+

Flow to Wells
+
Basic Assumptions

 Aquifer bounded on the bottom

 Horizontal Geologic Formations (with infinite extent)

 The potentiometric surface is horizontal and is steady prior to pumping

 Any changes in potentiometric surfaces are due to pumping

 Aquifer is homogeneous and isotropic

 All flow is radial towards the well

 Groundwater flow is horizontal

 Darcy’s Law is valid

 Water has constant density and viscosity

 Wells are fully penetrating

 Pumping well has infinitesimal diameter and 100 efficiency

+
Radial Flow

 Two dimensional flow in a confined aquifer

¶ 2h ¶ 2h w S ¶ h
+ + =
¶ x 2 ¶ y2 T T ¶ t
 Radial Transform
æyö
r= x +y
2 2 q = tan -1 ç ÷
èxø
 Axisymmetric
¶
=0
¶q
 Radial Coords Flow Equations
¶ 2h 1 ¶ h w S ¶ h
+ + =
¶ r r ¶ r T T ¶t
2
+
How about steady state conditions?

 We will see how transient tests can be used to assess

parameters, but steady state can provide a lot of information

 Assumptions
 Pumping well is screened in aquifer being test only
 Observation wells are also screen in that aquifer only
 Both pumping and observation wells are screened throughout the
entire aquifer thickness
+
Steady State Confined Aquifer
(Thiem solution)
+
Steady State Confined Aquifer
(Thiem solution)
 Governing equation

dh Q dr Q dr
Q = qA = K 2p rb dh = =
dr 2p Kb r 2p T r

 Integrating between two wells at different radii

Q æ r2 ö
h2 - h1 = ln ç ÷
2p T è r1 ø

Q æ r2 ö
T= ln ç ÷
2p ( h2 - h1 ) è r1 ø
+
Example Problem

 Two observation wells are placed at a distance on 2 and 10

meters from a well.

 The well is pumped at a flow rate of 800 litres/minute

 The depth to water is 15m and 12 m for the first and second
well respectively.

 What is the transmissivity of the well.

 If the depth of the aquifer is 10m, what is the hydraulic

conductivity
+
Steady State Unconfined Aquifer
+
Steady State Unconfined Aquifer
+
Steady State Unconfined Aquifer
(Thiem solution)
 Governing equation

db Q dr
Q = qA = K 2p rb bdb =
dr 2pK r

 Integrating between two wells at different radii

Q æ r2 ö
b -b =
2 2
ln ç ÷
p K è r1 ø
2 1

Q æ r2 ö
K= ln ç ÷
p ( b 2 - b1 ) è r1 ø
2 2
+
Example Problem

 Two observation wells are placed at a distance of 3 and 8

meters from a well in an unconfined aquifer

 The well is pumped at a flow rate of 500 litres/minute

 The depth to water is 5m and 2 m for the first and second

well respectively.

 If the depth of the aquifer is 10m, what is the hydraulic

conductivity?
+
Transients

 Now we will consider transient (time-varying) solutions.

 For some real situations it will simply take too long to reach
steady state (equilibrium – let me hear it Brandon)
conditions.

 We can use transient analysis to infer aquifer properties,

including storativity (which is not possible from steady
measurements)
+
Flow in a Completely Confined
Aquifer – Theis’ Solution
 Additional Assumptions
 Aquifer is confined on top and bottom
 No recharge
 Aquifer is compressible and water is released instantaneously
 Well is pumped at a constant rate
+
Schematic Setup
+
What does data typically look like?
+
Governing Equations

 Flow Equation

¶ 2 h 1 ¶h S ¶h
2 + =
¶r r ¶r T ¶t

 Initial Condition
h(r,t = 0) = h0

 Boundary Conditions

¶h
h(r ® ¥, t) = h0 lim rw ®0 2p rw bK =Q
¶ r r=r
w
+
Alternatively
 Define drawdown s(r, t) = h0 - h(r,t)

 Flow Equation
¶ 2 s 1 ¶s S ¶s
2 + =
¶r r ¶r T ¶t

 Initial Condition s(r,t = 0) = 0

 Boundary Conditions
s(r ® ¥,t) = 0 ¶s
lim rw ®0 2p rw bK = -Q
¶ r r=r
w
+
To Solve – Boltzman
Transformation
 Define
r2 T
x= D=
4Dt S

