Scholars' Mine

Masters Theses Student Theses and Dissertations

1966

Numerical solution of water hammer analysis

including friction losses
Lih Hsun Weng

Follow this and additional works at: https://scholarsmine.mst.edu/masters_theses

Part of the Civil Engineering Commons
Department:

Recommended Citation
Weng, Lih Hsun, "Numerical solution of water hammer analysis including friction losses" (1966). Masters Theses. 2946.
https://scholarsmine.mst.edu/masters_theses/2946

This Thesis - Open Access is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in Masters Theses by an
authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution
requires the permission of the copyright holder. For more information, please contact scholarsmine@mst.edu.
 NUMERICAL SOLUTION OF WAT~ HAMMER

ANALYSIS INCLUDING FRICTION LOSSES

BY
LIH HSUN WENG - t1 ~t,.

A
THESIS
submitted to the faculty ot
THE UNIVERSITY OF MISSOURI AT ROLLA
in partial fulfillment of the requirements tor the
Degree of
MASTER OF SCIENCE IB CIVIL ENGINEERING
Rolla, Mia JOur!
1966

Approved by

.; /, -
tr~ \> (ktaaa (advisor) / ' . . ·>-z._ ..

rfCaL¥ ~
1237 7
 ii

ABSTRACT

The water hammer phenomena is a very ~portant problem

1n pipe lines and gate operations. There are many analyt-
ical solutions tor water hammer problems including friction
losses, but moat ot them use a linear approximation tor the
friction etteet that is different from the actual situation.
Also the existing methods tor solving water hammer problems
involve ditticult mathematical operations· that are impractical
and inaccurate.
In order to tind a more rapid method tor solving water
hammer problems including friction losses and to obtain
closer agreement with the actual condition, the problem
ia solved here using the basic partial differential equations
v1th the inclusion ot friction loasea through a mathematical
transformation. A digital computer program is used to solve
the resulting equations.
 iii

ACKNOWLEDGMENT

The author wishes to express his appreciation to

Professor Jerry R. Bayless, Professor Paul R. Munger, and
Dr. Larry E. Farmer for their assistance in many phases of
this thesis. Also, the author is grateful for the assis-
tance given to him by Professor Ralph E. Lee, Director of
the Computer Center of the University of Missouri at Rolla.
 iv

TABLE OF CONTENTS
Page
ABSTRACT ••••••••••••••••••••••••••••••••••••••••••••••• ii
ACKNOWBBDGMENT ••••••••••••••••••••••••••••••••••••••••• · 111
LIST OF FIGURES •••••••••••••••••••••••••••••••••••••••• v
I. INTRODUCTION •••••••••••••••••••••••••••••••••••• 1
II • REV! EW OF LITERA TUBE • • • • • • • • • • • • • • • • • • • • • • • • • • • • 4
III. WATER HAMMER EQUATIONS NEGLECTING FRICTION LOSSES.l4

(A) Assumptions and Notation ••••••••••••••••••• 14

(B) Basic Equations •••••••••••••••••••••••••••• 16
{C) Physical Meaning of Water Hammer Equations •• 19
IV. WATER HAMMER EQUATIONS INCLUDING FRICTION LOSSES. 22
(J.) Numerical Analysis For Water Hammer Equations
Including Friction Losses •••••••••••••••••• 22
(.B) Practical Examples ••••••••••••••••••••••••• 30
V. CONCLUSIONS ••••••••••••••••••••••••••••••••••••• 40
BIBLIOGRAPHY •• • • • • • • • • • • •••• • • • • • • • • • • • • • • • • • • • • • • • • • • • 42
APPENDIX A The Computer Program ••••••••••••••••••••••• 44
APPENDIX B Water Hammer Analysis by The Laplace-Mellin
Transformation ••••••••••••••••••••••••••••• ~9

VITA •• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • .53
 v

LIST OF FIGURES
Figure Page
1. Water Hammer Wave Cycle ••••••••••••••••••••••• 5
2. Variation of Water Hammer Pressure with respect
to TLme at the Valva End •••••••••.•••••••••••• 11
3. The Specified Time Interval ••••••••••••••••••• 12
4. Dlfinition Sketch ••••••••••••••••••••••••••••• 16
5. Equilibrium of Momentum for a Water Column •••• 17
6. Condition of Cont~uity ••••••••••••••••••••••• 19
1. Water Hammer Wave Propagation ••••••••••.•••••• 20
8. Physical Meaning of Water Hammer Equations •••• 21
9. Point Network Showing Finite Differences •••••• 23
10. The Right End Boundary Condition •••••••••••••• 25
11. The Right End Boundary Point Network •••••••••• 27
12. The Left End Boundary Point Network ••••••••••• 28
13. Computation Procedure ••••••••••••••••••••••••• 29
14. Computer Flow Chart ••••••••••••••••••••••••••• 31
15. Variation of Water Hammer Pressure with Time at
Valve End. (Slow Valve Closure) ••••••••••••••• 33
16. Variation of Water Hammer Velocity with Time at
Reservoir End and Valve End. (Slow Valve
Closure) • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 34
17. Variation of Water Hammer Pressure with Time at
Valve End (Rapid Valve Closure) ••••••••••••••• 36
18. Variation of Water Hammer Velocity with Time at
Reservoir End and Valve End. (Rapid Valve
Clos-qre) • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 37
 1

I. INTRODUCTION

Most of the water hammer problems that have been solved

do not include the effect of friction losses. The solution
is applicable to the many situations in which friction losses
are small when compared with the head changes. However, this
effect sometimes cannot be neglected as in cases of slow gate
closure, in a long pipe, or when the friction factor is high.
A few investigators have developed graphical methods, (l) *
whereby the effect of friction losses are approximated by
placing hypothetical obstructions at certain selected loca-
tions along the pipe line with friction losses of each sec-
tion concentrated at these friction joints. These obstruc-
tions have the same total head loss as the entire pipe line
and are a first approximation of the effect of friction
losses. These methods produce an abrupt drop in pressure
at each obstruction, which is not the true case.
For turbulent flow, the friction losses in a pipe line
are actually distributed along the pipe line, and are assumed
to vary with the square of velocity. Thus the eff ect of
these losses can be included in the water hammer equations
by the introduction of a v 2 term in the basic water hammer

equations.
The direct analytical solution of the basic partial
di f ferential equations used in this thesis is one of the
methods capable of distributing the etfect or friction losses

* Numbers in parentheses refer to bibliography.

 2

along the pipe line 1n agreement with the actual situation.

