https://ntrs.nasa.gov/search.jsp?

R=19630003755 2019-02-03T06:17:35+00:00Z

I A LIQUID SLOSHING IN A
SPHERICAL TANK
I
,
FILLED TO AN ARBITRARY DEPTH

I
j
I .

0Y

II' *.
WEN-HWA CHU

I
I

ECHNICAL REPORT NO. 4

CONTRCICT NO. NAS8-1555
\., \
SWRI PROJECT NO. 6 - 1072-?
OTS PRICE ,

XER.OX

I5 DECEMBER 1962

II
i
j
I

SOUTHWEST RESEARCH INSTITUTE

II
SAN ANTONIO, TEXAS
I
 SOUTHWEST RESEARCH INSTITUTE
8500 C u l e b r a Road, San Antonio 6 , T e x a s

LIQUID SLOSHING IN A SPHERICAL TANK F I L L E D

TO AN ARBITRARY D E P T H

by

Wen-Hwa Chu

T e c h n i c a l Report No. 4
C o n t r a c t No. NAS8-1555
SwRI P r o j e c t N o . 6- 1072-2

Prepared for

National Aeronautics and Space Administration

George C . M a r s h a l l Space F l i g h t C e n t e r
Huntsville, Alabama

15 December 1962

A P P R 0V E D :

H. N o r m a n A b r a m s o n , D i r e c t o r
D e p a r t m e n t of Mechanical S c i e n c e s
 TABLE OF CONTENTS
Page

FORWARD iii

SUMMARY 1

INTRODUCTION 2

I. MATHEMATICAL FORMULATION 6

A. K e r n e l Function 6
B. Eigen Function 8
C. Sloshing F o r c e a n d P r e s s u r e in T r a n s l a t i o n a l
Oscil!ztinr? 9
D, The Moment Under T r a n s l a t i o n a l Oscillation 11
E. Pitching Oscillation 11

I I. NUMERICAL METHOD

A. Approximate Determination of the K e r n e l Function

at a F i n i t e Number of P o i n t s 14
B. Determination of Eigen V e c t o r s , fm’ and Natural
Frequencies 19
C. Evaluation of F o r c e 20
D. Precision Problem 21

111. EXAMPLE: FUEL SLOSHING I N A QUARTER-FULL

TANK 24

IV. CONCLUSIONS AND DISCUSSIONS 27

ACKNOWLEDGEMENT 28

REFERENCES 29

NOMENCLATURE 31

APPENDICES 34

Appendix I. Analytic E x p r e s s i o n f o r HeCp’rj 34

Appendix 11. Analytic E x p r e s s i o n f o r 3(f.f> 43
Appendix 111. Subroutines DKEF and N E F F
(WIZ P r o g r a m ) 50
Appendix IV. Derivation of nN 54
Appendix V. X - F o r c e Acting on the Tank by
Integration of P r e s s u r e 58
TABLE O F CONTENTS (Continued)

Page

FIGURES 60

1 G r a p h i c a l I l l u s t r a t i o n of Some N o m e n c l a t u r e s 60

2 S u r f a c e of I n t e g r a t i o n 61

3a Moment About Axis of Rotation 62

3b Motion of Tank i n Pitching (Rocking) O s c i l l a t i o n 63

4a C o m p a r i s o n of First Natural F r e q u e n c y with Data

by A b r a m s o n , e t al. , ( R e f . 15) 64

4b C o m p a r i s o n of the F i r s t T h r e e N a t u r a l F r e q u e n c i e s
with E x p e r i m e n t s by S t o f a n - A r m s t e a d (Ref. 1 0 ) 65

5 C o m p a r i s o n of F o r c e Response f o r Q u a r t e r - F u l l Tank
with E x p e r i m e n t s by A b r a m s o n , e t a l . , (Ref. 15) 66

6 N a t u r a l F r e q u e n c i e s Given by R i l e y - T r e m b a t h ( R e f , 1 7 ) 67

ii
 FORWARD

The work p r e s e n t e d in t h i s r e p o r t is the culmination of e f f o r t s

on the p a r t of the author that have extended o v e r a p e r i o d of
two y e a r s , with p a r t i a l support being r e c e i v e d f r o m s e v e r a l
s o u r c e s , The a n a l y s i s was originally undertaken a s p a r t of a
t h e o r e t i c a l study under Contract DA-23-072-ORD-1251 spon-
s o r e d by the A r m y Ballistic Missile Agency. The work was
continued, and l a r g e l y completed, under the p r o g r a m of i n t e r -
nal r e s e a r c h supported by Southwest R e s e a r c h Institute ( P r o -
j e c t 1059-2). Because of i t s relevance to e x p e r i m e n t a l work
a l r e a d y completed under the present p r o g r a m (NASA-MSFC
C o n t r a c t No. NAS8-1555) and published as Technical R e p o r t
No. 2 , i t was f e l t highly desirable that this w o r k should a l s o
be i s s u e d and d i s t r i b u t e d under a u s p i c e s of this c o n t r a c t .

The m a t e r i a l in this paper is also p r e s e n t e d in P a r t I , Section 3 ,

of a d i s s e r t a t i o n titled "Some Contributions to Unsteady Hydro-
dynamics in Engineering, '' submitted to the graduate faculty of
The Johns Hopkins University in p a r t i a l fulfillment f o r the
r e q u i r e m e n t s f o r the d e g r e e of Doctor of Philosophy,

iii
 LIQUID SLOSHING I N A SPHERICAL TANK F I L L E D
TO AN ARBITRARY D E P T H

SUMMARY

The k e r n e l function f o r liquid sloshing i n a s p h e r i c a l tank

fiiied to a n a r b i t r a r y depth i s shown to be r e i a t e d to the G r e e n ' s

function of the second kind and i s constructed s u c c e s s f u l l y by n u m e r -

ical means. Natural frequencies a r e then computed a s eigen values

of a m a t r i x . Eigen functions a r e obtained a t a finite number of points

a s the eigen v e c t o r s which a r e sufficient f o r approximate evaluation

of the f o r c e acting on the container. Simple f o r m u l a s of f o r c e and

m o m e n t a r e given f o r both pitching and t r a n s l a t i o n a l oscillation

under a fixed gravitational field. Finally, c o m p a r i s o n s of predicted

n a t u r a l f r e q u e n c i e s and f o r c e response with e x p e r i m e n t s f o r a

q u a r t e r - f u l l tank a r e a l s o given.

. .... L .., ,
 2 I

INTRODUCTION

Disturbances on a r o c k e t o r m i s s i l e can induce SiOS'ri'li-ig Gf

fuel in a p a r t i a l l y filled tank. It in t u r n e x e r t s excitation f o r c e s on

the vehicle and i n some c a s e s can be d e t r i m e n t a l to the t r a j e c t o r y o r

even r e s u l t s in l o s s of control. Sloshing in a c i r c u l a r cylindrical tank

h a s been widely investigated with and without damping. T o facilitate

dynamic a n a l y s i s , a n equivalent mechanical model f o r c i r c u l a r tank

i s given in R e f e r e n c e 2. F o r a s p h e r i c a l tank, a n ingenious s e m i -

n u m e r i c a l method was given i n Reference 1. However, the p r o b l e m

is only solved f o r t h r e e s p e c i a l c a s e s , namely, n e a r l y f u l l , n e a r l y

e m p t y , and half-full tanks. The r e s t r i c t i o n is due to the l a c k of the

G r e e n ' s function of the second kind (Neumann function) f o r the s p h e r -

i c a l bowl. Although the G r e e n ' s function of the f i r s t kind f o r the

s p h e r i c a l bowl is given in Reference 3 , it is doubtful t h a t a s i m p l e

e x p r e s s i o n f o r the G r e e n ' s function of the second kind e x i s t s in the

t o r o i d a l c o o r d i n a t e s , since the normal derivative on the s p h e r i c a l

c a p i s a combination of two derivatives in this coordinate s y s t e m .

