PDHonline Course C802 (2 PDH)

An Introduction to
Design of Dewatering Systems

J. Paul Guyer, P.E., R.A.

2015

PDH Online | PDH Center

5272 Meadow Estates Drive
Fairfax, VA 22030-6658
Phone & Fax: 703-988-0088
www.PDHonline.org
www.PDHcenter.com

An Approved Continuing Education Provider

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

An Introduction to Design of Dewatering Systems

J. Paul Guyer, P.E., R.A.

CONTENTS

1. ANALYSIS OF GROUNDWATER FLOW

2. MATHEMATICAL AND MODEL ANALYSES
3. FLOW NET ANALYSES
4. ELECTRICAL ANALOGY SEEPAGE MODELS
5. NUMERICAL ANALYSES
6. WELLPOINTS, WELLS, AND FILTERS
7. PUMPS, HEADERS, AND DISCHARGE PIPES
8. FACTORS OF SAFETY
9. DEWATERING OPEN EXCAVATIONS
10. DEWATERING SHAFTS AND TUNNELS
11. PERMANENT PRESSURE RELIEF SYSTEMS
12. FREEZING
13. CONTROL OF SURFACE WATER

(This publication is adapted from the Unified Facilities Criteria of the United States
government which are in the public domain, have been authorized for unlimited
distribution, and are not copyrighted.)

(Figures, tables and formulas in this publication may at times be a little difficult to read,
but they are the best available. DO NOT PURCHASE THIS PUBLICATION IF THIS
LIMITATION IS NOT ACCEPTABLE TO YOU.)

©2015 J. Paul Guyer Page 2 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

1. ANALYSIS OF GROUNDWATER FLOW.

1.1 DESIGN OF A DEWATERING and pressure relief or groundwater control system

first requires determination of the type of groundwater flow (artesian, gravity, or
combined) to be expected and of the type of system that will be required. Also, a
complete picture of the groundwater and the subsurface condition is necessary. Then
the number, size, spacing, and penetration of wellpoints or wells and the rate at which
the water must be removed to achieve the required groundwater lowering or pressure
relief must be determined.

1.2 IN THE ANALYSIS OF ANY DEWATERING SYSTEM, the source of seepage

must be determined and the boundaries and seepage flow characteristics of geologic
and soil formations at and adjacent to the site must be generalized into a form that can
be analyzed. In some cases, the dewatering system and soil and groundwater flow
conditions can be generalized into rather simple configurations. For example, the
source of seepage can be reduced to a line or circle; the aquifer to a homogeneous,
isotropic formation of uniform thickness; and the dewatering system to one or two
parallel lines or circle of wells or wellpoints. Analysis of these conditions can generall y
be made by means of mathematical formulas for flow of groundwater. Complicated
configurations of wells, sources of seepage, and soil formations can, in most cases, be
solved or at least approximated by means of flow nets, electrical analogy models,
mathematical formulas, numerical techniques, or a combination of these methods.

1.3 ANY ANALYSIS, EITHER mathematical, flow net, or electrical analogy, is not
better than the validity of the formation boundaries and characteristics used in the
analysis. The solution obtained, regardless of the rigor or precision of the analysis, will
be representative of actual behavior only if the problem situation and boundary
conditions are adequately represented. An approximate solution to the right problem is
far more desirable than a precise solution to the wrong problem. The importance of
formulating correct groundwater flow and boundary conditions, cannot be emphasized
too strongly.

©2015 J. Paul Guyer Page 3 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

1.4 METHODS FOR DEWATERING AND PRESSURE RELIEF and their suitability for
various types of excavations and soil conditions are described in the technical
literature. The investigation of factors relating to groundwater flow and to design of
dewatering systems are discussed in the technical literature. Mathematical, graphical,
and electroanalogous methods of analyzing seepage flow through generalized soil
conditions and boundaries to various types of dewatering or pressure relief systems
are presented in the technical literature.

1.5 OTHER FACTORS THAT have a bearing on the actual design of dewatering,
permanent drainage, and surface- water control systems are considered in this
discussion.

1.6 THE FORMULAS AND FLOW NET PROCEDURES presented in this discussion
are for a steady state of groundwater flow. During initial stages of dewatering an
excavation, water is removed from storage and the rate of flow is larger than required
to maintain the specified drawdown. Therefore, initial pumping rates will probably be
about 30 percent larger than computed values.

©2015 J. Paul Guyer Page 4 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

2. MATHEMATICAL AND MODEL ANALYSES.

2.1 GENERAL.

2.1.1 DESIGN. Design of a dewatering system requires the determination of the

number, size, spacing, and penetration of wells or wellpoints and the rate at which
water must be removed from the pervious strata to achieve the required groundwater
lowering or pressure relief. The size and capacity of pumps and collectors also depend
on the required discharge and drawdown. The fundamental relations between well and
wellpoint discharge and corresponding drawdown are presented in this discussion.
The equations presented assume that the flow is laminar, the pervious stratum is
homogeneous and isotropic, the water draining into the system is pumped out at a
constant rate, and flow conditions have stabilized. Procedures for transferring an
anisotropic aquifer, with respect to permeability, to an isotropic section are presented
in the technical literature.

2.1.2 SLOTS AND WELLS. The equations referenced are in two groups: flow and
drawdown to slots and flow and drawdown to wells. Equations for slots are applicable
to flow to trenches, French drains, and similar drainage systems. They may also be
used where the drainage system consists of closely spaced wells or wellpoints.
Assuming a well system equivalent to a slot usually simplifies the analysis; however,
corrections must be made to consider that the drainage system consists of wells or
wellpoints rather than the more efficient slot. These corrections are given with the well
formulas discussed below. When the well system cannot be simulated with a slot, well
equations must be used. The equations for slots and wells do not consider the effects
of hydraulic head losses H w in wells or wellpoints; procedures for accounting for these
effects are presented separately.

2.1.3 RADIUS OF INFLUENCE R. Equations for flow to drainage systems from a

circular seepage source are based on the assumption that the system is centered on
an island of radius R. Generally, R is the radius of influence that is defined as the

©2015 J. Paul Guyer Page 5 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

radius of a circle beyond which pumping of a dewatering system has no significant

effect on the original groundwater level or piezometric surface. Where there is little or
no recharge to an aquifer, the radius of influence will become greater with pumping
time and with increased drawdown in the area being dewatered. Generally, R is
greater for coarse, very pervious sands than for finer soils. If the value of R is large
relative to the size of the excavation, a reasonably good approximation of R will serve
adequately for design because flow and drawdown for such a condition are not
especially sensitive to the actual value of R. As it is usually impossible to determine R
accurately, the value should be selected conservatively from pumping test data or the
technical literature.

2.1.4 WETTED SCREEN. There should always be sufficient well and screen length
below the required drawdown in a well in the formation being dewatered so that the
design or required pumping rate does not produce a gradient at the interface of the
formation and the well filter (or screen) or at the screen and filter that starts to cause
the flow to become turbulent. Therefore, the design of a dewatering system should
always be checked to see that the well or wellpoints have adequate “wetted screen
length hws” or submergence to pass the maximum computed flow. The limiting flow q c
into a filter or well screen is approximately equal to

©2015 J. Paul Guyer Page 6 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 1
Head at center of fully and partially penetrating circular slots;
circular source; artesian flow.

©2015 J. Paul Guyer Page 7 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 2
Flow and drawdown at slot for fully and partially penetrating rectangular slots; circular
source; artesian flow.

©2015 J. Paul Guyer Page 8 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 3
Head within a partially penetrating rectangular slot; circular source; artesian flow

©2015 J. Paul Guyer Page 9 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

2.1.5 HYDRAULIC HEAD LOSS HW. Some equations in the technical literature do not
consider hydraulic head losses that occur in the filter, screen, collector pipes, etc.
These losses cannot be neglected, however, and must be accounted for separately.
The hydraulic head loss through a filter and screen will depend upon the diameter of
the screen, slot width, and opening per foot of screen, permeability and thickness of
the filter; any clogging of the filter or screen by incrustation, drilling fluid, or bacteria;
migration of soil or sand particles into the filter; and rate of flow per foot of screen.