However, the erteet of friction losaes, which are assumed to
vary with v2, makes the basic differential equations non-
linear. The simultaneous solution or this set of equations
he'd been considered impossible. Thus, approximations, some
or which appeared as graphical methods, some as linear
approx~ations and so forth, have been u•ed as substitutions.
S~me investigators have developed methods of est~ating

the friction losses by operational mathematics or other

analytical solutions.{2) They all used the linearization or
the friction term. Because the effect of friction ia expressed
1n terms of v2 1n the water hammer basic ditterential equa-
tions, and because most eases in engineering practice are
turbulent tlows, the linearization or the friction term ean
not exactly represent the actual situation, and thus the
methods are not wholly satisfactory.
The direct solution of the basic partial differential
equations seems to have greater accuracy over other methods.
~eeause the v2 term was included in the equations, it is
believed to be in better accordance with nature.
This thesis presents a general solution or water hammer
equations. The theoretical equations can be solved by using
a numerical solution which is programmed tor a digital com-
puter, however, some approxtmations are usually required.
The computer program solves the equations by uaing apecitied
time intervals.
 3

Some or the data that are available at any t~e during

the computations are the pressure head and velocity at some
definite point along the pipe line. From these data, curves
ean be plotted and the velocity and pressure distributions
ean be observed. This makes it easy to understand the water
hammer phenomena.
Work on water hammer analysis by numerical method has
been done by Victor L. Streeter and Chinto Lai, (3) who
give a good approx~ate solution to the water hammer equa-
tions. Their work was helpful to the author 1n making this
studyo
 4
II. REVIEW OF LITERATURE

rr the velocity of water flowing in a pipe is changed,

for example by rapidly closing a valve, presaure waves will
traverse the portion of the pipe between the valve and the
reservoir-·. These pressure waves continue until they are
damped out by reflections and friction.
Referring to Fig. 1, a large reservoir discharges through
a· horizontal pipe of constant diameter and wall thickness,
with a valve or gate in the downstream end. Initially the
valve is open and the water flows in a stea-dy state with a
velocity V0 • If it is· assumed that the valve is closed in-
stantaneously and friction losses are neglected, ~ediately

after the valve closure a· pre•sure wave is formed in the

pipe between the valve and reservoir. A conception of the
progress of the pressure waves in a pipe may be obtained if
the phenomenon is considered 1n four phases which repeat
indefinitely and without diminution in intensity. The wa•e
has the following four phases in its· full cycle(4):
Phase I. Diagrams (b) to (d). The velocity of the
first layer of water and adjacent to the closed valve is
abruptl1 reduced from V0 to zero and the kinetic energy of
the water is converted into pressure ene:rrgy, which results:
in the elastic work of compress·ing the water and stretching

the pipe wall. As soon as the first layer is stopped, the

 5

w. s. Hydraulic Gradient Line

Reservoir Gate
Ho open
1 '----------~
____.,.. v
(a) Initial Steady State Condition

w. s. H,:draulic Gradient Line ----z

Ho Wave front
j V=O
l V=V

(b) Transient Condition € -Seconds After Gate Closure

Hydraulic Gradient Line ---z

Wave front
~

(o) Transient Conditions at t= ..1!..

a
-e Seconds

Fig. 1. Water Hammer Wave Cycle

 6

Hydraulic Gradient Line

(d) Transient Conditions at t= -i- Seconds

Hydraulic Gradient Line

We.ve front
f~
I
V=•V

(e) Transient Conditions at b=+ + 15 Seconds

w. s. Hydraulic Gradient Line

Wave front
H

1 V=-V
0

2L
(t) Transient Conditions at t=
-a- -e Seconds

Fig. 1. Water Hammer Wave Cycle

 7

w. s. Hydraulic Gradient Line

Wave front
V=-V0

(g) Transient Condi tiona at 'b= ~ ·_ Seconds

a

w. s. Hydraulic Gradient Line ~

Wave front
V=-V0 ~ V=O
1
(h) Transient Conditions at 'b= Aa +e Seconds

w. s.
Hydraulic Gradient Line
front
V=O

-
(i) Transient Conditions at t= ~ -€ Seconds
a

w. s. Wave front
Hydraulic Gradient Line 2

(j) Transient Conditions at t= 1t Seconds

Fig. 1. Water Hammer Wave Cycle
 8

w. s.
Hydraulic Gradient Line
Wave front

(k) Transient Condi tiona at 'b= 3L +6 Seconds

a

w. s. Hydraulic Gradient Line

1 Wave front
1---
V=V0 1 V=O

(l) Transient Conditions at 'b:: 4L -e Seconds

a

w. s.
Hydraulic Gradient Line

Wave front

(m) Transient Conditions at b= 4L Seconds

a

Fig. 1. Water Hammer Wave Cycle

e in the above diagrams is used to denote an interval or

time which is very small in comparison with the round trip
wave travel ttme.
 9

process is repeated with the next layer. The high pressure

moves upstream as a wave. At the end of this phase, when
the wave front reaches the upstream end of the pipe, the
entire water column is instantly at rest, and all the kine-
tic energy has been converted into pressure energy. The
hydraulic gradient line is above the reservoir level for
the entire length of the pipe (diagram d). Because the
reservoir pressure is unchanged, an unbalanced force acts
on the layer in the upstream end of the pipe at the instant
of the arrival of the pressure wave. This condition is
unstable.
Phase 2. Diagrams (e) to (g). The pressure head in
the layer adjacent to the reservoir drops to the reservoir
level and the portion of pipe encasing this layer is no
longer stretched,_ but returns to its normal diameter. By
this reason it discharges into the reservoir the amount of
water that was stored in the space caused by the increase
in pipe diameter. In this manner the pressure energy of
the water and the pipe is reconverted into kinetic energy
of the water. The process of reconversion starts at the
reservoir and travels toward the downstream gate end as a
wave, and behind the wave front the pressure head stays at
the reservoir elevation and the water flows out of the
pipe into the reservoir. The wave of compression of phase
1 is thus reflected from the reservoir as a wave of expan-
sion of water and contraction of pipe diameter. At the
end of the phase, when the wave front reaches the closed
 10

gate, momentarily the entire water column has a velocity -V 0

and the pressure in the entire pipe is restored to normal
(diagram g)
Phase 3. Diagrame (h) to (j). The velocity in the por-
tion of pipe adjacent to the closed gate is reduced abruptly
to zero and the kinetic energy of the water is converted
into pressure energy. this results in the elastic work of
expending the water and contracting the pipe wall. A wave
of decreased pressure travels from the closed gate toward
the reservoir. The wave of expansion of phase 2 is reflect-
ed from the closed valve as a wave of expansion. In the por-