The sequence method given in Reference 4 i s convergent f o r G r e e n ' s

function of the first kind but m a y diverge f o r the second kind, One

m a y r e s o r t to Liouville-Neumann method ( s e r i e s method, Ref. 5 ) and

p r o v e it converges. But when the G r e e n ' s function on the boundary

 3

is d e s i r e d , the k e r n e l function i s singular; thus it b e c o m e s i n c r e a s -

ingly m o r e difficult to evaluate when m o r e t e r m s a r e needed. If we

do not employ the Neumaim f u n c t i m , a n illtegral equation on the f r e e

s u r f a c e i s a l s o obtained. Unfortunately, the eigen functions no longer

s a t i s f y the n e c e s s a r y orthogonal relationship (Ref. 6 ) , thus they a r e

the d e s i r e d eigen functions only if the Neumann function is employed

(Ref, l ) , In this p a p e r , a n u m e r i c a l s c h e m e is devised to d e t e r m i n e

the d e s i r e d k e r n e l function, which is one component of the Neumann

function, and then apply the s a m e p r o c e d u r e a s given in R e f e r e n c e 1

to evaluate the sloshing c h a r a c t e r i s t i c s . Considerably m o r e work is

r e q u i r e d to calculate the p r e s s u r e on the wall, although in principle

this c a n be done,

After the t h e o r y in the p r e s e n t p a p e r w a s developed, s o m e

other a p p r o a c h e s have b e e n published. One approach (Ref. 7 ) s e e k s

the v a r i a t i o n a l solution b a s e d on Hamilton's principle through Rayleigh-

Ritz method':'. Since only a n integrated f r e e s u r f a c e condition was im-

posed, it i s somewhat doubtful that a c c u r a t e prediction of f o r c e r e s p o n s e

o r p r e s s u r e c a n be a s s u r e d (Ref, 8 ) , although e r r o r in the lowest mode

frequency was l e s s than one p e r cent f o r a f l a t cylindrical tank. In

another approach (Ref, 9 ) finite difference techniques w e r e employed t o

__
-~
.
L
*''
T h i s method h a s been applied to s p h e r i c a l tank by Riley and T r e m b a t h
whose r e s u l t s a r e shown in F i g u r e 6 .
 4

s e e k eigen values in a boundary condition by t h r e e different methods.

Method I and Method 111 (Ref. 9 ) use e i t h e r Rayleigh quotient o r Rayleigh-

Ritz p r o c e d u r e , but a r e somewhat i n f e r i o r (Ref. 9 ) to the Rayleigh-Ritz

p r o c e d u r e applied to the continuous domain. Method I 1 (Ref. 9 ) c o n v e r t s

the p r o b l e m into a n equivalent m a t r i x eigen value p r o b l e m by eliminating

the points outside the f r e e surface through a n i n v e r s i o n of m a t r i x if the

n u m b e r of the other points i s s m a l l , o r through a n influence coefficient

type calculation if o t h e r w i s e . In the l a t t e r c a s e if t h e r e a r e N points

on the f r e e s u r f a c e , N boundary value p r o b l e m s should be f i r s t solved

( s a y by s u c c e s s i v e o v e r - r e l a x a t i o n ) before reduction to the eigen value

p r o b l e m of a N x N m a t r i x . Depending to a l a r g e extent on the number

of net points r e q u i r e d f o r a d e s i r e d a c c u r a c y ( s a y , 3 f i g u r e s in f r e q u e n -

c i e s and f o r c e r e s p o n s e ) , the computing t i m e ( b a s e d on e s t i m a t i o n on

a G E 225 computer)* of the l a s t method f o r a s p h e r i c a l tank s e e m s t o

be c o m p a r a b l e to the p r e s e n t method. On the o t h e r hand, although

f u r t h e r (significant) a c c e l e r a t i o n of the r a t e of convergence of the s u b -

r o u t i n e s in the p r e s e n t method in the p r e s e n t p r o b l e m m a y be quite

difficult, a n a l t e r n a t i v e n u m e r i c a l s c h e m e devised i s expected to reduce

,.
4.

It i s e s t i m a t e d under the assumption that t h e r e a r e 20 f r e e s u r f a c e

points and 300 total net points with 120 i t e r a t i o n s f o r e a c h boundary
value p r o b l e m (based on experience of a s i m i l a r p r o b l e m ) and
a v e r a g e speed f o r 5 multiplications , 4 additions , and one additional
multiplication o r division a t each point in e a c h i t e r a t i o n . T h e r e a r e
o t h e r e s t i m a t e s b a s e d on experiences which yield a p p r o x i m a t e l y the
s a m e magnitude of computing time.
 5

the computing t i m e to one-half o r f u r t h e r . F i n a l l y , R e f e r e n c e 10 h a s

a l s o b e e n published in which the kernel function i s c o n s t r u c t e d e m p i r i -

cally, b a s e d on knowledge for half-fu!! ar,d f u l l tank:

The purpose of the p r e s e n t p a p e r i s mainly to p r e d i c t the n a t u r a l

f r e q u e n c i e s and f o r c e r e s p o n s e and t o show how k e r n e l functions a r e r e -

lated t o the Neumann function on the boundary and can be c o n s t r u c t e d

n u m e r i c a l l y f o r a s p h e r i c a l tank. Analogous extension to o t h e r con-

figurations o r other p r o b l e m s m a y be possible but will not be t r e a t e d

in t h i s p a p e r .
 6

MATHEMATICAL FORMULATION

A. Kernel
- --
Function

Let G(P,Q) and G,(P,Q) be the G r e e n ' s function of the second

kind f o r the i n t e r i o r of the given s p h e r i c a l bowl ( F i g . 1 ) arid the s p h e r e ,

r e s p e c t i v e l y : ( a ) Both G(P, Q ) andG,(pO Q) possess continuous second

d e r i v a t i v e s and s a t i s f y the Laplace equation inside the bowl a n d the

s p h e r e , r e s p e c t i v e l y , e x c e p t the point P=a; ( b ) 30th G and Go possess

a unit s i n k , - -I
4iY RPQ
at P - Q inside the bowl; ( c ) %=-4ru2
- /
on the
-1
whole s u r f a c e of the s p h e r e , %=fit = - on the s u r f a c e of the
AR + A r

bowl, R and F; ( d ) G, b e t h a t given i n R e f e r e n c e 7 ; G(P, Q ) satisfies

the n o r m a l i z i n g condition J G ( q Q ) d ~=, 0 (Ref. 11). Following

R+F
t h e s e conditions, i t i s well known (Ref. 1 1 ) t h a t the Neumann function G

i s s y m m e t r i c as well as Go , i. e . , G(POG?) G(Q, P ) ,Go (eQ)= C,(Q, P)

When P, Q a r e both i n t e r i o r points, analogous t o t h e proof of s y m m e t r i c

p r o p e r t i e s , one h a s

which is a n i n t e g r a l equation governing G(P, Q) w h e r e P, Q i s inside

t h e bowl, not on F and R .

 7

F o r values of the G r e e n ' s function with P, Q (P#Q) both on F ,

not on R , apply d i r e c t l y the divergence t h e o r e m to the s u r f a c e shown

in F i g u r e 2. Since t h e r e is an infinitesimal s e m i - s p h e r e a r o u n d the

sinks at P and Q r e s p e c t i v e l y , one finds

121
P#r*Q
By making P and Q in Equation [ l ] a p p r o a c h P and Q on the f r e e s u r -

f a c e along i t s n o r m a l , Equation 117 c a n be r e d u c e d to Equation [ 2 ] .