2.1.6 WELL OR SCREEN PENETRATION W/D.

2.1.6.1 EQUATIONS AND GRAPHS for partially penetrating slots or wells are
generally based on those for fully penetrating drainage systems modified by model
studies and, in some instances, mathematical derivations. The amount or percent of
screen penetration required for effective pressure reduction or interception of seepage
depends upon many factors, such as thickness of the aquifer, distance to the effective
source of seepage, well or wellpoint radius, stratification, required “wetted screen
length,” type and size of excavation, and whether or not the excavation penetrates
alternating pervious and impervious strata or the bottom is underlain at a shallow
depth by a less pervious stratum of soil or rock. Where a sizeable open excavation or
tunnel is underlain by a fairly deep stratum of sand and wells are spaced rather widely,
the well screens should penetrate at least 25 percent of the thickness of the aquifer to
be dewatered below the bottom of the excavation and more preferably 50 to 100
percent. Where the aquifer(s) to be dewatered is stratified, the drainage slots or well
screens should fully penetrate all the strata to be dewatered. If the bottom of an
excavation in a pervious formation is underlain at a shallow depth by an impervious
formation and the amount of “wetted screen length” available is limited, the drainage
trench or well screen should penetrate to the top of the underlying less pervious
stratum. The hydraulic head loss through various sizes and types of header or
discharge pipe, and for certain well screens and (clean) filters is determined from
laboratory and field tests.

©2015 J. Paul Guyer Page 10 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

©2015 J. Paul Guyer Page 11 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

2.1.6.2 HEAD LOSSES IN THE SCREENED SECTION of a well Hs are based on

equal inflow per unit of screen surface and turbulent flow inside the well and is
equivalent to the entire well flow passing through one-half the screen length. Hydraulic
head loss within a wellpoint system can be estimated. Flow into a well can be impeded
by the lack of “wetted screen length,” in addition to hydraulic head losses in the filter or
through the screens and/or chemical or mechanical clogging of the aquifer and filter.

2.2 FLOW TO A DRAINAGE SLOT.

2.2.1 LINE DRAINAGE SLOTS. Equations presented in the technical literature can be
used to compute flow and head produced by pumping either a single or a double
continuous slot of infinite length. These equations assume that the source of seepage
and the drainage slot are infinite in length and parallel and that seepage enters the
pervious stratum from a vertical line source. In actuality, the slot will be of finite length,
the flow at the ends of the slot for a distance of about L/2 (where L equals distance
between slot and source) will be greater, and the drawdown will be less than for the
central portion of the slot. Flow to the ends of a fully penetrating slot can be estimated,
if necessary, from flow-net analyses.

©2015 J. Paul Guyer Page 12 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Note: A = artesian flow; G = gravity flow; C = combined artesian-gravity

flow; F = fully penetrating; P = partially penetrating.

Table 1
Index to figures for flow, head, or drawdown equations for given corrections

2.2.2 CIRCULAR AND RECTANGULAR SLOTS. Equations for flow and head or
drawdown produced by circular and rectangular slots supplied by a circular seepage
source are given in the technical literature. Equations for flow from a circular seepage
source assume that the slot is located in the center of an island of radius R. For many
dewatering projects, R is the radius of influence rather than the radius of an island,
and procedures for determining the value of R are discussed in the technical liter ature.

©2015 J. Paul Guyer Page 13 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Dewatering systems of relatively short length are considered to have a circular source
where they are far removed from a line source such as a river or shoreline.

2.2.3 USE OF SLOTS FOR DESIGNING WELL SYSTEMS. Wells can be substituted
for a slot; and the flow Q w, drawdown at the well (H-hw) neglecting hydraulic head
losses at and in the well, and head midway between the wells above that in the wells
ΔHm can be computed from the equations given in the technical literature for a (single)
line source for artesian and gravity flow for both “fully” and “partially” penetrating wells
where the well spacing a is substituted for the length of slot x.

2.2.4 PARTIALLY PENETRATING SLOTS. The equations for gravity flow to partially
penetrating slots are only considered valid for relatively high-percent penetrations.

2.3 FLOW TO WELLS.

2.3.1 FLOW TO WELLS FROM A CIRCULAR SOURCE.

2.3.1.1 EQUATIONS FOR FLOW and drawdown produced by a single well supplied
by a circular source are given in the technical literature. It is apparent that
considerable computation is required to determine the height of the phreatic surface
and resulting drawdown in the immediate vicinity of a gravity well (r/h less than 0.3).
The drawdown in this zone usually is not of special interest in dewatering systems and
seldom needs to be computed. However, it is always necessary to compute the water
level in the well for the selection and design of the pumping equipment.

2.3.1.2 THE GENERAL EQUATIONS for flow and drawdown produced by pumping a
group of wells supplied by a circular source are given in the technical literature. These
equations are based on the fact that the drawdown at any point is the summation of
drawdowns produced at that point by each well in the system. The drawdown factors F
to be substituted into the general equations appear in the equations for both artesian

©2015 J. Paul Guyer Page 14 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

and gravity flow conditions. Consequently, the factors given in the technical literature
for commonly used well arrays are applicable for either condition.

2.3.1.3 FLOW AND DRAWDOWN for circular well arrays can also be computed, in a
relatively simple manner, by first considering the well system to be a slot. However,
the piezometric head in the vicinity of the wells (or wellpoints) will not correspond
exactly to that determined for the slot due to convergence of flow to the wells. The
piezometric head in the vicinity of the well is a function of well flow Q w; well spacing a;
well penetration W; effective well radius r w; aquifer thickness D, or gravity head H; and
aquifer permeability k.

2.3.2 FLOW TO WELLS FROM A LINE SOURCE.

2.3.2.1 EQUATIONS ARE GIVEN in the technical literature for flow and drawdown
produced by pumping a single well or group of fully penetrating wells supplied from an
infinite line source were developed using the method of image wells. The image well (a
recharge well) is located as the mirror image of the real well with respect to the line
source and supplies the pervious stratum with the same quantity of water as that being
pumped from the real well.

2.3.2.2 THE EQUATIONS GIVEN in the technical literature for multiple-well systems
supplied by a line source are based on the fact that the drawdown at any point is the
summation of drawdowns produced at that point by each well in the system.
Consequently, the drawdown at a point is the sum of the drawdowns produced by the
real wells and the negative drawdowns produced by the image or recharge wells.

2.3.2.3 EQUATIONS ARE GIVEN in the technical literature for flow and drawdown
produced by pumping an infinite line at wells supplied by a (single) line source. The
equations are based on the equivalent slot assumption. Where twice the distance to a
single line source or 2L is greater than the radius of influence R, the value of R as
determined from a pumping test or from the technical literature should be used in lieu

©2015 J. Paul Guyer Page 15 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

of L unless the excavation is quite large or the tunnel is long, in which case equations
for a line source or a flow-net analysis should be used.

2.3.2.4 EQUATIONS FOR COMPUTING the head midway between wells above that
in the wells Ah, are not given in this discussion for two line sources adjacent to a
single line of wells. However, such can be readily determined from (plan) flow-net
analyses.

2.3.3 LIMITATIONS ON FLOW TO A PARTIALLY PENETRATING WELL.

Theoretical boundaries for a partially penetrating well (for artesian flow) are
approximate relations intended to present in a simple form the results of more rigorous
but tedious computations. The rigorous computations were made for ratios of R/D =
4.0 and 6.7 and a ratio R/r w = 1000. As a consequence, any agreement between
experimental and computed values cannot be expected except for the cases with
these particular boundary conditions. In model studies the flow from a partially
penetrating well was based on the formula:

Shown is some of the results obtained for wells of various penetrations centered inside
a circular source. Also presented are boundary curves computed for well-screen
penetrations of 2 and 50 percent. Comparison of data computed from model data with

©2015 J. Paul Guyer Page 16 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

that computed from the boundary formulas indicates fairly good agreement for well
penetrations > 25 percent and values of R/D between about 5 and 15 where R/r w >
200 to 1000. Other empirical formulas for flow from a partially penetrating well suffer
from the same limitations.

2.3.4 PARTIALLY PENETRATING WELLS. The equations for gravity flow to partially
penetrating wells are only considered valid for relatively high-percent penetrations.

©2015 J. Paul Guyer Page 17 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

3. FLOW-NET ANALYSES.

3.1 FLOW NETS ARE VALUABLE where irregular configurations of the source of
seepage or of the dewatering system make mathematical analyses complex or
impossible. However, considerable practice in drawing and studying good flow nets is
required before accurate flow nets can be constructed.

3.2 A FLOW NET IS A GRAPHICAL REPRESENTATION of flow of water through an

aquifer and defines paths of seepage (flow lines) and contours of equal piezometric
head (equipotential lines). A flow net may be constructed to represent either a plan or
a section view of a seepage pattern. Before a sectional flow net can be constructed,
boundary conditions affecting the flow pattern must be delineated and the pervious
formation transformed into one where k n = kv . In drawing a flow net, the following
general rules must be observed:

3.2.1 FLOW LINES AND equipotential lines intersect at right angles and form
curvilinear squares or rectangles.

3.2.2 THE FLOW BETWEEN any two adjacent flow lines and the head loss between
any two adjacent equipotential lines are equal, except where the plan or section
cannot be divided conveniently into squares, in which case a row of rectangles will
remain with the ratio of the lengths to the sides being constant.

3.2.3 A DRAINAGE SURFACE exposed to air is neither an equipotential nor flow line,
and the squares at this surface are incomplete; the flow and equipotential lines need
not intersect such a boundary at right angles.