tion of the pipe between the wave front and the closed valve
the velocity is zero and the pressure is below norma1 by an
amount corresponding to the excess pressure of phase 1. At
the end of the phase, when the wave front reaches the reser-
voir, the entire water column is momentarily at rest and
,the hydraulic gradient line is below the reservoir level
over the entire length of the pipe (diagram j).
Phase~. Diagrams (k) to (m). The condition in the
reservoir end of the pipe at the end of the third phase is
~nstable. The pressure head 1n the layer adjacent to the res-
ervoir ris.e s to the reservoir elevation and the pipe expands
to its normal diameter. By doing so it causes water to flow
from the reservoir into the pipe to fill the volume resul~ing

from increase of the diameter. A wave of increased pressure

travels from the reservoir toward the closed gate, and the
initial velocit1 and pressure are restored behind the wave
 11

front. At the end of the phase, when the wave front reaches
the closed gate, the entire water column has a velocity V0
and the pressure head in the entire pipe is at reservoir
level (diagram m), as at the beginning or phase 1.
Fig . 2 shows the variation or pressure with respect
to time at the valve for an instantaneous closure of the
valve . (5) Each phase described above extends over a period
or L/a seconds where L is the length of the pipe and a is
the velocity of the pressure wave.

P•

AP
2L 4L
I
"'1["""".
6L t
T a
Po

I API I I
Fig. 2. Variation of Water Hammer
Pressure with respect to Time
at the Valve End.

John Parmakian (1} has developed the basic water hammer

equations with friction losses neglected,

aH
ax
=-...L(..£!_
at * v
g
av
~X
) (1)

2 av
aH + v a H =- .....!,_ (2 )
~t i}X g ~
It is generally known that the friction factor is function
of Reynolds number Rn, that is
hr or
f= - - - - , . .2 -
L
(""1)
v
2g)
 12

2
h-rL v
rrr~ (3)
Victor L. Streeter solved those equations by the method
or characteristics (3), obtaining the characteristic equa-
tions as follows:

dt-~ =0 (4a)
V+-a

f 2
dVt-
t:1
-:-dH+·215 v dt =0
vdx
dt-

dV-
-· = 0

f"- dBt frt-v 2dt = 0

(5a)

(Sb)

These transform into difference equations (refer to Fig. 3):

(tp-tR)-( vfa )R(xp-xR) =0 (6)

( Vp-VR)+ t<Hp-HR}f- ( ~v2)R (tP-tR) =0 (7)

(8)
( t P -ts)-(..-..1-->s (~p-xs> :: o
V-a
(9)

t.
I+ I

8
t.t
Ax AX

0~----------------------~x

Fig. 3. The Specified Time Inte1~al

 13

The differential distances of IJ x and ~t in Fig. 3 are

lbnited by the characteristic lines, equations (4a) and ($a).
Assuming~x is known, then the position of P must lie within
the area PRS described by the characteristic lines. If the
At increment places point P outside this area, the values of
H and V will give a divergent solution.
Streeter used the MAD language (Michigan Algorithm De-
coder) to make the computer programs tor different ways or
gate closure and presented several practical examples and
obtained good approx~ations to the solutions.
 14

III • WATER HAMNER EQUATIONS NEGLECTING FRICTION LOSSES

(A) Assumptions and Notation

The analysis of water hammer in a closed conduit is

based on the elastic water column theory, therefore, it is

assumed that (l)
(l) The water in the pipe is compressible.
(2) The pipe walls deform elastically under the action
of a pressure change.
(3) The pipe r~ains full of water at all times.
(4) The hydraulic friction losses and velocity head are
negligible.
(5) The velocity is uniform over the cross section of
the pipe.
{6) The reservoir level remains constant.
The following notation is used (see Fig. 4 and Fig. 5)
A = cross-sectional area of pipe, in feet;
a = velocity of pressure wave, in feet per second;
~
b = thickness of pipe wall, in feet;
c = factor of pipe restriction;
1
cr= d~ensionless constant;
D = inside diameter of pipe, in feet;
E = modulus of elasticity of pipe wall, in pounds per
square foot;
e = roughness height, in feet;
r .X= external forces per unit mass exerted on the mass
ot the particle in the x-direction, in pounds per
slug;
t = friction factor;
g = acceleration or gravity, in feet per second2;
H = total pressure head at an arbitrary point along the
pipe line, in feet;
H
0
= head at the reservoir, in teet;
AH= head rise or drop at an arbitrary point, in feet;

hr= friction loss, 1n teet;

K = bulk modulus or elasticity of the water, in pounds
per square toot;
k = dimensionless number;
L : total length of pipe, in teet;
p = wetted perimeter or pipe, 1n teet;
p = pressure at any point in the pipe, in pounds per
square foot;
Rh= hydraulic radius, in feet;
R : Reynolds number;
n
r = radius or pipe, 1n teet;
Te• ttme of gate closure, in seconds;
t • ttme, 1n seconds;
v = velocity in pipe at a given t~e, in feet per second;

v0 = velocity in pipe for initial steady state conditions,

in teet per second;

w • specific weight of the water, 1n pounds per cubic foot;
Z = elevation, in teet;

Vx, v1 , Vz • the component velocities in the direction

ot the three coordinate axis.
 16

OC = angle of slope of pipe;

f> = density of the water, in slugs per cubic foot;
~o= shear force per unit area of pipe wall;
~ = viscosity, in pound-second per square foot;

1f= volume of water, in cubic feet;

~ = kinematic viscosity, in square feet per second;
cr = unit stress, in pounds per square foot;

J = elongation, in feet;

w. s.

Reservoi

wA(H 0 +A H)

Fig. 4. Definition Sketch

(B) Basic Equations
(l) The Dynamic Equation: Referring to Pig. 5, the
pressure difference between points band b', dx apart, is

;rr2p-1Cr2(P+ Udx)
ax
= -lt'r 2 fi-dx
According to Euler's equation<6)
 17

ft - ; P =fl ( Ovx. +v a vx +v d vx +v t:JVx

X ax at X aX y 3 1 z ~ Z )

or w 4' 2P w dV
-g.a.x- ~x =g-at (10)

dV _ V(x+dx. , t+dt)-V(x.,t) (_ay_dx+~dt)-L

dt - dt = (J X 3t dt

-(llvdt+LLdt)_!_
- ax at dt

=V~+ a v
ax
qt (11)

and the only external .force is the weight W?& r2dx of the
element of length dx; there is no force component 1n the
x-d1rect1on, that is fx=O, from Eq. (10):

..!!.. < v~ + av > --

- R.E_
g ax at ax
Substitute p=wH and dp=wdH 1n the above equation, then

dH
ax=-- _!_(.ll+v
g at av
ax ) <12 >

It may be assumed that the ~~ te~ is very small compared

with ~~ , and can be neglected.