In R e f e r e n c e 1 , f o r fuel sloshing in a s p h e r i c a l tank, only those

e i g e n functions proportional to cos 8 a r e needed: one s h a l l s e e in the

next section t h a t it 'is sufficient to know one component H(P, Q )

-- of the

Neumann function G(e (3) to d e t e r m i n e the sloshing c h a r a c t e r i s t i c s .

Let

Since G and G, a r e s y m m e t r i c functions, /-/ , H, and thus 4, are

s y m m e t r i c functions,
 8

- -
F o r points P , Q corresponding to P and Q r e s p e c t i v e l y , inside

the s p h e r i c a l bowl, Equation [ l ] can be i n t e g r a t e d t o yield

A, (eQ > =
/

-j- 3 (7a) H, (P, T ) d< -[%(E a) hl P,2) d<

F 7
f o r which the r e v e r s i n g of o r d e r s of integration a r e appiied and c a n be

justified by c a r r y i n g out the details. The function F is defined by

which i s a n o n s y m m e t r i c function a s -
a6

S i m i l a r l y , if both P and Q a r e on F , not on R , integration of

%=O , a l m o s t e v e r y w h e r e on F , hence f o r h a l f - s p h e r e hl(PPj=OJ #=He

which i s in a g r e e m e n t with Reference 1.

B. Eigen Functions

The eigen functions 6 a r e a s s u m e d to p o s s e s s the following

properties: (a) 4 i s r e g u l a r inside the bowl and V% =0

The l a s t condition i s a p p r o p r i a t e f o r t r a n s l a t i o n a l oscillation of the tank.

 9

Analogous to Equations[l] and [21

~ { P =) / G ( T P > % 5$(dds 17 3
F
when P is inside the s p h e r i c a l bowl, R + F.

5 #,(P)= {F G ( I , P) 2 $(I) dSz 18 1

when P i s on F.

Analogous to Equations 141 and 153 by integration

%(a=[fl(3 PI t m dG -
, P not on F.
F
b -
+g(p)= c/(p.rj zCpip+‘ , P on F.
0

--
T h i s shows only H ( P , Q ) i s needed f o r the p e r t i n e n t eigen functions.

C. Sloshing F o r c e-
and P r e s s u r e in T r a n s l a t i o n a l Oscillation

By introducing a displacement potential r e l a t i v e to the tank

%-2%(f)$(c$’&the sloshing f o r c e acting on the container i s derived

f r o m the Lagrangian’s equation i n R e f e r e n c e 1, namely

 10

The velocity potential

The p r e s s u r e on the container

within the a c c u r a c y of the linearized theory. Equation Ill] can also

be obtained d i r e c t l y by integration of p r e s s u r e (Appendix V ) .

Once 6 on F i s evaluated, one m a y employ ci, to obtain

+.,(PI from

%(P)=
F
d ( f , ~X) g(r)dg -J- ~ ( PIr ,g ( l ) c / ~
F

The i n t e g r a l on R dropped out a s

a&
-=0 dG = constant
on R and 2
3%
on R.

F o r P on R , not on F , the integrands of the i n t e g r a l s in

Equation [14b] a r e nonsingular, hence $ (P) c a n be calculated by

well-known n u m e r i c a l methods. F o r contact points both on R and F ,

the value of j$(fi) m a y be obtained by evaluation of the i n t e g r a l by

midpoint f o r m u l a .
 11

D. The Moment Under Translational Oscillation

-

F o r t r a n s l a t i o n a l oscillations, the velocity potential is propor -

tional to cos 6 and the flow i s a n t i s y m m e t r i c . It produces a horizon-

tal force < in the x - d i r e c t i o n and a couple C, about the c e n t e r of

the tank ( F i g . 3 a ) . T h e r e is no moment around z a x i s o r x a x i s by

symmetry. The m o m e n t about a fixed point 0’ on the z a x i s i s

It is not n e c e s s a r y to d e t e r m i n e 1 when the f o r c e

x
6 and the m o m e n t

a r e the d e s i r e d information in dynamic p r o b l e m s . F o r a sphere,

mt
all the p r e s s u r e f o r c e s acting on the s h e l l p a s s e s through its c e n t e r ,

hence produces no m o m e n t about it, i. e . ,

T h e r e f o r e the m o m e n t about 0’ i s s i m p l y

T h i s s t a t e m e n t c a n be e a s i l y shown by integration of the m o m e n t s due

to p r e s s u r e on the wall.

E. Pitching. Oscillation

Consider a pitching oscillation of amplitude 9 around a n a x i s

which is p a r a l l e l to y a x i s and at a v e r t i c a l distance Q

. below y a x i s

(Fig. 3b). In F i e u r e 3b. it i s c l e a r that

 The r a d i a l distance of any point ( x , y , z ) f r o m the axis of r o -

tation isd- . The velocity components on the s p h e r e due t o

rotation a r e :

The boundary condition on the wetted s p h e r e , R , i s

= &J ( S;@+ case)

T h i s i s equivalent to a translational oscillation of amplitude /J=$i
in the direction of x. Since the boundary condition on the f r e e s u r f a c e

i s the s a m e in the p r e s e n c e of a fixed gravitational f i e l d , the r e s u l t f o r

t r a n s l a t i o n a l oscillation c a n be applied. T h e r e is a n additional s t a t i c

tipping f o r c e which c a n be obtained by integrating the additional s t a t i c

pressure p’ o v e r R (Ref. 14)

/
P=PYXBy

T h i s f o r c e a c t s along a n x axis rotating with the tank.

 13

S i m i l a r l y , t h e r e is a n additional moment

The total f o r c e along a n axis x rotating with the tank is

The m o m e n t about 0’ i s approximately

The total f o r c e in the horizontal direction i s s t i l l 6 When t h e r e i s

tank fixed axial a c c e l e r a t i o n , the method of superposition p r e s e n t e d in

R e f e r e n c e 14 c a n be used to determine the x - f o r c e .

An equivalent mechanical mQdel f o r sloshing i n s p h e r i c a l tank

is given in R e f e r e n c e 15, but unfortunately the extrapolation to include

damping was not a s s u c c e s s f u l a s in the c a s e of a c y l i n d r i c a l tank

(Ref. 2 ) and could only be used f o r o r d e r of magnitude e s t i m a t e s (Ref. 15).

 NUMERICAL METHOD

A. --a t a
Approximate Determination of the K e r n e l Function
F i n i t e Number of Points

N u m e r i c a l q u a d r a t u r e formula will be used t o r e p l a c e the i n -

t e g r a l Equation [67 by a m a t r i x equation. T h e r e i s a m i n o r difficulty

- -
due to the p r e s e n c e of l o g a r i t h m ’ s s i n g u l a r i t y a t P = I or -6 ’
fi. -
t h e l a t t e r of which i s the integration v a r i a b l e . In the o r i g i n a l m a n u -

s c r i p t , a n a t t e m p t w a s m a d e to devise a m o r e sophisticated q u a d r a -

t u r e f o r m u l a , expecting higher a c c u r a c y . Unfortunately, i t s e e m s t o

contain i n t e g r a l s difficult t o e x p r e s s i n known functions, o r r e q u i r e

v e r y c a r e f u l p r o c e s s of taking limit under the i n t e g r a l s i g n s . F u r t h e r ,

the a p p a r e n t higher o r d e r t e r m s may b e actually v e r y l a r g e and not

negligible. To reduce total effort, the p r e s e n t n u m e r i c a l s c h e m e

b a s e d on midpoint f o r m u l a i s devised.