3.2.4 FOR GRAVITY FLOW, equipotential lines intersect the phreatic surface at equal
intervals of elevation, each interval being a constant fraction of the total net head.

©2015 J. Paul Guyer Page 18 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

3.3 FLOW NETS ARE LIMITED to analysis in two dimensions; the third dimension in
each case is assumed infinite in extent. An example of a sectional flow net showing
artesian flow from two line sources to a partially penetrating drainage slot is indicated.
An example of a plan flow net showing artesian flow from a river to a line of relief wells
is also indicated.

3.4 THE FLOW PER UNIT LENGTH (for sectional flow nets) or depth (for plan flow
nets) can be computed. Drawdowns from either sectional or plan flow nets can be
computed as well. In plan flow nets for artesian flow, the equipotential lines
correspond to various values of H h, whereas for gravity flow, they correspond to H 2-h2.
Since section equipotential lines for gravity flow conditions are curved rather than
vertical, plan flow nets for gravity flow conditions give erroneous results for large
drawdowns and should always be used with caution.

3.5 PLAN FLOW NETS GIVE erroneous results if used to analyze partially penetrating
drainage systems, the error being inversely proportional to the percentage of
penetration. They give fairly accurate results if the penetration of the drainage system
exceeds 80 percent and if the heads are adjusted as described in the following
paragraph.

3.6 IN PREVIOUS ANALYSES of well systems by means of flow nets, it was assumed
that dewatering or drainage wells were spaced sufficiently close to be simulated by a
continuous drainage slot and that the drawdown (H-hD) required to dewater an area
equaled the average drawdown at the drainage slot or in the lines of wells (H-h).
These analyses give the amount of flow Q T that must be pumped to achieve H-hD but
do not give the drawdown at the wells. The drawdown at the wells required to produce
H-ho downstream or within a ring of wells can be computed (approximately) for
artesian flow from plan flow nets by the equations shown in if the wells have been
spaced proportional to the flow. The drawdown at fully penetrating gravity wells can
also be computed.

©2015 J. Paul Guyer Page 19 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 4
Shape factors for wells of various penetrations centered inside a circular source

©2015 J. Paul Guyer Page 20 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

4. ELECTRICAL ANALOGY SEEPAGE MODELS.

4.1 THE LAWS GOVERNING FLOW of fluids through porous media and flow of
electricity through pure resistance are mathematically similar. Thus, it is feasible to use
electrical models to study seepage flows and pressure distribution for various seepage
conditions. Both two and three-dimensional models can be used to solve seepage
problems.

4.2 DARCY’S LAW FOR TWO-DIMENSIONAL flow of water through soil can be
expressed for unit length of soil formations as follows:

©2015 J. Paul Guyer Page 21 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 5
Flow and drawdown to wells computed from flow-net analyses

©2015 J. Paul Guyer Page 22 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Since the permeability in fluid flow is analogous to the reciprocal of the specific
resistance for geometrically similar mediums, the shape factors for Darcy’s law and
Ohm’s law are the same.

4.3 A TWO-DIMENSIONAL FLOW NET can be constructed using a scale model of

the flow and drainage system made of a conductive material representing the porous
media (graphite-treated paper or an electrolyte solution), copper or silver strips for
source of seepage and drainage, and nonconductive material representing impervious
flow boundaries. The electrical circuit consists of a potential applied across the model
and a wheatstone bridge to control intermediate potentials on the model. The flow net
is constructed by tracing lines of constant potential on the model, thus establishing the
flow-net equipotential lines after which the flow lines are easily added graphically. A
flow net constructed using an electrical analogy model may be analyzed in the same
manner as the one constructed.

4.4 THE EQUIPMENT FOR CONDUCTING three-dimensional electrical analog model

studies consists basically of a large plexiglass tank filled with diluted copper sulfate
solution and having a calibrated, elevated carrier assembly for the accurate positioning
of a point electrode probe anywhere in the fluid medium. A prototype is simulated by
fabricating appropriately shaped and sealed source and sink configurations and
applying an electrical potential across them. The model is particularly useful for
analyzing complex boundary conditions that cannot be readily analyzed by two-
dimensional techniques.

©2015 J. Paul Guyer Page 23 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

5. NUMERICAL ANALYSES.

5.1 MANY COMPLEX SEEPAGE PROBLEMS, including such categories as steady

confined, steady unconfined, and transient unconfined can be solved using the finite
element method.

5.2 A GENERAL COMPUTER CODE for analyzing partially penetrating random well
arrays has been developed based on results of three-dimensional electrical analogy
model tests. The computer code provides a means for rapidly analyzing trial well
systems in which the number of wells and their geometric configuration can be varied
to determine quantities of seepage and head distributions. Wells of different radii and
penetrations can be considered in the analysis.

©2015 J. Paul Guyer Page 24 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

6. WELLPOINTS, WELLS, AND FILTERS. Wells and wellpoints should be of a type

that will prevent infiltration of filter material or foundation sand, offer little resistance to
the inflow of water, and resist corrosion by water and soil. Wellpoints must also have
sufficient penetration of the principal water-carrying strata to intercept seepage without
excessive residual head between the wells or within the dewatered area.

6.1 WELLPOINTS. Where large flows are anticipated, a high-capacity type of wellpoint
should be selected. The inner suction pipe of self-jetting wellpoints should permit
inflow of water with a minimum hydraulic head loss. Self-jetting wellpoints should be
designed so that most of the jet water will go out the tip of the point, with some
backflow to keep the screen flushed clean while jetting the wellpoint in place.

6.1.1 WELLPOINT SCREENS. Generally, wellpoints are covered with 30- to 60-mesh
screen or have an equivalent slot opening (0.010 to 0.025 inch). The mesh should
meet filter criteria given in below. Where the soil to be drained is silty or fine sand, the
yield of the wellpoint and its efficiency can be greatly improved by placing a relatively
uniform, medium sand filter around the wellpoint. The filter should be designed in
accordance with criteria subsequently set forth in below. A filter will permit the use of
screens or slots with larger openings and provide a more pervious material around the
wellpoint, thereby increasing its effective radius (d below).

6.1.2 WELLPOINT HYDRAULICS. The hydraulic head losses in a wellpoint system

must be considered in designing a dewatering system.

6.2 WELLS. Wells for temporary dewatering and permanent drainage systems may
have diameters ranging from 4 to 18 inches with a screen 20 to 75 feet long depending
on the flow and pump size requirements.

6.2.1 WELL SCREENS. Screens generally used for dewatering wells are slotted (or
perforated) steel pipe, perforated steel pipe wrapped with galvanized wire, galvanized
wire wrapped and welded to longitudinal rods, and slotted polyvinyl chloride (PVC)

©2015 J. Paul Guyer Page 25 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

pipe. Riser pipes for most dewatering wells consist of steel or PVC pipe. Screens and
riser for permanent wells are usually made of stainless steel or PVC. Good practice
dictates the use of a filter around dewatering wells, which permits the use of fairly
large slots or perforations, usually 0.025 to 0.100 inch in size. The slots in well screens
should be as wide as possible but should meet criteria given in below.

6.2.2 OPEN SCREEN AREA. The open area of a well screen should be sufficient to
keep the entrance velocity for the design flow low to reduce head losses and to
minimize incrustation of the well screen in certain types of water. For temporary
dewatering wells installed in non-incrusting groundwater, the entrance velocity should
not exceed about 0.15 to 0.20 foot per second; for incrusting groundwater, the
entrance velocity should not exceed 0.10 to 0.20 foot per second. For permanent
drainage wells, the entrance velocity should not exceed about 0.10 foot per second.
As the flow to and length of a well screen is usually dictated by the characteristics of
the aquifer and drawdown requirements, the required open screen area can be
obtained by using a screen of appropriate diameter with a maximum amount of open
screen area.

6.3 FILTERS. Filters are usually 3 to 5 inches thick for wellpoints and 6 to 8 inches
thick for large-diameter wells. To prevent infiltration of the aquifer materials into the
filter and of filter materials into the well or wellpoint, without excessive head losses,
filters should meet the following criteria:

©2015 J. Paul Guyer Page 26 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

If the filter is to be tremied in around the screen for a well or wellpoint, it may be either
uniformly or rather widely graded; however, if the filter is not tremied into place, it
should be quite uniformly graded and poured in around the well in a heavy continuous
stream to minimize segregation.

©2015 J. Paul Guyer Page 27 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 6
Typical design of a filter for a well or wellpoint

6.4 EFFECTIVE WELL RADIUS. The “effective” radius r w of a well is that well radius
which would have no hydraulic entrance loss H w. If well entrance losses are
considered separately in the design of a well or system of wells, r w for a well or
wellpoint without a filter may be considered to be one-half the outside diameter of the
well screens; where a filter has been placed around a wellpoint or well screen, rw may
generally be considered to be one-half the outside diameter or the radius of the filter.