... a~ vt oH
=-g ~ x
<
13 >

w s.

Reservoir

Fig. 5 Bquilibriu. ot Momentum tor a Water Co1umn

 18

(2) Equation of Continuity (see Fig. 6)

(a) Increase in velocity: during the time inter-
val dt, the volume increased due to the increase of velocity
is (7)
(14)

Such a difference between inflow and out-flow is possible

only when either water is compressible under increased
pressure or the pipe line is expanded.
{b) Shortening of water column due to increase
in pressure: from Hooke's Law
and Ad= a-E J

'
•
I
d(dx) :: -k a
~t
P dtdx

~ dt means increase in pressure in the time interval dt.

T.he resultant change in volume is

dV2 ::1tr 2 d(dx) =1tr

2
~dtdx -f (1$)

(c) Expansion or pipe line due to pressure rise:

owing to the increase in pressure (~~)dt, the pipe radius
is increased by
dr = d(T'-j-
••• (F = p!.... and d (f'= c!.E....d ~
b at b

2
•• dr= *d~
the corresponding change in volume is

= 2Ttrdrdx = ~t 2 3 (16)
d?f3 dtdx E/,r
 19

dx

v- c'
c- -- -- ___.......,. (V- _adx)
~X

~ig. 6. Condition of Continuity

(d) Condition of continuity demands that

d t51=dlf +d tf
2 3
-7! r
2
av dtdx =nr 2 ~dtdx .1.. + ~dtdx 2 trr3
ax- --at" K at Eb

• _ av = __a_p (~ + ~) =w ~Ht ( ~ + ~ )
· · ax --aT K Eb o n. ~

let ~(...!.. + ~) = a~
g K Eb

or (17)

where a is the wave velocity in feet per second, then

av -~ aH
ax - a2. at (18)
Equations (13) and (18) are the approximate water hammer
equations, and their solutions are (see Appendix B):

H - H0 = F{t- 4> + f(t + !) (19)

a [F{t - ~)
a - f(t + ...!..)]
(20)
V - V
o = -.A. a
(C) Physical Meaning of Water Hammer Equations
When t = 0,
x = x 1 ; t = t, x = x 1 + a~
then F(t - ..!.) = F(t - XJ + at) = F( ~) = F(c) = constant
a a · a
 20

this means that the constant function F(xl/a) will move in

the direction or positive x with a velocity a at the instant
t.
Therefore the function F(t-2L) represents a pressure
a
wave of any shape, which ie propagated as a whole and with-
out deformation, at a velocity a, in the direction of posi-
tive x.
Similarly, the function f(t + ! ) represents a second
pressure wave of any shape, which is propagated without
deformation, at a velocity a, in the direction or negative
x.
Referring to Fig. 1 and from Eq. (19), H-H 0 , which re-
presents the excess pressure at point x and at t~e t, will
be the sum of two pre~sure waves F(t-~) a and t(t +~).
a
Sim.ilarly, from Eq. (20), V-V 0 , the change in velocity, is
e~al to the difference of the two pressure waves (g/a)f(t

+ ~ ) and (g/SJ)F(t-! ). Both F(t-! ) and t(t +! ) de-

pend on the boundary conditions at the ends of the pipe
line.

/
.
.,..__f( a:::t+x/a) F(t-x/a)
/ / '
/
I
/
/
-- ---..... ......
~-
~

Fig. 1. Water Hammer Wave Propagation

 21

From Eqs. (19) and (20), it is seen that F(t- !> is a

pressure height which is measured in the same unit as H 1n
teet ot water. (see Fig. 8). After a small interval o~

t~e dt bas passed

(t + dt) - (x ! dx) = (t + dt) - ( x ! edt) : t
This means that curve mn can also be used for the instant
ot time t + dt, provided dx •adt. Hence, F(t - ~) function
represents a pressure wave which moves in the direction o~

positive x with a velocity a, and t( t + ~ represents a

pressure wave which moves 1n the direction ot negative x
with a velocity e.

n' n

/
/
F( t- f-> at time t+dt
/
/
/
/
/
/
/
/
/
w.s. ...........
/

Fig. 8. Physical Meaning of Water Hammer Equations

 22

IV. NUMERICAL METHOD FOR WATER HAMMER EQUATIONS

INCLUDING FRICTION LOSSES

(A) Numerical Analysis For Water Hammer Equations In-

cluding Friction Losses
From Eqs. (3) and (12)

~! + + ~: = - ~ ( :! + v ~! ) (21)

The equation of continuity developed by John Parmakian(l)

will be used here:

(22)

Eqs. (21) and (22) can be solved simultaneously by the method

or finite difference approximations with the aid of a digital
computer, using specified time intervals as described in
the following procedure.
A part of the x-t plane is shown in Fig. 9. The ~V/ax
andoH/ax ~t point M can be approximated in terms or finite
differences as(B)

VR-VL
( all (23)
ax )M = 2 AX

aH I.lR-HL (24)
(
~X
)M = 2AX

at point P, a V/ at, and a H/ t!J t may be written as

( av
at )p = VP-VM
At
(25)

Hp -HM
( aH )p = (26)
at At
 23

t
• • • •

• • • • •

p
• • • • •

• L. R• •

Fig. 9. Point Netwo~k Showing Finite Differences

Substituting Eqs. (23), (24), (25) and (26) in Eqs. (21)
and (22), the following equations are obtained:

~-~ +

Solving for Hp and Vp, then

Hp= HM- ~ *[\fli(HR'B:r,) + f<vR-VL)J (27)

VJF VM- ~ ti [ VJo1(VR..\1t) + g(HR-HL) + ~V~x] (28)

It all values are ~no~ at time ~ t 0 , then the unknowns in

Eqs. (27) and (28) at tilne ~ t 0 +At are Hp and Vp, which
can be solved separatel1 tro~ these two equations and it is
possible to step forward an increment in time,~t, in both
pressure and valoci ty. For the next time increment, At,
Hp and Vp at time = t 0 tA t become HM and VM for time=t 0 + 2t\t.
The values of Hp and VP at time =t +2~ t
0 then become the HM
and VM at time = t 0+ 3A-t, etc.
Initial and Boundary Conditions:(9)
1. Initial Conditions: A necessary initial condition
for the solution is that all the pressure heads and ve-
locities along the pipe line, 0~ x f L, must be known at
time to=O. This is true before gate operation, when steady
state conditions exist, that is
V(x,O) = V0 and H(x,O) = H0
2. Boundary Conditions: The necessary boundary
conditions for the solution are those at x=O and ~L.

