\
\
D.b b;;.3 to 5c '\

C\
10
D
d.;; to 5e 1\ ~
--"'- 90
• 100 .. I I I
10
\
0.5
"" 0.1 0.05 0.01 0.005
100
0.001
GRAIN SIZE IN MIWMETERS
Figure D-2. Illustration of the design of a graded filter

EM 1110-2-1901
30 Apr 93
Change 1
•
filter for sufficient seepage discharge capacity is done by applying Darcy's
Law, Q - kia, and one example is presented in Chapter 8 of the main text.
D-10. Construction. EM 1110-2-1911 and EM 1110-2-2300 provide guidance for

construction. Major concerns during construction include:
a. Prevention of contamination of drains and filters by runoff contain-

ing sediment, dust, construction traffic, and mixing with nearby fine-grained
materials during placement and compaction. Drain and filter material may be
kept at an elevation higher than the surrounding fine-grained materials during
construction to prevent-contamination by sediment-carrying runoff.
b. Prevention of segregation, particularly well-graded filters, during

handling and placement.
c. Proper in-place density is usually required to be an average of

85-percent relative density with no area less than SO-percent relative
density. Granular materials containing little or no fines should be saturated
during compaction to prevent "bulking" (low density) which can result in
settlement when overburden materials are placed and the drain is subsequently
saturated by seepage flows.
d. Gradation should be monitored closely so that designed filter cri- •.'~

teria are met.
e. Thickness of layers should be monitored to ensure designed discharge

capacity and continuity of the filter.
Thus, quality control/assurance is very important during filter construction

because of the critical function of this relatively small part of the
embankment.
D-11. Monitoring. Monitoring of seepage quantity and quality (see Chapter 13

for method of monitoring seepage) once the filter is functioning is very
important to the safety of the embankment. An increase in seepage flow may be
due to a higher reservoir level or may be caused by cracking or piping. The
source of the additional seepage should be determined and action taken as
required (Chapters 12, 13, and 14). Decreases in seepage flows may also sig-
nal dangers such as clogging of the drain(s) with piped material, iron oxide,
calcareous material, and effects of remedial grouting. Again, the cause
should be determined and appropriate remedial measures taken. Drain outlets
should be kept free of sediment and vegetation. In cold climates, design or
maintenance measures should be taken to prevent clogging of drain outlets by
ice.
D-10

Em - 1110-2-1901 Seepage Analysis and PDF

Uploaded by

Copyright:

Available Formats

Em - 1110-2-1901 Seepage Analysis and PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Em - 1110-2-1901 Seepage Analysis and PDF

Uploaded by

Copyright:

Available Formats

CECW-EG Department of the Army EM 1110-2-1901

U.S. Army Corps of Engineers

Engineering and Design

SEEPAGE ANALYSIS AND

Distribution Restriction Statement

Engineering and Design

FOR TilE COMMANDER:

ENGINEERING AND DESIGN

SEEPAGE ANALYSIS AND

DEPARTMENT OF THE ARMY

Engineer Manual 30 September 1986

Engineering and Design

1. Purpose. This manual presents the fundamental design principles and

2. Applicability. This manual is applicable to all HQUSACE/OCE elements and

FOR THE COMMANDER:

This manual supersedes EM 1110-2-1901 dated February 1952

Engineer Manual 30 September 1986

Engineering and Design

Subject Paragraph Page

CHAPTER 2. DETERMINATION OF PERMEABILITY OF SOIL

CHAPTER 3. DETERMINATION OF PERMEABILITY OF ROCK

Permeabilities of Rock Masses-------------- 3-1 3-1

This manual supersedes EM 1110-2-1901, Feb 52

Subject Paragraph Page

CHAPTER 4. SEEPAGE PRINCIPLES

General Considerations---------------------- 4-1 4-1

CHAPTER 5. CONFINED FLOW PROBLEMS

General Considerations---------------------- 5-1 5-1

CHAPTER 6. UNCONFINED FLOW PROBLEMS

Introduction -------------------------------- 6-1 6-1

CHAPTER 7. SEEPAGE TOWARD WELLS

Use of Wells -------------------------------- 7-1 7-1

Subject Paragraph Page

CHAPTER 8. SEEPAGE CONTROL IN EMBANKMENTS

General------------------------------------- 8-1 8-1

CHAPTER 9. SEEPAGE CONTROL IN EARTH FOUNDATIONS

General------------------------------------- 9-1 9-1

CHAPTER 10. SEEPAGE CONTROL THROUGH EARTH ABUTMENTS

Through Earth Abutments--------------------- 10-1 10-1

CHAPTER 11. SEEPAGE CONTROL IN ROCK FOUNDATIONS AND

General Considerations---------------------- 11-1 11-1

CHAPTER 12. REMEDIAL SEEPAGE CONTROL

General Considerations---------------------- 12-1 12-1

Subject Paragraph Page

Upstream Impervious Blanket----------------- 12-5 12-4

CHAPTER 13. MONITORING PERFORMANCE OF SEEPAGE CONTROL

General Considerations---------------------- 13-1 13-1

CHAPTER 14. INSPECTION, MAINTENANCE, AND REHABILITA-

Introduction -------------------------------- 14-1 14-1

APPENDIX B. APPROXIMATE METHODS FOR ANALYSIS OF FLOW

Introduction-------------------------------- B-1 B-1

APPENDIX C. ANALYSIS OF PRESSURE INJECTION TESTS

Subject Paragraph Page

APPENDIX D. FILTER AND DRAIN DESIGN AND CONSTRUCTION

General--------------------------------------- D-1 D-1

1-2. Applicability. The provisions of this manual are applicable to all

1-3 References. Appendix A contains a list of Government and non-Government

2-1. Darcy's Law.

a. Development of Darcy's Law. Henry Darcy, a French engineer, con-