The i n t e g r a l s a r e divided into N equal p a r t s ( N = 2 0 will b e

u s e d ) and the field point i s one of the c e n t e r s of t h e i n t e r v a l s . A

s i m p l e midpoint f o r m u l a will not be applicable when the logarithmic

s i n g u l a r i t y a p p e a r s a t the midpoint, but if the i n t e r v a l i s subdivided

into f o u r i n t e r v a l s ( o r m o r e ) the e r r o r m a y become a c c e p t a b l e . For

e x a m p l e , c o n s i d e r the i n t e g r a l
 15

= A&& - ?.5;30/ A
rw
The e r r o r is . 1 6 6 , F o r d = 1 / 2 0 ( N = 2 0 , b = l ) , the r e l a t i v e e r r o r

i s l e s s than 0.1870.

Let b-1,

(note the o r d e r of i & j ) 9.3

then Equation 161 can be rewritten a s

Two s i m i l a r n u m e r i c a l s c h e m e s will be p r e s e n t e d . The f i r s t s c h e m e

was actually employed in the example, while the second s c h e m e is the

a l t e r n a t i v e s c h e m e r e q u i r i n g much l e s s computer t i m e . In the f i r s t

s c h e m e one evaluates 3dFn. and 8'

'd
a t N x 4N points, a s s u m i n g four

point midpoint f o r m u l a . ith point on the f r e e s u r f a c e i s located a t the

midpoint of the ith i n t e r v a l ( i = 1 , 2 , ... , N). j r e p r e s e n t s the i n t e -

gration variable located a t the midpoint of the s u b i n t e r v a l ( j = 1 , 2 , . ~ ~ ,4N).

 16

Thus

T o d e s c r i b e the second ( a l t e r n a t e ) s c h e m e , c o n s i d e r the whole

s q u a r e domain to be c o m p o s e d of N x N s q u a r e s u b d o m a i n s . In all t h e

diagonal s q u a r e s , 3.- (0 )

d
and e. (0)

t t
a r e evaluated a s in the f i r s t s c h e m e

(four values in e a c h s q u a r e ) , but they will t a k e the value of the functions

a t the c e n t e r of e a c h s q u a r e in off-diagonal d o m a i n , which a r e a l s o e v a l -

uated. T h e s e d a t a w i l l b e denoted by 7
9.
-(b)
, H..
v
. Total n u m b e r of

evaluation (both 4(6).

d
and $'6 ) a r e N x 4N i n the f i r s t s c h e m e , b u t

N x N + 3N in the s e c o n d s c h e m e ( 4 point midpoint f o r m u l a f o r diagonal

i n t e g r a l ) . F o r instance consider

In the f i r s t s c h e m e , being given by Equations ( 2 la], 12 lb],

P -pk
 ~ ~~

17

given by Equation

Equations [23a] - [23d] c a n be condensed into the single f o r m u l a with

f , given by Equation [21b] and one finds

d

S i m i l a r l y , the second i n t e g r a l on the r i g h t s i d e of Equation 1 2 0 1 is

where
 In the second s c h e m e , replace % (0)
by si*
(*)
and 44. (0)
by
-
ey to)
,

The i n t e g r a l Equation [ 6 ] i s t h e r e f o r e reduced approximately to the

m a t r i x equation,

f[C]=-[M3 --IC] [D] [Z 52

I
the solution of which i s I

(11 i s a unit m a t r i x , [ C ] , [MJ, ID] a r e s q u a r e m a t r i c e s of which the

e l e m e n t s of ith r o w and lth column a r e C y , MV , DEp , r e s p e c t i v e l y .

As a check of a c c u r a c y , the s y m m e t r i c p r o p e r t y =c/.i

C,6..
bl
should

hold approximately, Then we can u s e the a v e r a g e value f o r the c o r -

r e c t i v e t e r m in the k e r n e l function, i. e . ,

The k e r n e l function i s t h e r e f o r e

1273
-
c.. being known a t d i s c r e t e points c o r r e s p o n d i n g to both
d d

given by equation [Zla] . The difficulty of the p r o b l e m , however, l i e s

i n the a c c u r a t e and r a p i d evaluation of the function 3(d

v
and e&- (0)
.
B. Determination of Eigen Vectors, yfn’ and-N- atural Frequencies

The eigen function t a k e s the value E ( f ) on F , which is governed

by

There

-
JZ,= G2Q
9
,9 being effective gravitational a c c e l e r a t i o n

Analogous to Section A , the m a t r i x approximation of Equation 1291 i s

where the f a c t o r 1 / 2 on the right-hand side is in a g r e e m e n t with R e f e r -

ence 1, since the s t r e n g t h of the G r e e n ’ s function h a s not been doubled

in t h i s p a p e r . The e l e m e n t s of the m a t r i x A i s

where i , k c o r r e s p o n d s to /9., &, both given by Equation [21a] and

both v a r y f r o m 1 to N.
 20
I

(0)
In the f i r s t s c h e m e , ff{# i s not evaluated at the c e n t e r of any s q u a r e

subdomain t o r e d u c e computing time and is a p p r o x i m a t e d by

I
where

(property of s y m m e t r y )
I (0, . (0)
i n the second s c h e m e , r e p i a c e dq by A$ j i = l to N , j = 1 to 4Nj.

The l a r g e s t eigen values N of Equation [30] yields the

b
2J2,x
l e a s t r e s o n a n t frequency p a r a m e t e r s . 3, . And the eigen v e c t o r s will ~

be employed in evaluation of the force r e s p o n s e . ~

C. Evaluation
- - of F o r-
ce

The sloshing f o r c e f o r translational oscillation i s

where

-1
13
 21

The nondimensional f o r c e

D. Precision Problem

The functions 4 ,H, have been f i r s t e x p r e s s e d in t e r m s of

complete e l l i p t i c a l i n t e g r a l s of the f i r s t kind, the second kind, and

the t h i r d kind and of s i m p l e e l e m e n t a r y functions (Appendix I , 11).

The e l l i p t i c a l i n t e g r a l s of the third kind a r e e x p r e s s e d in t e r m s of

Heuman’s L a m d a function A,, (Ref. 12), which i s again e x p r e s s e d

e i t h e r in a s e r i e s f o r m o r in a close f o r m of incomplete and c o m -

plete e l l i p t i c a l i n t e g r a l s of the f i r s t and second kind, i. e . ,

In 3j and HO3 , a s e r i o u s precision p r o b l e m o c c u r s due to a l m o s t

complete l o s s of significant f i g u r e s in s u b t r a c t i o n s f o r f i , fd - both

small. At f i r s t , the s e r i e s f o r m of the L a m d a function was u s e d ,

but it was found that the s e r i e s i s very slowly convergent when the

p a r a m e t e r i s n e a r unity, especially if double p r e c i s i o n o r twelve

significant f i g u r e s a r e sought. Then i t i s r e s o r t e d to the i t e r a t i v e

methods f o r evaluating elliptic i n t e g r a l s (Ref. 13), which c o n v e r g e s

to within f o u r o r five iterations. Although the complete elliptic

i n t e g r a l s c a n be computed v e r y rapidly, the subroutine N E F F (Ap-

pendix 111) f o r incomplete integrals and a difference r e l a t e d to i t

 22

c o n s u m e s 8 seconds (used t w i c e ) , while the total t i m e f o r evaluating

J,, H, i s only 25 secords at each point, a l l on the GE 225 c o m p u t e r .