(a) The Right End (gate end): Fig. 10 shows a portion

of a pipe with a partially closed gate. Assume that the
partial closure of the gate occurs instantaneously and the
velocity of water upstream from the gate is reduced from V0
to v1 • A portion of the initial kinetic energy of water is
converted into pressure energy. A wave of increased pressure
is formed at the upstream face of the gate. The wave front
travels upstream with high velocity and results in the
elastic work of compressing the water and stretching the
pipe wall. A wave of reduced pressure is formed at the down-
stream face of the gate. The front of this wave travels
downstream. Assume that the pipe on the downstream side of
 25

b a

t.fodt t
Vo t-.- Vl
• PoA t:=plA
a ____...+V
dx +x ..-..
x+-dx

Fig. lO(a)

Wave front~ a
I

Fig. lO(b) Fig. lO(c)

Fig. 10. The Right End Boundary Condition
the gate is short and disregard the water hammer wave in
this portion of the pipe.
Let x be the distance between gate and the wave front
in Fig. lO(a) , at the time t after the change in gate opening.
The abscissa x, measured from the point A at the gate is
positive toward the inlet at reservoir end, while the ve-
locity of water is positive in the opposite direction. The
front ot the pressure wave advances a distance dx during
the time interval dt. The velocity of the wave front is
dx
(1"£'""".

In Fig. lO(b) the wave front is at section a-a at the

 26

beginning of the time interval dt. The mass ~Adx of the

water column between sections a-a and b-b has a momentum in
the direction of positive x of - ~dxV 0 • At ·the end of time
interval dt the wave front is at section b-b as shown in
Fig. lO(c). The mass of water described above has a mo-
w
mentum - gAdxVl•
The impulse or the acting force system in the direction

of positive (i.e. the horizontal resultant force (p 1 -p 0 )A

X

acting on the mass element, multiplied by the time interval

dt) is equal to (pl-p 0 )Adt.
Equating the impulse during time interval dt to the
change in momentum during the same t1me interval dt, we
obtain
Ft = MV
(p1 - p 0 )A dt = - iA dx (v1 - V0 ) or
Pl- Po= ~ l!<vo- Vl) (29)

This equation shows that in the region passed by the

wave front the pressure exceeds that of the undisturbed
flow by
v
A p = ga .A V, where
a is the velocity of the pressure wave in a pipe line.
Substitution of P= wH gives
(30)

(31)
which shows that the head rise is proportional to the ve-
locity reduction.
Referring to Fig. 11, let x = xM• L be the boundary
 ?7

t
Boundary line

• • • • •

• • • • •

t 0 + £1 t • • • • •

t = to L M X
1. AX .Jx•L
' I

Fig. 11. The Right End Boundary Point Network

line and use a system ot grid points at the boundary.
When t <T 0 , from Eq. ( 30) and

VM -6V =r
Vo P
J HM +LlH
Ho (32)

1n which rp is the ratio ot the effective gate opening at

tfme P during the gate closure to the effective gate open-
ing at tbne zero<S>, and

(33)

By solving Eqs. (30) and (32) simultaneously tor ~H

and L) V, then
Vp = VM -AV (34)
Hp: HM .f. llH (3$)
For t:>T 0 (= time of closure), Vp • o.
(b) The Lett End (reservoir end): The pressure
bead at reservoir end remains constant during the gate move-
ment time, that is H(O,t)= H0 • Let X•XM=o be the boundary
 28

Boundary line
t

• • • • •

• • • •

t 0+.At • • •

Fig. 12. The Left End Boundary Point Network

line as shown in Fig. 12, and use a system of grid points
at the boundary. From Eqs. (19) and (20), since H-H~O,

then F( t-L/a) = -f('btL/a) and

V - V0 = -2 f- F(t-L/a) (36)
this equation means that when a direct pressure wave reaches
the reservoir, the change in velocity of water at that lo-
cation is twice that produced by the direct wave at other
points in pipe line. There.f ore
Vp= VM-2d V ( 37 )

Program Flow Chart

Fig. 13 shows the computation procedure, first, H0 and
V0 will be given at A0 , A1, A2 , •••••• A9 as shown in Fig.
13, then using Eqs. (27) and (28), Hand Vat B1, B2 , ••••
Ba can be evaluated; next, using Eqs. (34), (35) and (37),
H and V at B9 and B0 can be· evaluated; in a similar manner,
 t

~M •
• • • • • • •

• • • • • • • •
.A.t I AX

• • • • • • • • •

co c • c2 .c3 .c4 .c5 06 .c7 c8 .c9

• 1 • •

K=2 Bo ,Bl • B2 .B3 .B4 .B5 • B6 .B7 .Be .B9

'b=:t , K.::l 1Ao ,nl --2 .--3 --4 -S .--6 · 7 ··H • 9 •

0 I=l
~o ~L

Fig. 13. Computation Procedure

(\)
....!)
 30

they can be obtained at c1 , c2 , ••••• ca, and then c9 and

C0 ; etc. So that the computation can be performed as far
as desired. Fig. 14 shows a flow chart for the general
program procedure for solution of water hammer including
hydraulic friction losses in a simple pipe line of constant
cross section. As pictured in Fig. 14, the computer proce-
dure is as follows:
(1) All data and parameters for the program are set
to initial values.
(2) Starting at the first time increment, ( ~2),
t = t 0 +At, all friction factors, pressure heads and veloci-
ties at each point along the pipe line are evaluated beginning
with the point I = 2, through the next to the last point,
I = N-1.
(3) After this is completed, the left end, (I=l), and
right end, (l=N), boundary values are computed tor the given
time, t.
(4) The values obtained from steps (2) and (3) are
then printed as output data for time, t.
(5) The program sets t = ~At and sets the initial
values of pressure heads and velocities for the next time
increment. Steps (2) through (5) are repeated until the
final time increment, t = M, is completed, where M~Tc•

(B) Practical Examples

Example 1. A 5-tt.-diameter steel pipeline, 1.0 in.
thick and 3730 ft. long, carries water at 5.0 ft/sec and
at 70°F. A valve at the downstream end of the pipe has a
 31

READ IN INITIAL VALUES

OF DATA AND PARAMETERS
SET I =
2, t =
l.

COMPUTE THE PRESENT VALUES OF

FRICTION FACTOR (EQ.3), PRESSURE
,HEAD AND VELOCITY ( EQ$. 27 and 28 )
FOR THE PRESENT POINT, I 1 AND
TIME, t.