6.5 WELL PENETRATION. In a stratified aquifer, the effective well penetration usually
differs from that computed from the ratio of the length of well screen to total thickness
of the aquifer.

6.6 SCREEN LENGTH, PENETRATION, AND DIAMETER. The length and

penetration of the screen depends on the thickness and stratification of the strata to be
dewatered. The length and diameter of the screen and the area of perforations should
be sufficient to permit the inflow of water without exceeding the entrance velocity given

©2015 J. Paul Guyer Page 28 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

above. The “wetted screen length h ws” (or hw for each stratum to be dewatered) is
equal to or greater than Q w/qc . The diameter of the well screen should be at least 3 to
4 inches larger than the pump bowl or motor.

©2015 J. Paul Guyer Page 29 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

7. PUMPS, HEADERS, AND DISCHARGE PIPES. The capacity of pumps and piping
should allow for possible reduction in efficiency because of incrustation or mechanical
wear caused by prolonged operation. This equipment should also be designed with
appropriate valves, crossovers, and standby units so that the system can operate
continuously, regardless of interruption for routine maintenance or breakdown.

7.1 CENTRIFUGAL AND WELLPOINT PUMPS.

7.1.1 CENTRIFUGAL PUMPS CAN be used as sump pumps, jet pumps, or in

combination with an auxiliary vacuum pump as a wellpoint pump. The selection of a
pump and power unit depends on the discharge, suction lift, hydraulic head losses,
including velocity head and discharge head, air-handling requirement, power available,
fuel economy, and durability of unit. A wellpoint pump, consisting of a self-priming
centrifugal pump with an attached auxiliary vacuum pump, should have adequate air-
handling capacity and be capable of producing a vacuum of at least 22 to 25 feet of
water in the headers. The suction lift of a wellpoint pump is dependent on the vacuum
available at the pump bowl, and the required vacuum must be considered in
determining the pumping capacity of the pump. Characteristics of a typical wellpoint
pump vacuum unit are shown in figure 7. Sump pumps of the centrifugal type should
be self-priming and capable of developing at least 20 feet of vacuum. Jet pumps are
high head pumps; typical characteristics of a typical 6-inch jet pump are shown in
figure 8.

©2015 J. Paul Guyer Page 30 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 7
Characteristics of typical vacuum unit for wellpoint pumps

©2015 J. Paul Guyer Page 31 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 8
Characteristics of 6-inch jet pump

7.1.2 EACH WELLPOINT PUMP SHOULD be provided with one connected standby
pump so as to ensure continuity of operation in event of pump or engine failure, or for
repair or maintenance. By overdesigning the header pipe system and proper
placement of valves, it may be possible to install only one standby pump for every two
operational pumps. If electric motors are used for powering the normally operating
pumps, the standby pumps should be powered with diesel, natural or LP gas, or
gasoline engines. The type of power selected will depend on the power facilities at the
site and the economics of installation, operation, and maintenance. It is also advisable
to have spare power units on site in addition to the standby pumping units. Automatic
switches, starters, and valves may be required if failure of the system is critical.

7.2 DEEP-WELL PUMPS.

7.2.1 DEEP-WELL TURBINE OR SUBMERSIBLE PUMPS are generally used to

pump large-diameter deep wells and consist of one or more stages of impellers on a

©2015 J. Paul Guyer Page 32 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

vertical shaft. Turbine pumps can also be used as sump pumps, but adequate stilling
basins and trash racks are required to assure that the pumps do not become clogged.
Motors of most large-capacity turbine pumps used in deep wells are mounted at the
ground surface. Submersible pumps are usually used for pumping deep, low-capacity
wells, particularly if a vacuum is required in the well.

7.2.2 IN THE DESIGN OF DEEP-WELL pumps, consideration must be given to

required capacity, size of well screen and riser pipe, total pumping head, and the
lowered elevation of the water in the well. The diameter of the pump bowl must be
determined before the wells are installed, as the inside diameter of the well casing
should be at least 3 to 4 inches larger in diameter than the pump bowl. Approximate
capacities of various turbine pumps are presented in table 2.

7.2.3 SUBMERSIBLE PUMPS REQUIRE either electric power from a commercial

source or one or more motor generators. If commercial power is used, 75 to 100
percent of (connected) motor generator power, with automatic starters unless
operational personnel are on duty at all times, should be provided as standby for the
commercial power. Spare submersible pumps, generally 10 to 20 percent of the
number of operating pumps, as well as spare starters, switches, heaters, and fuses,
should also be kept at the site.

7.2.4 DEEP-WELL TURBINE PUMPS can be powered with either electric motors or
diesel engines and gear drives. Where electric motors are used, 50 to 100 percent of
the pumps should be equipped with combination gear drives connected to diesel
(standby) engines. The number of pumps so equipped would depend upon the
criticality of the dewatering or pressure relief needs. Motor generators may also be
used as standby for commercial power. For some excavations and subsurface
conditions, automatic starters may be required for the diesel engines or motor
generators being used as backup for commercial power.

©2015 J. Paul Guyer Page 33 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

7.3 TURBOVACUUM PUMPS. For some wellpoint systems requiring high pumping
rates, it may be desirable to connect the header pipe to a 30- or 36-inch collection tank
about 20 or 30 feet deep, sealed at the bottom and top, and pump the flow into the
tank with a high-capacity deepwell turbine pump using a separate vacuum pump
connected to the top of the tank to produce the necessary vacuum in the header pipe
for the wells or wellpoints.

7.4 HEADER PIPE.

7.4.1 HYDRAULIC HEAD LOSSES caused by flow through the header pipe, reducers,
tees, fittings, and valves should be computed and kept to a minimum (1 to 3 feet) by
using large enough pipe. These losses can be computed from equivalent pipe lengths
for various fittings and curves.

7.4.2 WELLPOINT HEADER PIPES should be installed as close as practical to the

prevailing groundwater elevation and in accessible locations. Wellpoint pumps should
be centrally located so that head losses to the ends of the system are balanced and as
low as possible. If suction lift is critical, the pump should be placed low enough so that
the pump suction is level with the header, thereby achieving a maximum vacuum in
the header and the wellpoints. If construction is to be performed in stages, sufficient
valves should be provided in the header to permit addition or removal of portions of the
system without interrupting operation of the remainder of the system. Valves should
also be located to permit isolation of a portion of the system in case construction
operations should break a swing connection or rupture a header.

7.4.3 DISCHARGE LINES should be sized so that the head losses do not create
excessive back pressure on the pump. Ditches may be used to carry the water from
the construction site, but they should be located well back of the excavation and
should be reasonably watertight.

©2015 J. Paul Guyer Page 34 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

8. FACTORS OF SAFETY.

8.1 GENERAL. The stability of soil in areas of seepage emergence is critical in the
control of seepage. The exit gradient at the toe of a slope or in the bottom of an
excavation must not exceed that which will cause surface raveling or sloughing of the
slope, piping, or heave of the bottom of the excavation.

8.2 UPLIFT. Before attempting to control seepage, an analysis should be made to

ensure that the seepage or uplift gradient is equal to or less than that computed from
the following equations:

In stratified subsurface soils, such as a course-grained pervious stratum overlain by a

finer grained stratum of relatively low permeability, most of the head loss through the
entire section would probably occur through the finer grained material. Consequently,
a factor of safety based on the head loss through the top stratum would probably
indicate a more critical condition than if the factor of safety was computed from the

©2015 J. Paul Guyer Page 35 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

total head loss through the entire section. Also, when gradients in anisotropic soils are
determined from flow nets, the distance over which the head is lost must be obtained
from the true section rather than the transformed section.

8.3 PIPING. Piping cannot be analyzed by any rational method. In a study of piping
beneath hydraulic structures founded on granular soils, it was recommended that the
(weighted) creep ratio Cw should equal or exceed the values shown in the technical
literature for various types of granular soils.

where H - he represents vertical distance from the groundwater table to the bottom of
the excavation. These criteria for piping are probably only applicable to dewatering of
sheeted, cellular, or earth-dike cofferdams founded on granular soils. Once piping
develops, erosion of the soil may accelerate rapidly. As the length of seepage flow is
reduced, the hydraulic gradient and seepage velocity increase, with a resultant
acceleration in piping and erosion. Piping can be controlled by lowering the
groundwater table in the excavated slopes or bottom of an excavation, or in either less
critical situations or emergencies by placement of filters over the seepage exit surface
to prevent erosion of the soil but still permit free flow of the seepage. The gradation of
the filter material should be such that the permeability is high compared with the
aquifer, yet fine enough that aquifer materials will not migrate into or through the filter.
More than one layer of filter material may be required to stabilize a seeping slope or
bottom of an excavation in order to meet these criteria.