Longer t i m e would be r e q u i r e d f o r higher p r e c i s i o n a s the n u m b e r of

iterations increase s .
T o i n c r e a s e the p r e c i s i o n , analytic s u b t r a c t i o n s a r e m a d e SO

that no significant s u b t r a c t i o n r e m a i n s , if possible. Noniterative sub-

t r a c t i o n s in which f o u r o r l e s s figures a r e l o s t a r e acceptable if f o u r

o r m o r e significant f i g u r e s out of eight (single regular p r e c i s i o n on

the m a c h i n e ) i s d e s i r e d . The technique c a n be i l l u s t r a t e d by the

I
following c a s e s :

(1) L e t (A - B ) , the difference of A and B i s s m a l l but c a n be

e x p r e s s e d analytically without subtraction. Then, f o r ex-

ample,
I
- --
1
should be evaluated f r o m -
-/ (A-0)
JiiG nfi JX+6

(2) Let & I s be small (positive) quantities containing no sub-

t r a c t i o n , then (/+$,)(/+i&) ....-(/t&,,n)

-/ should be evaluated

by repeated application of the s i m p l e r e l a t i o n that

(I+A)(/*Rd-/ =&,it& f4g2

(3) T o s u b t r a c t a d e s i r e d quantity f r o m a known function m a y

r e q u i r e a new subroutine f o r this function p e r f o r m i n g s i g -

nificant subtraction analytically, e. g., N E F F . (Appendix 111)

 23

Aside f r o m relatively mechanical operations, the device of DKEF and

N E F F s u b r o u t i n e s , the following relation was expedient (Appendix IV).

/7N and a r e defined by Equations [IV-63, [IV-81, r e s p e c t i v e l y .

It i s noted t h a t , a f t e r a small manipulation, d i r e c t n u m e r i c a l

integration of the i n t e g r a l s 5 ,Hod a t sampling points of the e n t i r e

domain of fL., p6 w a s a l s o computed by Weddle's r u l e . Although f o u r

o r m o r e significant f i g u r e s can be obtained, it i s d e e m e d too slow o v e r

the m a j o r p a r t of the domain. F o r i n s t a n c e it took about 5 and 2 - 1/ 2

minutes r e s p e c t i v e l y f o r 6 and e3 on a GE 225 computer with 384

i n t e r v a l s , o r a r e l a t i v e e r r o r of about otherwise a t a point n e a r

the right lower c o r n e r of the domain ( ~ , p dnear unity). T h e s e values

a t sampling points a r e valuable a s they s e r v e a s a good check on the

p r e s e n t c o m p u t e r p r o g r a m , which evaluates F I , F2, F 3 , H O I , HO2,

H03, a l l together a t a r a t e of 25 seconds p e r net point (on the s a m e

c o m p u t e r with a n a c c u r a c y of four o r m o r e significant f i g u r e s ) .

 24

E X A M P L E : F U E L SLOSHING I N A QUARTER-FULL TANK

First, 30 and Hoare generated, then the m a t r i x Equation c 2 5 1 is

solved, The c o r r e c t i v e p a r t Ci. to the k e r n e l function obtained i s s y m -

d
m e t r i c a l m o s t to four f i g u r e s (Table I ) . The relative e r r o r s i n the s a m p l e

points a r e l e s s than 0.370 o r b e t t e r . Since t h e s e values a r e quite r e p r e -

sentative, the values of c.d a t other points a r e not shown in the table.

Next, the eigen values and eigen v e c t o r s of Equation 1307 and then

the f o r c e r e s p o n s e of Equation [31] a r e calculated. The calculated f i r s t

f o u r e i g e n values a r e 9.48863, 2.0591201, 1.2003387, 0.84773955, re-

spectively, The c o r r e s p o n d i n g frequency p a r a m e t e r s a r e c o m p a r e d with

e x p e r i m e n t s in F i g u r e s 4a and 4b. It s e e m s that the values a r e well within

possible e x p e r i m e n t a l e r r o r , although it m a y be slightly l e s s than the

actual value, noting that n a t u r a l frequencies a r e somewhat s m a l l e r f o r

l a r g e r amplitudes of oscillation.

The constants needed to calculate the f o r c e r e s p o n s e a r e c o m p a r e d

with graphical values given by Reference 1 in Table 11. Since the coefficient

9;/c; i s i n a g r e e m e n t with Budiansky’s value, the m a i n difference l i e s

i n the value of f i r s t n a t u r a l frequency f o r frequency r a n g e in i t s neighbor-

hood. Since graphically interpolated value i s l e s s r e l i a b l e , which i s a l s o

c o n f i r m e d experimentally in this c a s e , only the p r e s e n t theory is c o m p a r e d

with e x p e r i m e n t s (Ref. 15) in Figure 5. The d i f f e r e n c e between t h e o r y and

e x p e r i m e n t s , p e r h a p s , is essentially due to finite amplitude effect. But

the a g r e e m e n t s e e m s to be quite reasonable.

 25

w
k

a
>
.-I
c-,
V
al
s; jr; k
k
0
u
w
0
m
al

W
..

L
N 0
4
I=
 26

n l s,, = --ct
14;

- L- , I Y
I
I3udi 2: sky I Budiarisky
Chu I ( F i g . 9 , Ref. 1 ) Chu ( F i g . 10, Ref. I )
--

0.249586

2 I 5.6077376 I 5.4443 1
1
0.44687 x 10-3 0.4172 x 10-3
1
-i- - -- - - -- -- -__-___ __ - -_ -_I_ --_ -

3 9.61'37893 9.2325 0.44758 x 0.7664 x

1
I

'
' I
4 , 13.620935
I
--- 0. 1 1 2 6 3 3 x ---
I I A

T A B L E 11: Comparison of C o n s t a n t s w i t h Data

from R e f e r e n c e I
 27

CONCLUSIONS AND DISCUSSIONS

The p r e s e n t theoyy arid compiiter prograr;; see: t~ yield

s a t i s f a c t o r y predictions of n a t u r a l frequencies and f o r c e r e s p o n s e

in c o m p a r i s o n with e x p e r i m e n t s for a q u a r t e r - f u l l s p h e r i c a l tank.

The c o m p u t e r p r o g r a m i s expected to be applicable to other liquid

d e p t h s , although not beyond improvement in efficiency. The r e s u l t s

a l s o c o n f i r m the theory that the kernel function i s r e l a t e d to the

Neumann function on the boundary and that this function c a n be con-

s t r u c t e d by adding a c o r r e c t i v e part to a known G r e e n ' s function

n u m e r i c a l l y f o r p r a c t i c a l applications. Extensions to o t h e r p r o b -

l e m s m a y be possible, but one m u s t r e s o l v e the p r e c i s i o n p r o b l e m

if it e x i s t s and one m a y a l s o find a m o r e sophisticated n u m e r i c a l

s c h e m e t o be m o r e d e s i r a b l e , either in a c c u r a c y o r in efficiency.
 28

ACKNOWLEDGEMENT

The author would like to e x p r e s s h i s deep a p p r e c i a t i o n t o

R o b e r t Gonzales f o r h i s industrious e f f o r t in p r o g r a m m i n g a s well

a s checking m o s t p a r t s of the p r e s e n t t h e o r y ; and to T h o m a s J a c k s o n

f o r h i s consultation and a s s i s t a n c e in p a r t of the p r o g r a m m i n g . A l s o ,

the author would like t o e x p r e s s his a p p r e c i a t i o n t o D r . Thein Wah

and William S q u i r e f o r t h e i r informative d i s c u s s i o n s , Finally, the

author would like to thank D r . H. N . A b r a m s o n f o r h i s d i s c u s s i o n s

and h i s continued s u p p o r t , without which the p r e s e n t r e s e a r c h could

not have been completed.

 29

REFERENCES

1. Budiansky, B. , "Sloshing of Liquids in C i r c u l a r C a n a l s and

S p h e r i c a l Tanks," J. A e r o . S c i . , Vol. 27, No. 3, pp. 161-
173, M a r c h 1960.