DOES I = N - 1? '--------41

SET INITIAL
VALUES FOR COMPUTE THE LEFT END BOUNDARY
HAND V POINTS (EQ. 37) AND THE RIGHT END
BOUNDARY POINTS (EQS. 34 and 35)
FOR THE PRESENT TIME, t.

PRINT OUT THE VALUES OF t 1 H,

AND V AT PRESENT TIME, t.

DOES t = M'l

Fig. 14. Computer Flow Chart

 32

head of 236 ft. across it initially. For valve area

closure in time tc=B.o sec., find the pressure intensity
at the valve and velocity variation at both the valve end
and reservoir end. (slow valve closure)
(Solution) L = 3730 ft.
D= ,5.0 ft.
Ho= 236 f't.
V::
0
,5.0. :Ct/sec.
T= 70°F

Tc= B.o sec.

a= 3730 ft/sec.
L/a = 1.0 sec.
From the computer program output, the
answers are shown in the following table.
(see APPENDIX A, Program 1.)
VR ft/sec. llv f:U. vv ft/sec.
tin L/a sec.
0 ,5.00 236.00 ,5.00
l 4.23 267.01 4· 7.5
2 3 • .53 292.81 4.27
3 2.80 332.70 3.71

4 2.06 3.58.88 3.14

5 1.33 414.26 2.43
6 0.59 440.34 1.73
7 -0.14 .511.73 0.89
8 -0.87 543.46 o.oo
9 -1.60 62.5.33 o.oo
H
Fig. 15. Variation of Water Hammer Pressure with Time
ft. at Valve End. T0 = 8.0 sec.

/000

400

200

OL_ I I I I I I I I I I ~
0 I 2 .j 4 5 6 7 8 9 10
t in L/a sec.

\.-"
\.-"
 Fig. 16. Variation of Water Hammer Velocity with Time
at Reservoir End and Valve End. Tc = 8.0 sec.

v .5
ft;/sec.
4

.3

2.

I I > >J I I ..

0 1- I I I I <!:o: - - in L/a sec.

_,
-2

-3

w
+=-
 35

Example 2. A 8-tt.-diameter steel pipeline, 1.0 inch

thick and 820 ft. long, carries water at 10.0 ft/sec and at
70°F. A valve at the downstream end of the pipe has a head
of 195 ft. across it initially. For valve area closure in
t~e tc=0.7 sec., find the pressure intensity at ~be valve
end and velocity variation at both the valve end and res-
ervoir end (rapid valve closure.)
(Solution) L = 820 ft
D = 8.0 ft
H0 = 195 ft
V0 = 10.0 ft/sec

T = 70°F
Te= 0.1 sec.
• = 4680 ft/sec.
2L/a· = 0.35 sec.
t VRft/sec. H ft. V tt/sec.
in L/2a: sec. v v

0 10.00 19.5.00 10.00

1 9.71 320.75 9.19
2 9.42 281.11 8.65
3 9.13 470.96 7.62
4 8.84 303.67 7.46
5 8.56 867.86 4.73
6 8.27 373·. 32 4.25
7 7.98 1235.01 1.73
8 7.69 628.80 o.oo
9 7.40 156.3.61 o.oo
 1000

H, ft.
800

600

400

eoo

0 I z 3 4 s 6 7 8 9 t in L/2a sec.

Fig. 17. Variation of Water Hammer Pressure with Ttme

\.t.J
at Valve End. Tc = 0.7 sec. 0'
 /0

.9

v
ft/sec.
7

"

0~--L---L---~--~--~--~--~~~--~--~------------~
0 I 2 3 4 s 6 7 8 8 10 t in L/2a sec.

Fig. 18. Variation of Water Hammer Velocity with T~e

at Reservoir End and Valve End. T0 = 0.7 sec.

\.,..)
--3
 38

Example 3. A 0.0365 ft.-diameter steel pipeline, 0.5

inch thick and 200 ft. long, carries water at 2.11 ft/8ec
and at 70°F. A valve at the downstream end of the pipe
has a head of 335 ft. across it initially. For valve area
closure 1n time t 0 =0.09 see., find the pressure intensity
at the valve end and velocity variation at both the valve
end and reservoir end. {rapid valve closure.)
(Solution) See APPENDIX A, Program 2.
L = 200 ft.
D = 0.036.5 ft.
H
0
= 335 ft.
V0 = 2.11 ft/sec.
T = 70°F
T c= 0.09 sec.
a = 4444.44 ft/sec.
2L/a = 0.09 sec.
t v ft/ 80• H ft
v
V ft/sec.
v
1n L/2a sec. R

0 2.77 335.00 2.77

1 2.66 425.79 2.10
2 2.55 405.59 1.58
3 2.44 537.34 0.77

4 2.32 512.00 o.oo

5 2.15 687.26 o.oo
6 0.93 826.00 o.oo
7 13.32 1331.00 o.oo
8 27.2.3 1828.20 o.oo
9 22.67 2$42.25 o.oo
 39

In example 3, the increments are too large, which makes

the finite difference equations divergent and the answers
become unreasonable after the seventh interval of time.
Both the velocity and pressure head increased irregularly.
This example shows that the increments should be made as
small as possible to make the answers closer to theo-
retical answers and improve the approximations.
 40

V. CONCLUSION

The direct solution ot the basic partial differential

equations is the better method capable of distributing the
effect of friction losses along the pipe line 1n agreement
with the actual situation. Because the v2-ter.swaa includ-
ed 1n the equations, it seems to have greater accuracy
over other methods. This numerical method uses the con-
tinuity and dynamic equations, and transfo~ them into
finite difference equations, which with initial conditions,
boundary conditions and other parameters make the basic
pa·r tial differential equations solvable with the aid of a
digital computer. From the computer output data, several
curves were plotted to illustrate the phenomena of water
hammer. Figures 15 and 17 show the variation of water
hammer pressure with t~e at the valve end. With different
types of gate operations, the pressure rose in different
ways.
For any fixed point along the pipe ltne, the velocity
and pressure can be traced following the ttme variation.
This method is capable ot adequately handling many differ-
ent gate operations. After a computer program tor one type
ot pipe system is made, the same type of proble~ can be
solved as many times as desired by s~ply changing the in-
put: data.
·! Dee the IBM 1620 digital computer doea not haYe big
enough capacity, the answers can not be tu.rther improved by
 41

increasing the number of points in the network. If one

wants to make further investigations for this method, he
may make the increment less in both the tLme interval and
distance along the pipe line, then the result will be a
better approximation of the .exact solution.
 BIBLIOGRAPHY