8.4 DEWATERING SYSTEMS. As in the design of any works, the design of a

dewatering system should include a factor of safety to cover the variations in
characteristics of the subsurface soils, stratification, and groundwater table; the
incompleteness of the data and accuracy of the formulations on which the design is

©2015 J. Paul Guyer Page 36 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

based; the reduction in the efficiency of the dewatering system with time; the frailties of
machines and operating personnel; and the criticality of failure of the system with
regard to safety, economics, and damage to the project. All of these factors should be
considered in selecting the factor of safety. The less information on which the design is
based and the more critical the dewatering is to the success of the project, the higher
the required factor of safety. Suggested factors of safety and design procedures are as
follows:

(1) Select or determine the design parameters as accurately as possible from existing
information.
(2) Use applicable design procedures and equations set forth in this manual.
(3) Consider the above enumerated factors in selecting a factor of safety.
(4) Evaluate the experience of the designer.
(5) After having considered steps l-4, the following factors of safety are considered
appropriate for modifying computed design values for flow, drawdown, well spacing,
and required “wetted screen length.”

In addition to these factors of safety being applied to design features of the system,
the system should be pump-tested to verify its adequacy for the maximum required
groundwater lowering and maximum river or groundwater table likely (normally a
frequency of occurrence of once in 5 to 10 years for the period of exposure) to occur.

©2015 J. Paul Guyer Page 37 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

©2015 J. Paul Guyer Page 38 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

9. DEWATERING OPEN EXCAVATIONS. An excavation can be dewatered or the

artesian pressure relieved by one or a combination of methods. The design of
dewatering and groundwater control systems for open excavations, shafts, and tunnels
is discussed in the following paragraphs.

9.1 TRENCHING AND SUMP PUMPING.

9.1.1 THE APPLICABILITY OF TRENCHES and sump pumping for dewatering an

open excavation is discussed in the technical literature. Where soil conditions and the
depth of an excavation below the water table permit trenching and sump pumping of
seepage, the rate of flow into the excavation can be estimated from plan and sectional
flow analyses or formulas presented.

9.1.2 WHERE AN EXCAVATION EXTENDS into rock and there is a substantial inflow
of seepage, perimeter drains can be installed at the foundation level outside of the
formwork for a structure. The perimeter drainage system should be connected to a
sump sealed off from the rest of the area to be concreted, and the seepage water
pumped out. After construction, the drainage system should be grouted. Excessive
hydrostatic pressures in the rock mass endangering the stability of the excavated face
can be relieved by drilling 4-inchdiameter horizontal drain holes into the rock at
approximately lo-foot centers. For large seepage inflow, supplementary vertical holes
for deep-well pumps at 50- to 100-foot intervals may be desirable for temporary
lowering of the groundwater level to provide suitable conditions for concrete
placement.

9.2 WELLPOINT SYSTEM. The design of a line or ring of wellpoints pumped with
either a conventional wellpoint pump or jet-eductors is generally based on
mathematical or flow-net analysis of flow and drawdown to a continuous slot.

9.2.1 CONVENTIONAL WELLPOINT SYSTEM. The drawdown attainable per stage of

wellpoints (about 15 feet) is limited by the vacuum that can be developed by the pump,

©2015 J. Paul Guyer Page 39 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

the height of the pump above the header pipe, and hydraulic head losses in the
wellpoint and collector system. Where two or more stages of wellpoints are required, it
is customary to design each stage so that it is capable of producing the total
drawdown required by that stage with none of the upper stages functioning. However,
the upper stages are generally left in so that they can be pumped in the event pumping
of the bottom stage of wellpoints does not lower the water table below the excavation
slope because of stratification, and so that they can be pumped during backfilling
operations.

(a) The design of a conventional wellpoint system to dewater an open excavation, as

discussed, is outlined below.

Step 1. Select dimensions and groundwater coefficients (H, L, and k) of the formation
to be dewatered based on investigations.
Step 2. Determine the drawdown required to dewater the excavation or to dewater
down to the next stage of wellpoints, based on the maximum groundwater level
expected during the period of operation.
Step 3. Compute the head at the assumed slot (he or ho) to produce the desired
residual head hD in the excavation.
Step 4. Compute the flow per lineal foot of drainage system to the slot Q P.
Step 5. Assume a wellpoint spacing a and compute the flow per wellpoint, Q W = aQp.
Step 6. Calculate the required head at the wellpoint h w corresponding to Q w.
Step 7. Check to see if the suction lift that can be produced by the wellpoint pump V
will lower the water level in the wellpoint to h,(p) as follows:

where

V= vacuum at pump intake, feet of water

M = distance from base of pervious strata to pump intake, feet

©2015 J. Paul Guyer Page 40 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

HC = average head loss in header pipe from wellpoint, feet

HW = head loss in wellpoint, riser pipe, and swing connection to header pipe, feet
Step 8. Set the top of the wellpoint screen at least 1 to 2 feet or more below hw-Hw to
provide adequate submergence of the wellpoint so that air will not be pulled into the
system.

(b) If an excavation extends below an aquifer into an underlying impermeable soil or

rock formation, some seepage will pass between the wellpoints at the lower boundary
of the aquifer. This seepage may be intercepted with ditches or drains inside the
excavation and removed by sump pumps. If the underlying stratum is a clay, the
wellpoints may be installed in holes drilled about 1 to 2 feet into the clay and backfilled
with filter material. By this procedure, the water level at the wellpoints can be
maintained near the bottom of the aquifer, and thus seepage passing between the
wellpoints will be minimized. Sometimes these procedures are ineffective, and a small
dike in the excavation just inside the toe of the excavation may be required to prevent
seepage from entering the work area. Sump pumping can be used to remove water
from within the diked area.

9.2.2 JET-EDUCTOR (WELL OR) WELLPOINT SYSTEMS. Flow and drawdown to a

jet-eductor (well or) wellpoint system can be computed or analyzed as discussed. Jet-
eductor dewatering systems can be designed as follows:

Step 1. Assume the line or ring of wells or wellpoints to be a drainage slot.

Step 2. Compute the total flow to the system for the required drawdown and
penetration of the well screens.
Step 3. Assume a well or wellpoint spacing that will result in a reasonable flow for the
well or wellpoint and jet-eductor pump.
Step 4. Compute the head at the well or wellpoint h w required to achieve the desired
drawdown.
Step 5. Set eductor pump at M = h w-H w with some allowance for future loss of well
efficiency.

©2015 J. Paul Guyer Page 41 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(a) If the soil formation being drained is stratified and an appreciable flow of water
must be drained down through the filter around the riser pipe to the wellpoint, the
spacing of the wellpoints and the permeability of the filter must be such that the flow
from formations above the wellpoints does not exceed

Substitution of small diameter well screens for wellpoints may be indicated for stratified
formations. Where a formation is stratified or there is little available submergence for
the wellpoints, jet-eductor wellpoints and risers should be provided with a pervious
filter, and the wellpoints set at least 10 feet back from the edge of a vertical
excavation.

(b) Jet-eductor pumps may be powered with individual small high-pressure centrifugal
pumps or with one or two large pumps pumping into a single pressure pipe furnishing
water to each eductor with a single return header. With a single-pump setup, the water
is usually circulated through a stilling tank with an overflow for the flow from wells or
wellpoints. Design of jet eductors must consider the static lift from the wells or
wellpoints to the water level in the recirculation tank; head loss in the return riser pipe;
head loss in the return header; and flow from the wellpoint. The (net) capacity of a jet-
eductor pump depends on the pressure head, input flow, and diameter of the jet
nozzle in the pump. Generally, a jet-eductor pump requires an input flow of about 2 to
2½ times the flow to be pumped depending on the operating pressure and design of
the nozzle. Consequently, if flow from the wells or wellpoints is large, a deep-well
system will be more appropriate. The pressure header supplying a system of jet

©2015 J. Paul Guyer Page 42 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

eductors must be of such size that a fairly uniform pressure is applied to all of the
eductors.

9.2.3 VACUUM WELLPOINT SYSTEM. Vacuum wellpoint systems for dewatering

fine-grained soils are similar to conventional wellpoint systems except the wellpoint
and riser are surrounded with filter sand that is sealed at the top, and additional
vacuum pump capacity is provided to ensure development of the maximum vacuum in
the wellpoint and filter regardless of air loss. In order to obtain 8 feet of vacuum in a
wellpoint and filter column, with a pump capable of maintaining a 25 foot vacuum in
the header, the maximum lift is 25 – 8 = 17 feet. Where a vacuum type of wellpoint
system is required, the pump capacity is small. The capacity of the vacuum pump will
depend on the air permeability of the soil, the vacuum to be maintained in the filter, the
proximity of the wellpoints to the excavation, the effectiveness of the seal at the top of
the filter, and the number of wellpoints being pumped. In very fine-g-rained soils,
pumping must be continuous. The flow may be so small that water must be added to
the system to cool the pump properly.