2. A b r a m s o n , H. N. , Chu, W . H . , and R a n s l e b e n , G. E . , J r . ,
"Representation of F u e l Sloshing in C y l i n d r i c a l T a n k s by a n
Equivalent Mechanical Model," Tech. Rept. No. 8 , C o n t r a c t
N o . DA-23-072-ORD-1251, SwRI P r o j e c t No. 43-768-2,
Southwest R e s e a r c h Institute, June 1960. A l s o , ARS J o u r n a l ,
Vol. 31, No. 12, pp. 1697-1704, D e c e m b e r 1961.

3. B a t e m a n , H. , "Partial Differential Equations of Mathematical

P h y s i c s , ! I Dover Publications, New York, pp. 46 1-468, (1944).

4. Boley, B. A. , "A Method f o r the Construction of G r e e n ' s

F u n c t i o n s , " Q . Appl. Math., Vol. X I V , No. 3 , pp. 249-257,
October 1956.

5. Whittaker, E. T . , and Watson, G o N:, "Modern A n a l y s i s , "

A m e r i c a n E d . , MacMillan C o . , New York, pp. 221-222 (1948).

6. T r e m b a t h , N. W . , "Fluid-Sloshing in T a n k s of A r b i t r a r y
Shape,'' Rept. GM45.3-378, Space Technology L a b o r a t o r i e s ,
I n c . , L o s Angeles, C a l i f . , p. 15, (1957).

7. L a w r e n c e , H. R . , Wang, C . J . , and Reddy, R . B . , "Variational

Solution of F u e l Sloshing Modes," Jet P r o p u l s i o n , Vol. 28, NO. 11
PP,. 729 - 7 3 6 , November 1958.

8. Chu, W . H . , "On the Role of N a t u r a l Boundary Conditions i n the

Rayleigh-Ritz Method," (to a p p e a r in - - ).
J. Appl. Mech.

9. E h r l i c h , L. W . , Riley, J . D . , S t r a n g e , W . G . , and T r o e s c h ,
B. A . , "Finite-Difference Techniques f o r a Boundary P r o b l e m
with a n Eigenvalue in a Boundary Condition," J. SOC. Indust.
Appl. Math, Vol. 9 , No. 1, M a r c h 1961.

10. Stofan, A. J . , and A r m s t e a d , A . L. , "Analytical and E x p e r i -

m e n t a l Investigation of F o r c e s and F r e q u e n c i e s Resulting F r o m
Liquid Sloshing i n a Spherical T a n k , ' ' NASA TN D-1281, July
1962.
 30

11. B e r g m a n , S . , and Schiffer, M . , " K e r n e l F u n c t i o n s a n d

Elliptic Differential Equations in M a t h e m a t i c a l P h y s i c s , I '
A c a d e m i c P r e s s , N e w York, (1953).

12. S y r d , P. F . , and F r i e d m a n , M. D. , "Handbook of Elliptic

I n t e g r a l s f o r E n g i n e e r s and P h y s i c i s t s , " S p r i n g - V e r l a g
B e r l i n , (1954).

13. Hancock, H. , "Elliptic I n t e g r a l s , " Dover P u b l i c a t i o n s , Inc. ,

1 s t Edition, New York (1958 p r i n t i n g ) .

14. Chu, W. H . , " F r e e S u r f a c e Condition for Sloshing R e s u l t i n g

f r o m P i t c h i n g and Some C o r r e c t i o n s , " ARS J o u r n a l , Vol, 30,
No. 11, pp. 1093-1094, November 1960.

15. A b r a m s o n , H . N . , Chu, W . H . , and G a r z a , L. R . , "Liquid

Sloshing in S p h e r i c a l Tanks," Tech. Rept. No. 2 , C o n t r a c t
N o . NAS8-1555, SwRI P r o j e c t No. 6-1072-2, M a r c h 1962,
( t o a p p e a r in ARS J o u r n a l ) .

16. Kellog, 0. D. , "Foundations of P o t e n t i a l T h e o r y , " F r e d e r i c k

Ungar Publishing C o . , New York, p. 257.

17. Riley, J. D . , a n d T r e m b a t h , N . W . , "Sloshing of Liquids in

S p h e r i c a l T a n k s , " J. A e r o / S p a c e S c i e n c e s , V o l . 28, No. 3 ,
I

pp. 245-246, M a r c h 1961.

 NOMENCLATURE

a = r a d i u s of the s p h e r i c a l tank

'F
-- 'rea of undisturbed free s u r f a c e

a r e a of wetted s u r f a c e of s p h e r e
AR =

b = m a x i m u m value of p , r a d i u s of f r e e s u r f a c e

d = tank d i a m e t e r , 2a

( c . f . Appendix 111)

D K E F = a function in the computer p r o g r a m ( c . f . Appendix 111)

F = the u n d i s t u r b e d f r e e s u r f a c e

6 = h o r i z o n t a l f o r c e acting on the tank due t o f u e l s l o s h i n g

.r,(p,pj = i n t e g r a t e d k e r n e l function r e l a t e d to
J = effective gravitational a c c e l e r a t i o n

87 - h,(pz.fp
G(eQ) = G r e e n ' s function of the second kind f o r the s p h e r i c a l bowl

G.(f?Q) = G r e e n ' s function of the second kind f o r a s p h e r e

h(t?Q) - additional p a r t of G r e e n ' s function f o r s p h e r i c a l tank o t h e r

than half -f ull

h,(Pf') = i n t e g r a t e d k e r n e l function r e l a t e d to /t (f? Q)

H ( / ?f?) = i n t e g r a t e d k e r n e l function r e l a t e d t o G ( P , 0)
 32

point of i n t e g r a t i o n , e x c e p t I being t h e unit n i a t r i x

c o m p l e t e elliptic i n t e g r a l s of f i r s t and second kind,

r e s p e c t ive l y

t o t a l m a s s of liquid (fuel)

n = outer normal

a function i n the computer p r o g r a m ( c ,f . Appendix I I I )

( c . f . Appendix 111)

a r i n g c o r r e s p o n d i n g to p ( C $!,e)

2 W
9, =
p t P’
2a
9, =
./cpP-bzf+$(f -f’f+4pph‘
6 ,f analogous t o P but r e l a t e d to Q and I, r e s p e c t i v e l y

Qs defined by Equation [I- 1Oa]

spherical coordinates

the wetted s p h e r i c a l s u r f a c e b e f o r e s l o s h i n g u n l e s s defined b y 11-21

defined by Equations [I-lOh] , [I-lob] , [11-5b], respectively

d i s t a n c e between the points P and Q

e l e m e n t of s u r f a c e

dshe + f d f on F
h o r i z o n t a l displacement of c o n t a i n e r in the x - d i r e c t i o n

v e r t i c a l distance of f r e e s u r f a c e f r o m c e n t e r of s p h e r e ;
positive upward

defined by Equation [12b] , 1&P(I)d%

F
 33

Bn defined by Equation [lZa]

c o s 2 = angle between the v e c t o r s 9a n d op,(

, respectively

= H e u m a n ' s l a m b d a function (Ref. 9 )

= c o m p l e t e elliptic integrals of the t h i r d kind (Ref. 9 )

= r a d i a l d i s t a n c e f r o m a point on the f r e e s u r f a c e to the

cep~ternf the f r e e surface

- P of i n t e g r a t i o n variable

= d e n s i t y of liquid (fuel)

= velocity potential, v$"= $ , 9 being the velocity vector

= nth e i g e n function

= nth i n t e g r a t e d eigen function r e l a t e d t o 4

= q SPnp
= f r e q u e n c y of oscillation

= nth r e s o n a n t frequency
'= Un2Qh , nth resonant f r e q u e n c y p a r a m e t e r

S
--u-b-s c r i p t s
F related to surface F

i,j,k related to pi, 4, f# , respectively

I r e l a t e d t o integration v a r i a b l e s

P r e l a t e d to the point P(G$?!e) o r P(&)

Q r e l a t e d to the point Q(q $ 8 ) o r a(/&)

R related to surface R
 APPENDIX I. ANALYTIC EXPRESSION F O R 11, r
(f; ' ) ~

The G r e e n ' s function of the second kind f o r a whole s p h e r e I

(Ref. 16) i s I

whe r e i

When P and P' both on F , I

Making u s e of a new v a r i a b l e /s= *

F--
cT
2
and a new p a r a m e t e r P,= -
247
P+P'
 35

where

[I-6a]

where D(,9,) is D K E F ( 3 ) given b y Appendix 111.