1. Par.makian, John, Waterhammer Analysis, Prentice-Hall,

Inc., New York, (1955) 161 pages.
2. Rich, George R., Water-Hammer Analysis by the Laplaca-
Mellin Transformation, Trans. ASME 1945, paper 44-A38,
p. 316
3. Streeter, Victor L. and Lai, Chintu., Water Hammer
Analysis Including Fluid Friction, ASCE Volume 88,
Hy3, ( 1962).
4. Rich, George R., Handbook of Applied Hydraulics, Mc-
Graw-Hill Book Company, New York, p.638-652. (1952)
5. Streeter, Vietor L •• Fluid Mechanics, McGraw-Hill
Book Company, New York, ( 1958·) Second Edi t1 on,
p. 363-374.
6. Rouse, Hunter, Elementary Mechanics of Fluids, John
Wiley & Sons, Inc., New York, (1946) Second Edition,
p. 201-213.
7. Den Hartog, J. P., Advanced Strength of Materials,
McGraw-Hill Book Company, New York, (1952) p. 90-99.
B. McCracken, Daniel D. and Dorn, William s., Numerical
Methods and Fortran Programming with Applications in
Engineering and Science, John Wiley and Sons, Inc.,
New York, (1965) P• 365-404.
9. Rich, George R., Hydraulic Transients, McGraw-Hill
Book Company, Inc., New York, (1951) First Edition,

p. l-24·
 43

10. Churchill, Ruel v., Fourier Series And Boundary Value

Problems, McGraw-Hill Book Company, New York, Second
Edition, p.77-l25.
11. Ellis, D. s., Elements of Hydraulic Engineering, D.
Van Nostrand Company, New York, (1947) p. 230-243.
12. Hildebrand, F. B., Introduction To Numerical Analysis,
McGraw-Hill Book Company, New York, (1956) p. 128-160.
13. Hunsaker, J. c •• and Rightmire, B. G. Engineering
Application of Fluid Mechanics, McGraw-Hill Book
Company, New York, (1947) p. 122-126.
14. Vennard, John K., Elementary Fluid Mechanics, John
Wiley & Sons, (1954) Third Edition, p. 188-204.
 44

APPENDIX A
THE COMPUTER PROGRAM

The program for computing pressure and velocity varia-

tion in water hammer analysis is listed here for the bene-
fit of others who wish to do further investigation on the
water hammer problems. The program is the "Fortran II"
:program and use the IBM 1620 digital computer. With the
Change of input data, most of the general problems in water
1

lnammer can be solved. The program is as following:

*LIST PRINTER
*ALL STATEMENT MAP
C C***CEX029 WENG L H FORTRAN 2
C PROGRAM l PRESSURE AND VELOCITY VARIATION IN WATER
C HAMMER ANALYSIS
DIMENSION H(25.25),V(25,25),DEH(25,25),DELH(25,25)
READ lOO,PL,A.D,G,FH
READ lOO,H(l,l),V(l,l),SH,TC
PRINT lOO,PL,A,D,G,F.H
PRINT lOO,H(l,l),V(l,l),SH,TC
DO 50 I=l, 19
H(I+l, l)::H(I ,l)+SH
V(I+ 1,1 )=V(I, 1)

50 PRINT 200,H(I+l,l),V(I+l,l)
DO 60 1";:1, 19

B(l,:&l)=H(l,K)
 45

60 PRINT 300, H(l,~l)

DT:: PL/A

DX s PL/20 .o

R = DT/DX
I= 2
K =1
10 Y= (2.0~FH*D)/(PL*V(I,K)**2)

DV = V(I+l,K)-Y.(I-l,K)
DH = H(I+l,K)-H(I-l,K)
H(I,~l) = H(I,K)-0.5*R*(V{I,K)*DB+(A**2*DV)/G)
V(I ,Kf.l) = V(I ,K)-0.5-J$-R*(V(I ,K)*DVf.G*OO. (Y*V(I ,K)**2*DX)/D)
PRINT 400,H(I,K+l) ,V(I,K+l)
I=I+1
IF(I-19)10,10,20
20 V(l,K+L)=V(l,K)-2.0*(V(2,K)-V(2,~1))
PRINT 500 I V(l,IQ-1)

DVR=V(l9 ,K)-V(l9 ,Ia-1)

0= (H(20,K)+ (A*DVR)/G)/H(20,1)

W=S~TF(C)

B=K

P'D=B*DT
DEN=V(20,K)-((TC-PT)*V(20,l)*WYTC
DEH(20,K)=(A*DEV)/G
DELH(20,l)=O.O
DELH ( 20,91 )=DEH ( 20 ,K) -DELH ( 20 ,K~J
H ( 20, IQ-1 )::H {20 ,K )+ DELH ( 20 I & l)
IF(TC-PT)l4,l4,12
 46
14 V( 20 ,IQ-1 )=O.O
GO TO 13

12 V(20,~l)~V(20,K)-DEV

13 PRINT 600,DEH( 20,K) ,DELH(20 ,K+ 1) ,H(20 ,~ 1), V(20 ,IQ-1)

~&1

IF(K-20)30,90,90
30 I::::2
GO TO 10

100 FORMAT (2X,5Fl4.4)

200 FORMAT (6X,4Hall=,Fl2.4,6X,4HVIl=,F12.4)
300 FORMAT (6X,4HalL=,Fl2.4)
400 FORMAT (4X,3HH~ 1 El8.8,4X,3HVP.:,El8.8)

500 FORMAT (6X,3H~ 1 El8.8)

600 FORMAT (2X4HD~ 1 El2.5,2X4HD~,El2.5,2X3~,El2.5,

2X3HVR= ,El2.5)
90 CALL EXIT

END
C PROGRAM 2. PRESSURE AND VELOCITY VARIATION IN WATER

C HAMMER ANALYSIS
DIMENSION H(l2,17),V(l2,17),DEH(l2,17),DELH(l2,17)

READ lOO,PL,~,D,G,FH

READ lOO,S(l,l),V(l,l),SH,TC
PRINT lOO,PL,A,D,G,FH
PRINT lOO,H(l,l),V(l,l),SH,TC

DO 50 I::::l,9
ll (I+ 1, 1 )=S (I , l )+ SH
 47

50 V (I+ l, 1) = V (I , l)

DO 60 If:=l, 14
60 H ( l,K+ 1 )=H( 1 ,K)
D'D=PL/ ( 2. O*A)
DL=PL/10.0
~DT/DX

1=2

K=l

10 Y= (64.4*FH*D)/(PL*V(I ,K)-)HE-2)
DV=V(I+ l,K) -V(I-l,K)
DH=H(I+l,K)-H(I-l,K)

H(I ,IQ-1 )=H(I ,K)-0 • .5*R*(V(I ,K)*DHt (AH-2*DV)/G)