9.3 ELECTROOSMOSIS.

9.3.1 AN ELECTROOSMOTIC DEWATERING SYSTEM consists of anodes (positive

electrodes, usually a pipe or rod) and cathodes (negative electrodes, usually
wellpoints or small wells installed with a surrounding filter), across which a d-c voltage
is applied. The depth of the electrodes should be at least 5 feet below the bottom of
the slope to be stabilized. The spacing and arrangement of the electrodes may vary,
depending on the dimensions of the slope to be stabilized and the voltage available at
the site. Cathode spacings of 25 to 40 feet have been used, with the anodes installed
midway between the cathodes. Electrical gradients of 1.5- to 4-volts-per-foot distance
between electrodes have been successful in electroosmotic stabilization. The electrical
gradient should be less than about 15 volts per foot of distance between electrodes for
long-term installations to prevent loss in efficiency due to heating the ground. Applied

©2015 J. Paul Guyer Page 43 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

voltages of 30 to 100 volts are usually satisfactory; a low voltage is usually sufficient if
the groundwater has a high mineral content.

9.3.2 THE DISCHARGE OF A CATHODE WELLPOINT may be estimated from the

equation:

Current requirements commonly range between 15 and 30 amperes per well, and
power requirements are generally high. However, regardless of the expense of
installation and operation of an electroosmotic dewatering system, it may be the only
effective means of dewatering and stabilizing certain silts, clayey silts, and clayey silty
sands. Electroosmosis may not be applicable to saline soils because of high current
requirements, nor to organic soils because of environmentally objectionable effluents,
which may be unsightly and have exceptionally high pH values.

9.4 DEEP-WELL SYSTEMS.

9.4.1 THE DESIGN AND ANALYSIS of a deep-well system to dewater an excavation

depends upon the configuration of the site dewatered, source of seepage, type of flow
(artesian and gravity), penetration of the wells, and the submergence available for the
well screens with the required drawdown at the .wells. Flow and drawdown to wells
can be computed or analyzed as discussed

©2015 J. Paul Guyer Page 44 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

9.4.2 METHODS ARE PRESENTED whereby the flow and drawdown to a well system
can be computed either by analysis or by a flow net assuming a continuous slot to
represent the array of wells, and the drawdown at and between wells computed for the
actual well spacing and location.

9.4.3 THE SUBMERGED length and size of a well screen should be checked to
ensure that the design flow per well can be achieved without excessive screen
entrance losses or velocities. The pump intake should be set so that adequate
submergence (a minimum of 2 to 5 feet) is provided when all wells are being pumped.
Where the type of seepage (artesian and gravity) is not well established during the
design phase, the pump intake should be set 5 to 10 feet below the design elevation to
ensure adequate submergence. Setting the pump bowl below the expected drawdown
level will also facilitate drawdown measurements.

9.5 COMBINED SYSTEMS.

9.5.1 WELL AND WELLPOINT SYSTEMS. A dewatering system composed of both

deep wells and wellpoints may be appropriate where the groundwater table has to be
lowered appreciably and near to the top of an impermeable stratum. A wellpoint
system alone would require several stages of wellpoints to do the job, and a well
system alone would not be capable of lowering the groundwater completely to the
bottom of the aquifer. A combination of deep wells and a single stage of wellpoints
may permit lowering to the desired level. The advantages of a combined system, in
which wells are essentially used in place of the upper stages of wellpoints, are as
follows:

(a) The excavation quantity is reduced by the elimination of berms for installation and
operation of the upper stages of wellpoints.
(b) The excavation can be started without a delay to install the upper stages of
wellpoints.

©2015 J. Paul Guyer Page 45 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(c) The deep wells installed at the top of the excavation will serve not only to lower the
groundwater to permit installation of the wellpoint system but also to intercept a
significant amount of seepage and thus reduce the flow to the single stage of
wellpoints.

9.5.2 SAND DRAINS with deep wells and wellpoints. Sand drains can be used to
intercept horizontal seepage from stratified deposits and conduct the water vertically
downward into a pervious stratum that can be dewatered by means of wells or
wellpoints. The limiting feature of dewatering by sand drains is usually the vertical
permeability of the sand drains itself, which restricts this method of drainage to soils of
low permeability that yield only a small flow of water. Sand drains must be designed so
that they will intercept the seepage flow and have adequate capacity to allow the
seepage to drain downward without any back pressure. To accomplish this, the drains
must be spaced, have a diameter, and be filled with filter sand so that:

Generally, sand drains are spaced on 5- to 15-foot centers and have a diameter of 10
to 18 inches. The maximum permeability kv of a filter that may be used to drain soils
for which sand drains are applicable is about 1000 to 3000 x 10-4 centimeters per
second or 0.20 to 0.60 feet per minute, Thus, the maximum capacity Q D of a sand
drain is about 1 to 3 gallons per minute. The capacity of sand drains can be
significantly increased by installing a small (1- or 1½-inch) slotted PVC pipe in the
drain to conduct seepage into the drain downward into underlying more pervious strata
being dewatered.

9.6 PRESSURE RELIEF SYSTEMS.

©2015 J. Paul Guyer Page 46 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

9.6.1 TEMPORARY RELIEF of artesian pressure beneath an open excavation is

required during construction where the stability of the bottom of the excavation is
endangered by artesian pressures in an underlying aquifer. Complete relief of the
artesian pressures to a level below the bottom of the excavation is not always required
depending on the thickness, uniformity, and permeability of the materials. For uniform
tight shales or clays, an upward seepage gradient i as high as 0.5 to 0.6 may be safe,
but clay silts or silts generally require lowering the groundwater 5 to 10 feet below the
bottom of the excavation to provide a dry, stable work area.

9.6.2 THE FLOW TO A PRESSURE RELIEF SYSTEM is artesian; therefore, such a

system may be designed or evaluated on the basis of the methods for artesian flow.
The penetration of the wells or wellpoints need be no more than that required to
achieve the required drawdown to keep the flow to the system a minimum. If the
aquifer is stratified and anisotropic, the penetration required should be determined by
computing the effective penetration into the transformed aquifer as described in
appendix E.

9.7 CUTOFFS. Seepage cutoffs are used as barriers to flow in highly permeable
aquifers in which the quantity of seepage would be too great to handle with deepwell
or wellpoint dewatering systems alone, or when pumping costs would be large and a
cutoff is more economical. The cutoff should be located far enough back of the
excavation slope to ensure that the hydrostatic pressure behind the cutoff does not
endanger the stability of the slope. If possible, a cutoff should penetrate several feet
into an underlying impermeable stratum. However, the depth of the aquifer or other
conditions may preclude full penetration of the cutoff, in which case seepage beneath
the cutoff must be considered. Figure 9 illustrates the effectiveness of a partial cutoff
for various penetrations into an aquifer. The figure also shows the soils to be
homogeneous and isotropic with respect to permeability. If, however, the soils are
stratified or anisotropic with respect to permeability, they must be transformed into an

©2015 J. Paul Guyer Page 47 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

isotropic section and the equivalent penetration computed by the method given in
appendix E before the curves shown in the figure are applicable.

9.7.1 CEMENT AND CHEMICAL GROUT CURTAIN. Pressure injection of grout into a
soil or rock may be used to reduce the permeability of the formation in a zone and seal
off the flow of water. The purpose of the injection of grout is to fill the void spaces with
cement or chemicals and thus form a solid mass through which no water can flow.
Portland cement, fly ash, bentonite, and sodium silicate are commonly used as grout
materials. Generally, grouting pressures should not exceed about 1 pound per square
inch per foot of depth of the injection.

(a) Portland cement is best adapted to filling voids and fractures in rock and has the
advantage of appreciably strengthening the formation, but it is ineffective in
penetrating the voids of sand with an effective grain size of 1 millimeter or less. To
overcome this deficiency, chemical grouts have been developed that have nearly the
viscosity of water, when mixed and injected, and later react to form a gel which seals
the formation Chemical grouts can be injected effectively into soils with an effective
grain size D10 that is less than 0.1 millimetre. Cement grout normally requires a day or
two to hydrate and set, whereas chemical grout can be mixed to gel in a few minutes.

(b) Cement grouts are commonly mixed at water- cement ratios of from 5:1 to 10:1
depending on the grain size of the soils. However, the use of a high water- cement
ratio will result in greater shrinkage of the cement, so it is desirable to use as little
water as practical. Bentonite and screened fly ash may be added to a cement grout to
both improve the workability and reduce the shrinkage of the cement. The setting time
of a cement grout can be accelerated by using a 1:1 mixture of gypsum-base plaster
and cement or by adding not more than 3 percent calcium chloride. High-earlystrength
cement can be used when a short set time is required.