By taking limiting p r o c e s s ,

[I-Sb]

Similarly

[I-77

where

Also, t h i s c a n be evaluated m o r e a c c u r a t e l y by

[I-7a]
 36

r
When p or f is z e r o

From i n t e g r a t i o n by p a r t s ,

r 1-81
 37

where

By u s i n g p a r t i a l f r a c t i o n s , the variable and the definition of r ( d ,2g z )

( c . f . Ref. 9),one finds

 [I-8b]

I
where the well-known complete elliptic i n t e g r a l of the t h i r d kind i s

defined b y

11-91
 39

where

Q1-h2 = z;
Hence
 /
When f‘ o r f i s zero,

It was found that t h e r e is a p r e c i s i o n p r o b l e m i n Equationl1-8j

when P and
/
a r e near zero. T h i s might be anticipated a s t h e r e i s

a v e r y s m a l l denominator proportional to ‘ff’ ’* and the r e s u l t i s

expected to be s m a l l in view of Equation [I-S’J, After somewhat l a b o r -

ious manipulations with Equation [ I - 8 9 to r e s o l v e the p r e c i s i o n p r o b l e m ,

i s obtained n the following f o r m (with b = 1):

[I-101

where

(I- 1Ob]

[I- 1oc)
 [I - 1od3

[I-lOf]

[I-lOh]

R4=-2 ?&- [I-lOi]

A, > I , A= 40
[I-lojl

It m u s t be noted that (A,-/,) ,(Ai-/),(fl2-1) a r e evaluated not

by d i r e c t s u b t r a c t i o n s , but by accurate f o r m u l a s .

[I- 1Ok]

1
 42

[I- 1 0 11

(p-f') can be calculated without loss of significant f i g u r e . It i s

a l s o noted that f o r s m a l l zF , one should r e p l a c e (h4-6'a") = L2(1c4t)

by -$b or (1-a'~ by -$ with b =/.

It i s r e c a l l e d that Budiansky's technique of differentiation

under i n t e g r a l s i g n does not s e e m to l e a d to s i m p l e r e s u l t s , due t o

the p r e s e n c e of n o n - z e r o IF , the r e l a t i v e depth m e a s u r e d f r o m the

c e n t e r of the s p h e r i c a l tank.
 APPENDIX I I. ANALYTIC EXPRESSION FOR 3@,9 j

The o u t e r n o r m a l derivative of G,(P,Pj on the f r e e s u r f a c e i s

[II- 11

= 3+J,+33 [II - 31

Using the s a m e technique .as in Appendix I, one f i n d s

[II-4]

where Q3- TZU = Q(b’-f’) 1

 44

It i s noted t h a t a s p e c i a l c a s e of the reduction f o r m u l a in R e f e r e n c e 1 3

c a n be used to evaluate the following i n t e g r a l , which o c c u r r e d in $(p,p?

Or, u s e

which c a n be checked e a s i l y by differentiation.

When p orp' is zero,

To i n c r e a s e a c c u r a c y in n u m e r i c a l evaluation, Equation [II-41 is

r e p l a c e d by (with b = 1 )

[TI - 4b3

b-p 9 p-{ and 6?-pp' can be evaluated a c c u r a t e l y f o r known

d i s c r e t e values of p-p' and is evaluated by

(11-4c]

Next, 5 will be e x p r e s s e d in t e r m s of e l e m e n t a r y functions.

 45

Some f u r t h e r manipulation i s required t o avoid p r e c i s i o n p r o b l e m s

for p, p' small or + small. One finds

Finally, 3 will be e x p r e s s e d i n closed f o r m a s follows;
 47

[I1 - 6)

T h e r e i s a s e r i o u s precision p r o b l e m f o r f,p /
small in

Equation 111-6). After manipulations, the p r e c i s i o n p r o b l e m is r e -

solved by employing the following equivalent f o r m .

where
 48

[I1 - 7 b]

[II-7c]

XF' fG,
Where the t e r m s in the i n n e r bracket could be r e p l a c e d by
z; tp2+ /g*
f o r higher p r e c i s i o n , which s e e m s u n n e c e s s a r y as the e r r o r in 8, is

sufficiently small in the c r i t i c a l range due to the f a c t o r f'tf' .

[II-7d]

[I1 - 7 e]
 49

where f u r t h e r manipulation m a y be needed f o r v e r y small ZF to

avoid p r e c i s i o n p r o b l e m s i n the domain of s m a l l p and p’ ,

[II-7 f ]

?do, 4
/

When p or + 0.

It i s i m p o r t a n t to note that whether p >?’ or p>p the s u m of

j: , 4 and & always approaches z e r o a s + + 0. Therefore,

f o r a half-full tank /+=He which is in a g r e e m e n t with Budiansky’s

k e r n e l function a s i d e f r o m a n apparent f a c t o r of two difference

mentioned previously ,
 50

APPENDIX 111. SUBROUTINES DKEF AND N E F F ( W I Z PROGRAM)

DKEF = DKEF (k,k, 1 , 1, 1 )

N E F F = NEFF ( ~ , X , X ' , l , l )
-.
l h e unity a r g u m e n t s a r e actually d u m m i e s , while the five a r g u -

m e n t s r e p r e s e n t five outputs. F o r D K E F , the outputs a r e K ( k ) = DKEF ( l ) ,

E ( k ) = DKEF ( 2 ) , ( E ( k ) - K ( k ))/k2= DKEF ( 3 ) , K ( k ) - W / 2 = D K E F ( 4 ) , and the

n u m b e r of i t e r a t i o n s = DKEF ( 5 ) . DKEF ( 3 ) i s not obtained f r o m DKEF (1)-

DKEF ( 2 ) but i s obtained a f t e r a significant analytic s u b t r a c t i o n i n the p r o -

gram.
v F o r N E F F the outputs a r e F( p , k - k ) = N E F F ( l ) , E ( &, ~ IC = K )

= N E F F (21, ( E ( f , k ) - k s i n p ) = N E F F ( 3 ) , the n u m b e r of i t e r a t i o n s f o r

evaluating F ( ,k)= N E F F ( 4 ) , the number of i t e r a t i o n s f o r evaluating

N E F F ( 3 ) = N E F F ( 5 ) . N E F F ( 3 ) i s evaluated a f t e r a significant analytic

s u b t r a c t i o n in the p r o g r a m while N E F F ( 2 ) is s i m p l y obtained f r o m NEFF(3)