 The flow equation and IC/BCs now become

d 2s ds
x 2 + (1+ x ) = 0 s (x ® ¥) = 0
dx dx
ds Q
limx ®0 x =-
dx 4p bK
+
Solution
ds
 Let u=
dx

 Our equation becomes

du
x + (1+ x) u = 0
dx

 Solution
Q e -x
u=-
4pT x
+
Solution

Q e -x
 Solution u=-
4pT x

 Recall ds
u=
dx

 Therefore ¥
Q e -a
s=
4pKb
ò a
da
r 2S
4 Kbt

This is called the exponential integral

and can be called from Matlab or any
good maths program
+
Data from a pumping test

Known Q, known radius

+

 Approximately (for large times)

Q æ æ r 2 S öö
s= ç -0.5722 - lnç ÷÷
4pKb è è 4Kbt øø

4p Kb æ 4Kbt ö æ Kbt ö
s = -0.5722 + ln ç 2 ÷ = 2.303log10 ç 2.246 2 ÷
Q è rS ø è r Sø

 Using the graph on previous page (two unknowns – straight

lines tells all) Time when line intersects x axis

Q Tt *
Kb = T = 0.1833 S = 2.246 2
Ds10 r
Change in drawdown over one decade on log scale
+
Alternatively

 We use the well function (tabulated in appendix 1 of the text

by Fetter)
Q r2S
s= W (u) u=
4p T 4Tt
+
Why use the tabulated function?
Plot drawdown vs time in log log
+ Overlay the Two and pick a match point (does not
have to be on the curves – usually 1,1 on well
function curve)

Q
T= W ( u)
4 p ( h - h0 )
Using the match point, apply
4Tut
S= 2
r
+
Example Problem (Matlab)
Solve Both Ways
 You are provided with the following pumping curve

 The flowrate is 10m^3/s. The radius of the well is 10m

 What is the transmissivity and storativity of the aquifer?

+
What if the aquifer is leaky?
 We include some form of flux from a confining layer
¶ 2 h 1 ¶ h ( h0 - h) K ' S ¶ h
+ - =
¶r r ¶r
2
Tb' T ¶t

OR
¶ 2 s 1 ¶ s sK ' S ¶ s
+ + =
¶ r r ¶ r Tb' T ¶ t
2
+
What if the aquifer is leaky?

 Solution is given in terms of a tabulated function

Q
s= W ( u, r / B)
4p T

where

r2S
u= B = Tb'/ K '
4Tt

W(u,r/B) is called the artesian well function and can be looked up

in Appendix 3 of Fetter
+
What does the Artesian Well
Function look like?
+
Unconfined Aquifers

 Assumptions

 Aquifer is unconfined
 Vadose zone has no influence on drawdown
 Water initially pumped comes from instantaneous release of
water from elastic storage
 Eventually water from from storage due to gravity drainage
 Drawdown is negligible relative to saturated aquifer thickness
 The specific yield is at least 10 times the elastic storativity
 The aquifer can be anisotropic
+
Unconfined Aquifers

 Again, a tabulated results exists (water table aquifer

pumping function - appendix 6)

Q
h 0 -h = W (uA , uB , G)
4p T

r2S r 2 Sy r 2 Kv
uA = uB = G= 2
4Tt 4Tt b Kh
+
What does this pumping function
look like?
+

 Graphical Methods with Match Point discussed in text book.

They work just fine and I recommend that those of you who
are interested read about them, but they are a little outdated
with current computer capabilities.
+
Slug Tests – The Poor Man’s
Alternative
 Pumping Tests are expensive for many many reasons (labor
costs, well drilling costs, equipment, etc.). Sometime one way
also not actually wish to extract water from an aquifer for fear
that it may be contaminated.

 Slug Tests (or their counterpart bail-down tests) are a cheap and
quick alternative

 A known quantity of water is quickly added or removed from a

well and the response of water level in the well is measured.

 Water does not have to be added – instead a slug of known

volume can be thrown in, displacing a known volume of water.