(c) Chemical grouts, both liquid and powderbased, are diluted with water for injection,
with the proportions of the chemicals and admixtures varied to control the gel time.

©2015 J. Paul Guyer Page 48 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(d) Injection patterns and techniques vary with grout materials, character of the
formation, and geometry of the grout curtains. (Grout holes are generally spaced on 2-
to 5-foot centers.) Grout curtains may be formed by successively regrouting an area at
reduced spacings until the curtain becomes tight. Grouting is usually done from the top
of the formation downward.

(e) The most perplexing problem connected with grouting is the uncertainty about
continuity and effectiveness of the seal. Grout injected under pressure will move in the
direction of least resistance. If, for example, a sand deposit contains a layer of gravel,
the gravel may take all the grout injected while the sand remains untreated. Injection
until the grout take diminishes is not an entirely satisfactory measure of the success of
a grouting operation. The grout may block the injection hole or penetrate the formation
only a short distance, resulting in a discontinuous and ineffective grout curtain. The
success of a grouting operation is difficult to evaluate before the curtain is complete
and in operation, and a considerable construction delay can result if the grout curtain
is not effective. A single row of grout holes is relatively ineffective for cutoff purposes
compared with an effectiveness of 2 or 3 times that of overlapping grout holes.

9.7.2 SLURRY WALLS. The principal features of design of a slurry cutoff wall include:
viscosity of slurry used for excavation; specific gravity or density of slurry; and height
of slurry in trench above the groundwater table. The specific gravity of the slurry and
its level above the groundwater table must be high enough to ensure that the
hydrostatic pressure exerted by the slurry will prevent caving of the sides of the trench
and yet not limit operation of the excavating equipment. Neither shall the slurry be so
viscous that the backfill will not move down through the slurry mix. Typical values of
specific gravity of slurries used range from about 1.1 to 1.3 (70 to 80 pounds per cubic
foot) with sand or weighting material added: The viscosity of the slurry for excavating
slurry wall trenches usually ranges from a Marsh funnel reading of 65 to 90 seconds,
as required to hold any weighting material added and to prevent any significant loss of
slurry into the walls of the trench. The slurry should create a pressure in the trench

©2015 J. Paul Guyer Page 49 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

approximately equal to 1.2 times the active earth pressure of the surrounding soil.
Where the soil at the surface is loose or friable, the upper part of the trench is
sometimes supported with sand bags or a concrete wall. The backfill usually consists
of a mixture of soil (or a graded mix of sand-gravel-clay) and bentonite slurry with a
slump of 4 to 6 inches.

©2015 J. Paul Guyer Page 50 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

Figure 9
Flow beneath a partially penetrating cutoff wall.

©2015 J. Paul Guyer Page 51 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

9.7.3 STEEL SHEET PILING. Seepage cutoffs may be created by driving a sheet pile
wall or cells to isolate an excavation in a river or below the water table. Sheet piles
have the advantage of being commonly available and readily installed. However, if the
soil contains cobbles or boulders, a situation in which a cutoff wall is applicable to
dewatering, the driving may be very difficult and full penetration may not be attained.
Also, obstructions may cause the interlocks of the piling to split, resulting in only a
partial cutoff.

(a) Seepage through the sheet pile interlocks should be expected but is difficult to
estimate. As an approximation, the seepage through a steel sheet pile wall should be
assumed equal to at least 0.01 gallon per square foot of wall per foot of net head
acting on the wall. The efficiency of a sheet pile cutoff is substantial for short paths of
seepage but is small or negligible for long paths.

(b) Sheet pile cutoffs that are installed for longterm operation will usually tighten up
with time as the interlocks become clogged with rust and possible incrustation by the
groundwater.

9.7.4 FREEZING. Freezing the water in saturated porous soils or rock to form an ice
cutoff to the flow of groundwater may be applicable to control of groundwater for shafts
or tunnels where the excavation is small but deep.

©2015 J. Paul Guyer Page 52 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

10. DEWATERING SHAFTS AND TUNNELS.

10.1 THE REQUIREMENTS and design of systems for dewatering shafts and tunnels
in cohesionless, porous soil or rock are similar to those previously described for open
excavations. As an excavation for a shaft or tunnel is generally deep, and access is
limited, deepwells or jet-eductor wellpoints are considered the best method for
dewatering excavations for such structures where dewatering techniques can be used.
Grout curtains, slurry cutoff walls, and freezing may also be used to control
groundwater adjacent to shafts or tunnels.

10.2 WHERE THE SOIL OR ROCK FORMATION is reasonably homogeneous and

isotropic, a well or jet-eductor system can be designed to lower the water table below
the tunnel or bottom of the shaft using methods and formulas presented. If the soil or
rock formation is stratified, the wells must be screened and filtered through each
pervious stratum, as well as spaced sufficiently close so that the res idual head in each
stratum being drained is not more than 1 or 2 feet. Dewatering stratified soils
penetrated by a shaft or tunnel by means of deep wells may be facilitated by sealing
the wells and upper part of the riser pipe and applying a vacuum to the top of the well
and correspondingly to the filter. Maintenance of a vacuum in the wells and
surrounding earth tends to stabilize the earth and prevent the emergence of seepage
into the tunnel or shaft.

10.3 IN COMBINED WELL-VACUUM SYSTEMS, it is necessary to use pumps with a

capacity in excess of the maximum design flow so that the vacuum will be effective for
the full length of the well screen. Submersible pumps installed in sealed wells must be
designed for the static lift plus friction losses in the discharge pipe plus the vacuum to
be maintained in the well. The pumps must also be designed so that they will pump
water and a certain amount of air without cavitation. The required capacity of the
vacuum pump can be estimated from formulas for the flow of air through porous media
considering the maximum exposure of the tunnel facing or shaft wall at any one time to
be the most pervious formation encountered, assuming the porous stratum to be fully

©2015 J. Paul Guyer Page 53 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

drained. The flow of air through a porous medium, assuming an ideal gas flowing
under isothermal conditions, is given in the following formula:

©2015 J. Paul Guyer Page 54 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

15- to 20-foot centers have been used to dewater caissons and mine shafts 75 to 250
feet deep.

©2015 J. Paul Guyer Page 55 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

10.4 IN DESIGNING A WELL SYSTEM TO DEWATER a tunnel or shaft, it should be

assumed that any one well or pumping unit may go out of operation. Thus, any
combination of the other wells and pumping units must have sufficient capacity to
provide the required water table lowering or pressure relief. Where electrical power is
used to power the pumps being used to dewater a shaft or tunnel, a standby generator
should be connected to the system with automatic starting and transfer equipment or
switches.

©2015 J. Paul Guyer Page 56 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

11. PERMANENT PRESSURE RELIEF SYSTEMS. Permanent drainage or pressure

relief systems can be designed using equations and considerations previously
described for various groundwater and flow conditions. The well screen, collector
pipes, and filters should be designed for long service and with access provided for
inspection and reconditioning during the life of the project. Design of permanent relief
or drainage systems should also take into consideration potential encrustation and
screen loss. The system should preferably be designed to function as a gravity system
without mechanical or electrical pumping and control equipment. Any mechanical
equipment for the system should be selected for its simplicity and dependability of
operation. If pumping equipment and controls are required, auxiliary pump and power
units should be provided. Piezometers and flow measuring devices should be included
in the design to provide a means for controlling the operation and evaluating its
efficiency.

©2015 J. Paul Guyer Page 57 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

12. FREEZING.

12.1 GENERAL.

12.1.1 THE CONSTRUCTION of a temporary waterstop by artificially freezing the soil

surrounding an excavation site is a process that has been used for over a century, not
always with success and usually as a last resort when more conventional methods had
failed. The method may be costly and is time-consuming. Until recent years far too
little engineering design has been used, but nowadays a specialist in frozen-soil
engineering, given the site information he needs, can design a freezing system with
confidence. However, every job needs care in installation and operation and cannot be
left to a general contractor without expert help. A favorable site for artificial freezing is
where the water table is high, the soil is, e.g., a running sand, and the water table
cannot be drawn down because of possible damage to existing structures of water (in
a coarser granular material). The freezing technique may be the best way to control
water in some excavations, e.g., deep shafts.

12.1.2 FROZEN SOIL NOT ONLY is an effective water barrier but also can serve as
an excellent cofferdam. An example is the frozen cofferdam for an open excavation
220 feet in diameter and 100 feet deep in rubbishy fill, sands, silts, and decomposed
rock. A frozen curtain wall 4000 feet long and 65 feet deep has been successfully
made but only after some difficult problems had been solved. Mine shafts 18 feet in
diameter and 2000 feet deep have been excavated in artificially frozen soils and rocks
where no other method could be used. Any soil or fractured rock can be frozen below
the water table to form a watertight curtain provided the freeze-pipes can be installed,
but accurate site data are essential for satisfactory design and operation.