V(I,~1)=V(I,K)-0.5*R*(V(I,K)*DVtG*DE+(Y*V(I,K)**2*DX)/D)

PRINT 400 ,H(I ,IQ-1), V(I ,K+ l)

I=I+1
IF(I-9 )10 ,10 ,20
20 V( 1 1 !9-·1 )=V( l,K) -2.0*(V( 2,K) -V( 2,19-l))

PRINT ,500,V(l,K+l)
DVR=V(9,K)-V(9,K+l)
C=(H(lO,K)+ (Ait-DVR)/G)/H(lO,l)

W=S~TF(C)

B=K

PT=B*DT
DEV=V( 10 ,K)- ( ( TC-PT)·U·V(lO, 1 )*W)/TC
DEH(lO,K)=(A*DEV)/G
I

DELH( 10, 1)=0.0

DELH( 10 ,K+1)=DEH (10,K)+ DELH(lO,K)
 48

H(lO,K-l)=H(lO,K)-DELH(lO,K-1)
IF(TC-PT)l4,14,12
14 V(lO,K-1)=0.0
GO TO 13
12 V(10,K-1)=V(lO,K)-DEV
13 PRINT 600,DEH(lO,K),DELH(lO,K-l),H(lO,K-l),V(lO,K-l)
K=K-1
IF(K-15)30,90,90
30 1=2
GO TO 10
100 FORMAT (2X,5Fl4.4)
400 FORMAT (ij.X,3HHP= ,El6.8 ,4JC,3HVP= ,El8.8)
500 FORMAT (6X,3HVL=,El8.8)
600 FORMAT (2X4HDEH=,El2.5,2X4HDLH=,El2.5,2X3~,El2.5,

2X3HVR=,El2.-5)

90 CALL EXIT
END
 49

APPENDIX B

WATER HAMI"'ER ANALYSIS BY THE LAPLACE-MELLIN

TRANSFORMATION, Trans. ASME 1945, paper 44·A38, p.316.

Let H= l ~ -H"tadt and

v :e [,DUe-Htvdt

a2 ~) = ~:~ (l)

Multiply through Eq. (l) by e-Ht and integrate from

0 to oo:

a2
a 2!Jart:l<l e -HtHdt =laO
o o e
-Ht a < a H
at at )dt
F
:Jo e
-Ht a H
C3(at)
X o

a2
0:~ H .. (e-Ht,~~~)J:: H[e-HtHJ +i['"e-HtHdt

LarJ.(t;:·o .;.u(a] t=o+ H~

= - (aH
= O-H•H 0 + H2 H
~ 2
dli H -H + H H - 0
-- 2 o-
a x2 a2 a
~!g~~! to ordinary differential equation:

d~- - H
H +-HH 0 =02
-
dx2 a2 a2
X X H.
Solving: H= A cosh La + B sinh La + ...2... (2)
H

(3)
.;, .elL= .L 2 l
ax g at
Multiply through Eq. (3) by e-Ht and integrate :from 0 to cu:
 50

a LO<J 10()e -Ht a V

0 X ~:> e
-Ht
Hdt = 1g )rove -Ht (8t
0
av
)dt =
l
g 0

oR
ax ::: g
l { (
e-
Ht
V o
)O(J + H Jo
(o<.J Ht
e- Bdt
}

V H V
-
_.!..Q+-
g g

v +
=- " ..!.Q... !LY... (4)
g g

Boundary condi tiona.:

(a) When x =0 1 H = H0
-
H=;(~ e -Ht__
0
Ho
J1 0 dt = H: 1 So A=O

(b) When x = L V = 0; so 1 V =0
From Eqs. (2) and (4):
B = .....liiig.!..._Vo
__
H L
H cos~
a
There i = ......i_vo;;__ _ sinh~ + Ho
a H
H cosh HL
a
cosh .1!!...
- Vo a
v = n-< 1 - ----- )
cosh .1!...1!
a

Inverse transform:

H = H 0+ aVn
.-..........
1
• L
...::L. ~ (-.ltn-l
_ (2n-1) 1t'a t·
sin ( 2~lmx
'-
g i t n= 2n- 2L s 1n
1

oe::;- ~ 1 )n-1
O<J _
V = V0 n:!b ~12n-lJ cos -
(2n-l)1l at
-a cos
(2
n2L1)1rx

From the trigonometrical formulas:

 51

sin mx sin nx = ~ [cos (m-n)x - cos (mt-n)x J

oos mx cos nx =~ ( cos (m-n)x + cos (mt-n)x J
Then

sin (2p-l)1C at sin (2~l)trx

2L

= ~ cos ( 2n-l) 1t" ( at-x) _ cos ( 2n-l) 11: ( abt x)

t:. 2L 2L
cos (2n.l)itat cos (2n 1)7Cx
2L 2L

= !(cos ( 2n-l) 7t (at-x)

2L
+ cos . {2n-l )7C (a'bt- x)J
2t
and
.!..!,g, .JJ:~ (-l)n-1 { (2n-1)1t(at-x)
H = Ho + g 1t L.. 2 ( 2n-l) cos ..---.....::;;;.._;2L~--.-...--..
n-1

_ cos (2n-l~(abtx)}

V = Vo - &_~ (-l)n-l {cos (2p·l)1£{at-x)

1'"C n==l 2(2n-l) 2L

+ cos (2n-l) Tr(at + x)}

2L
Let (2n-l)~(t-x/a)
cos 2aL

DO
-F( t+ ~) = aVo .Jt Z (-l)n-1
cos
(2n-1)1t'(t + x/a)
2at
a g It' n=l 2( 2n-l)
 52

then H-H 0 = f( t- ...!...) + F( t + ...i.)

a e.

V-V 0 = L[t(t-
e.
...!.) - F(t +
e. a
..!..)]
 53

VITA

The author was born on January 15, 1938, in Amoy,

Fukien, China. He received his primary and secondary
education in Chungli, Taiwan, China; his college education
at Taipei Institute of Technology in Taipei, Taiwan, China,
and Cheng Kung University, Taiwan, China. In June of 1964,
he received the degree of Bachelor of Science in Civil
Engineering from Cheng Kung University. He has been
enrolled in the Graduate School of the University of
Missouri at Rolla since January, 196.5.

123707