 Slug test responses can be overdamped or underdamped and

different and appropriate methods must be chosen to properly
analyse data.
+
Overdamped
Cooper-Bredehoeft-Papadopulos
+
Overdamped
Cooper-Bredehoeft-Papadopulos
 We can show that (F is tabulated in appendix 2 of book)

H
= F (h, m )
Where H0

h = Tt / rc2 m = rs2 S / rc2

rc – radius of casing rs – radius of screen

Overlay data to identify what values of eta and mu correspond to your

setup and

T = rc2 / t1 S = ( rc2m ) / rs2

t1 is the time where on the type curve Tt/rc2=1

+
Type Curves for Cooper-
Bredehoeft-Papadopulos
+
Sample Problem (Matlab)

 Consider the following dataset and using the CBP method

evaluate the Transmissivity and Storativity. Well and screen
radius is 5 cm.

T(s) H H0 H/H0
0.01 96.5 100 0.965
0.05 86 100 0.86
0.1 74.5 100 0.745
0.2 58 100 0.58
0.5 29 100 0.29
1 10.8 100 0.108
5 6.2 100 0.062
10 2.7 100 0.027
20 1.3 100 0.013
+
Data from Example
+
Overlay Graphics
+
Overlay Graphics
+
Overlay Graphics
+
Overlay Graphics
Tt/rc2=1 on type curve plot
+
Remove the Type Curve, but keep
vertical line

When overlaid on
Figure 5.19
We identify
mu=1e-6
(see figure
below)
t1=0.1

t1=0.1
+
Overdamped
Hvorslev Method
+
Overdamped
Hvorslev Method Interpretation
 If the length of the piezometer is more than 8 times the radius
of the well screen, i.e. Lc/R>8 then

r 2 ln ( Le / R)
K=
2Let37

K – hydraulic conductivity
r – radius of the well casing
R – radius of the well screen
Le – length of the well screen
t37 – time it take for the water level to rise or fall to 37% of the
initial change.
+
Example Problem
t(s) H H0 (cm) H/H0
0.5 100 0.9
1 100 0.82
2 100 0.67
4 100 0.37
5 100 0.25
7 100 0.2
8 100 0.165
9 100 0.135

R=0.1m (radius of screen)

r=0.1m (radius of casing
L=5m (length of well screen)
+
Data
+
+
Overdamped
Bouwer and Rice Method
+
Overdamped
Bouwer and Rice Interpretation
rc2 ln ( Re / R) 1 æ H 0 ö
K= ln ç ÷
2Le t è Ht ø
-1
é 1.1 A + B ln ë( h - Lw ) / Rû
é ù ù
ln ( Re / R) = ê + ú
êë ln ( Lw / R) Le / R úû

K – hydraulic conductivity
rc – radius of the well casing
R – radius of the gravel envelope
Re – effective radius over which hear is disipated
Le – length of screen over which water can enter
H0 – drawdown at time 0
Ht – drawdown at time t
t – time
Lw – distance from water table to the bottom of the bore hole
A,B – constants on figure 5.25

Valid for Lw<h (the saturated thickness of the aquifer)

+
Underdamped
Van der Kamp
+
Underdamped Van der Kamp
 Consider the following decay formula

H (t ) = H 0e-g t cos (wt )

 Transmissivity given by

T = c + a ln (T )

rc2 g / L g L=
g
c = -a ln éë0.79rs2 S g / L ùû a= d=
v 2 +g 2
8d g/L
T – transmissivity
rc – radius of well casing
rs – radius of well screen
S – storage coefficient
L – effective length of the water column
g – gravity
w – angular frequency
+ Steps in Van der Kamp

 Calculate angular frequency w

2p 2p
w= = Dt – time between succesive
peaks t2 - t1 Dt

 Calculate damping factor

ln éë H ( t1 ) / H ( t2 )ùû
g=
t2 - t1

 Solve implicit equation (iteratively)

T0 = c + a ln ( c)
T1 = c + a ln (T0 )
T2 = c + a ln (T1 )
Continue until converged
+
Example Problem
Interpret the following curve
rc=0.5 m and rs=0.5m

H(cm)
+
Cautions and Guidelines for Slug
Tests
 Skin Effects can yield underpredictions

 Geological Survey Guidelines

 Three or more slug tests should be performed on a given well
 Two or more different initial displacement should be used
 The slug should be introduced as instantaneously as possible
 Good data acquisition equipment should be used
 An observation well should be employed for storage estimation
 Analysis method should be consistent with site
 Study results carefully and reassess analysis method if necessary
 Appropriate well construction parameters should be used
+
Effects of Hydrogeologic
Boundaries