12.2 DESIGN. As with the design of any system for sub surface water control, a
thorough site study must first be made. Moving water is the factor most likely to cause
failure; a simple sounding-well or piezometer layout (or other means) must be used to
check this. If the water moves across the excavation at more than about 4 feet per

©2015 J. Paul Guyer Page 58 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

day, the designer must include extra provisions to reduce the velocity, or a curtain wall
may never close. If windows show up in the frozen curtain wall, flooding the excavation
and refreezing with added freeze-pipes are nearly always necessary. A knowledge of
the creep properties of the frozen soils may be needed; if the frozen soil is used as a
cofferdam or earth retaining structure, such can be determined from laboratory tests.
Thermal properties of the soils can usually be reliably estimated from published data,
using dry unit weight and water content.

12.3 OPERATION. The ground is frozen by closed-end, steel freeze-pipes (usually

vertical, but they can be driven, placed, or jacked at any angle) from 4 to 6 inches in
diameter, spaced from 3 to 5 feet in one or more rows to an impervious stratum. If
there is no impervious stratum within reach, the soil may be completely frozen as a
block in which the excavation is made, or an impervious stratum may be made
artificially. In one project, a horizontal disk about 200 feet across and 24 feet thick was
frozen at a minimum depth of 150 feet. Then, a cylindrical cofferdam 140 feet in
diameter was frozen down to the disk, and the enclosed soil was excavated without
any water problem.

(1) Coaxial with each freeze-pipe is a 1½ to 2-inch steel, or plastic, supply pipe
delivering a chilled liquid (coolant) to the bottom of the closed freeze pipe. The coolant
flows slowly up the annulus between the pipes, pulls heat from the ground, and
progressively freezes the soil. After a week or two, the separate cylinders of frozen soil
join to form the barrier, which gradually thickens to the designed amount, generally at
least 4 feet (walls of 24-foot thickness with two rows of freeze-pipes have been frozen
in large and deep excavations in soft organic silts). The total freeze-time varies from 3
to 4 weeks to 6 months or more but is predictable with high accuracy, and by
instrumentation and observation the engineer has good control. Sands of low water
content freeze fastest; fine-grained soils of high water content take more time and total
energy, although the refrigeration horsepower required may be greater than for sands.

©2015 J. Paul Guyer Page 59 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(2) The coolant is commonly a chloride brine at zero to -20 degrees Fahrenheit, but
lower temperatures are preferable for saving time, reducing the amount of heat to be
extracted, and minimizing frost heave effects (which must be studied beforehand). In
recent years, liquid propane at -45°F has been used in large projects, and for small
volumes of soil, liquid nitrogen that was allowed to waste has been used. (These
cryogenic liquids demand special care-they are dangerous.) Coolant circulation is by
headers, commonly 8-inch pipes, connected to a heat-exchanger at the refrigeration
plant using freon (in a modern plant) as the refrigerant. The refrigeration equipment is
usually rented for the job. A typical plant requires from 50 horsepower and up; 1000
horsepower or more has sometimes been used. Headers should be insulated and are
recoverable. Freeze-pipes may be withdrawn but are often wasted in construction;
they are sometimes used for thawing the soil back to normal, in which case they could
be pulled afterward.

12.4 IMPORTANT CONSIDERATIONS. The following items must be considered when

the freezing technique is to be used:

(1) Water movement in soil.

(2) Location of freeze-pipes. (The spacing of freeze-pipes should not exceed the
designed amount by more than 1 foot anywhere along the freeze wall.)

(3) Wall closure. (Freeze-pipes must be accurately located, and the temperature of the
soil to be frozen carefully monitored with thermocouples to ensure 100 percent closure
of the wall. Relief wells located at the center of a shaft may also be used to check the
progress of freezing. By periodically pumping these wells, the effectiveness of the ice
wall in sealing off seepage flow can be determined.)

(4) Frost-heave effects-deformations and pressures. (Relief wells may be used to

relieve pressures caused by expansion of frozen soil.)

©2015 J. Paul Guyer Page 60 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(5) Temperature effects on buried utilities.

(6) Insulation of aboveground piping.

(7) Control of surface water to prevent flow to thefreezing region.

(8) Coolant and ground temperatures. (By monitoring coolant and soil temperatures,
the efficiency of the freezing process can be improved.)

(9) Scheduling of operations to minimize lost time when freezing has been completed.

(10) Standby plant. (Interruption of coolant circulation may be serious. A standby plant
with its own prime movers is desirable so as to prevent any thaw. A c ontinuous
advance of the freezing front is not necessary so that standby plant capacity is much
less than that normally used.)

©2015 J. Paul Guyer Page 61 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

13. CONTROL OF SURFACE WATER.

13.1 RUNOFF OF SURFACE WATER from areas surrounding the excavation should
be prevented from entering the excavation by sloping the ground away from the
excavation or by the construction of dikes around the top of the excavation. Ditches
and dikes can be constructed on the slopes of an excavation to control the runoff of
water and reduce surface erosion. Runoff into slope ditches can be removed by
pumping from sumps installed in these ditches, or it can be carried in a pipe or lined
ditch to a central sump in the bottom of the excavation where it can be pumped out.
Dikes at the top of an excavation and on slopes should have at least 1 foot of
freeboard above the maximum elevation of water to be impounded and a crown width
of 3 to 5 feet with side slopes of 1V on 2-2.5H.

13.2 IN DESIGNING A DEWATERING system, provision must be made for collecting

and pumping out surface water flooded. Control of surface water within the diked area
will not only prevent interruption of the dewatering operation, which might seriously
impair the stability of the excavation, but also prevent damage to the construction
operations and minimize interruption of work. Surface water may be controlled by
dikes, ditches, sumps, and pumps; the excavation slope can be protected by seeding
or covering with fabric or asphalt. Items to be considered in the selection and design of
a surface water control system include the duration and season of construction, rainfall
frequency and intensity, size of the area, and character of surface soils.

13.3 THE MAGNITUDE OF THE rainstorm that should be used for design depends on
the geographical location, risk associated with damage to construction or the
dewatering system, and probability of occurrence during construction. The common
frequency of occurrence used to design surface water control sumps and pumps is a
once in 2-to 5-year rainfall. For critical projects, a frequency of occurrence of once in
10 years may be advisable
.

©2015 J. Paul Guyer Page 62 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

13.4 IMPOUNDING RUNOFF on excavation slopes is somewhat risky because any

overtopping of the dike could result in overtopping of all dikes at lower elevations with
resultant flooding of the excavation.

13.5 AMPLE ALLOWANCE for silting of ditches should be made to ensure that
adequate capacities are available throughout the duration of construction. The grades
of ditches should be fairly flat to prevent erosion. Sumps should be designed that will
minimize siltation and that can be readily cleaned. Water from sumps should not be
pumped into the main dewatering system.

13.6 THE PUMP AND STORAGE requirements for control of surface water within an
excavation can be estimated in the following manner:

Step 1. Select frequency of rainstorm for which pumps, ditches, and sumps are to be
designed.
Step 2. For selected frequency (e.g., once in 5 years), determine rainfall for 10-, 30-,
and 60-minute rainstorms at project site.
Step 3. Assuming instantaneous runoff, compute volume of runoff VR (for each
assumed rainstorm) into the excavation or from the drainage area into the excavation
from the equation:

where
c = coefficient of runoff
R = rainfall for assumed rainstorm, inches
A = area of excavation plus area of drainage into excavation, acres

©2015 J. Paul Guyer Page 63 of 64

www.PDHcenter.com PDHonline Course C802 www.PDHonline.org

(The value of c depends on relative porosity, character, and slope of the surface of the
drainage area. For impervious or saturated steep excavations, c values may be
assumed to range from 0.8 to 1.0.)
Step 4. Plot values of VR versus assumed duration of rainstorm.
Step 5. Plot pumpage rate of pump to be installed assuming pump is started at onset
of rain.

13.7 THE REQUIRED DITCH and sump storage volume is the (maximum) difference
between the accumulated runoff for the various assumed rainstorms and the amount
of water that the sump pump (or pumps) will remove during the same elapsed period
of rainfall. The capacity and layout of the ditches and sumps can be adjusted to
produce the optimum design with respect to the number, capacity, and location of the
sumps and pumps.

13.8 CONVERSELY, the required capacity of the pumps for pumping surface runoff
depends upon the volume of storage available in sumps, as well as the rate of runoff.
For example, if no storage is available, it would be necessary to pump the runoff at the
rate it enters the excavation to prevent flooding. This method usually is not practicable.
In large excavations, sumps should be provided where practicable to reduce the
required pumping capacity. The volume of sumps and their effect on pump size can be
determined graphically or can be estimated approximately from the following equation:

©2015 J. Paul Guyer Page 64 of 64