$- k sin? . Although k' does not a p p e a r in the functions sought,

it i s calculated f r o m a f o r m u l a without s u b t r a c t i o n , a s one c a n e a s i l y s e e

significant f i g u r e s of k1 would be lost if k i s n e a r unity. The b a s i c f o r m u l a e

a r e a l l given in R e f e r e n c e 13. F o r complete elliptic i n t e g r a l s , the i t e r a t i v e

method b a s e d on g e o m e t r i c and a r i t h m e t i c m e a n s was employed. F o r in-

complete elliptical i n t e g r a l s , the iterative method b a s e d on i n v e r s e o r d e r

of t r a n s f o r m a t i o n was employed in o r d e r to c o n s t r u c t N E F F ( 3 ) . The p r o -

g r a m s a r e written in l 1 W I Z l 1language f o r GE 2 2 5 c o m p u t e r s , which i s

analogous to "FORTRAN" f o r IBM c o m p u t e r s , and a r e given on the following

pages:
 51

W I 2 SOURCE PROGRAM

SEQ -LABL- - TY- STATEME NT C- ZE -NZE - PL - MI -ANY

-ARGPBKEF(2)
-VA#l ,VB#ARGP,PI #3.1415926536

-Qf?OD#ARG*ARG/( (ldrARGP)*(l&ARGP)) ,KN#QROD.NO&- -

-KKI#0.5*dROD ,SUN#-O.5*( IMK)

-KP#VB/VA,SUM#SUN,PN# 2*SQRT. (VB*VA)/(VA&VB) --

-KN#KN*KN/( ( lM")*( 1&PN) ) ,ROD2#KN ,ROD1 &ROD

-wODRaoo1 dCRoD2dcR001*Roo2
-KK#KK*0.5*KN ,SUN#SWO,5%K

-VATI#0.5* ( V A W B ) ,VBT#SQRT. (VA*VB)

-V A ~ A ,TV S # ~ B T,NO#NO&~
-ABS. ( (wm-~ca)/QROD)-DELTA
-ABS. ( (SUN-SUM)/SUN)-DELTA
-KNPHkY)ROD*PI *O. 5 ,FKKNPHBQ. 5*P I

-DKEF( 1 ) # f ~ WK)

-DKEF( 3) #SUN*FK $( (E(K)-K(K))/K/K)- -

-- -DKEF 2)#FK&ARG*ARG*SUN*FK $E(K)

- - -DKEF 4 ) #KNPH $K ( K )-PI /2

- - -OKEF 5)#NO
 52

W I Z SOURCE PROGRAM

STATEMENT -C- ZE -NZE - PL - M I -ANY

 53

W I 2 SOURCE PROGRAM

STATEMENT

587- -Kt<MK*K:V ,W&*W

590- -SSN#SSN&AC*SSY&AS*CSH, N02#N02&1 $$

59'- -OELN@ELN*DELN/( (1 WN)+ ( iMN@*RTKN) )

592- -KN&*RTKN/( I&N)

593- -VN#--MM*OELN/( KK*KN )
594- -WJ#-ir,MM*DSN/SQEiif. (KK)

595, -ENK3#ENKTMKK*VN-KAPA*UN $ E-K*SI N. (BETA) --

596- -ABS. ( (ENKS-ENKT)/ENKS)-DELTA - -++ - -*+ -
597- -N E F F ( ~ ) # I F $
. F(BETA,K.#KAPA) --
598- -NEFF ( 2)#ENKS&KAPA*SSO $ E(BETA,K)

599, -NEFF( 3 ) #fNKS ,NEFF (4) #?dol,NEfF ( 5 ) #NO2 -- - - - -

9 -

- - $ENKS I S E(BETA.K)-K*SIN(BETA),K~APA
 54

A P P E N D I X IV. DERIVATION O F TN

For n(+%2,9') the f o r m u l a s 410.01 and 411.01 of R e f e r e n c e 12

a r e a p p l i c a b l e , in which dP= -
-2 <o /lQ=9, , i.e. ,
-1- )Ip

[IV- 11

where

[IV- la]

and

[IV- 23

To exploit the p o s s i b i l i t y of gain in significant f i g u r e s , K- Kfd)

/- +=
will not be computed by s i m p l e s u b t r a c t i o n of
K(@
d*
from either

of the above equations. T h i s difference i s defined a s flN.

F r o m s i m p l e a l g e b r a i c manipulation of Equations (IV- 1) and ( I V - 2 ) ,

one f i n d s
 55

For pi, p~ small, p i s near 3 , k is n e a r z e r o , thus

n,(p,fi) i s n e a r but l e s s than unity. Also f o r p., o;, n e a r unity, k

i s n e a r unity. Equation (IV-3) may s t i l l l o s e too m a n y significant

f i g u r e s through A, - / o r / - 102 . One c a n f u r t h e r apply the addition

f o r m u l a (#153.01, Ref. 12) r e s t r i c t e d t o the condition t h a t R tanp-

t a n V = 1 , i. e . ,

where iV2- /-

Eliminating A,(fi A ) -1 f r o m (IV-4), (IV- l a ) , ( I V - 2 a ) , one finds

The ref o r e

[IV-6;

Applying the addition f o r m u l a (IV-4) a g a i n , one finds

 [IV- 83

where

F1v-9~~1

[lV-9b]
T h e r e i s apparently a gain of significant f i g u r e s of n, when f; ,

a r e s m a l l (d, R s m a l l ) if equation (IV-7)is u s e d , provided that the

f i r s t t e r m in the b r a c k e t c a n be evaluated a s a c c u r a t e l y a s t h e second

term. This i s achieved by employing the subroutines D K E F and N E F F

f o r equation (IV-9).
 APPENDIX V . X-FORCE ACTING O N THE TANK B Y
INTEGRATION O F PRESSURE

A s s u m e a velocity potential

[v- 1)
w h e r e the f i r s t t e r m i s a p a r t i c u l a r solution s a t i s f y i n g the n o r m a l

d e r i v a t i v e condition on t h e s p h e r e . & a r e the e i g e n functions which

have iio contribution to t h e n o r m a i veiocity on t h e s p h e r e . In o r d e r

t o s a t i s f y the f r e e s u r f a c e condition f o r s i n u s o i d a l o s c i l l a t i o n s

one h a s

since 6; a r e orthogonal on F , &=L 4'd.S and &=?ix 2

8,dS

The p r e s s u r e

The x - f o r c e c a n be obtained b y d i r e c t i n t e g r a t i o n of p r e s s u r e
 59

since

Equations [V-2, - 4 1 a r e the s a m e r e s u l t s a s t h a t obtained in

R e f e r e n c e 1 through L a g r a n g i a n s ’ equations,
 z

, x

FIGURE 1. G r a p h i c a l Illustration of S o m e N o m e n c l a t u r e s
 61

Q
i
P

F I G U R E 2. Surface of Integration
 62

c
i

L
FIGURE 3a. M o m e n t About Axis of Rotation
 63

II X - z

center 0

L rocking a x i s parallel t o IJ

J ?

F I G U R E 3b. Moment of Tank in Pitching (Rocking) Oscillation

 64
0

a
6
a
W
L
I
I!
0
z 5
0
0
W
v)
2
l
(z
W
t-
W 4
E
a
a
a
(L
t
0
z
W
3 3
Ea
LL
-I
a
a
3
t- 2(
a
z

-0.5 0 0.5 I .o

LIQUID DEPTH %
FIGURE 4a. C o m p a r i s o n of First N a t u r a l Frequency with
Data by Abramson, e t a l . , (Ref. 15)
 65

4.413-
4.0
I
0
A

Q
WATER
MERCURY REF IO
THEORY
THEORY, BUOIANSKY *

-1.2 -0.8 - 0.4 0 0.4 0.8 1.2

LIQUID DEPTH %

FIGURE 4b. C o m p a r i s o n of the First T h r e e N a t u r a l F r e q u e n c i e s

with E x p e r i m e n t s of Stofan- A r m s t e a d (Ref. 10)
 66

ZF/d = -//2 , U/d = 0.00828

specific
gravity
A Water Y = / p = LO centipoise

v Cloride Z = 1.34 p = Q44~entipoise

+Present rhcory

0 2 4 6 8
w: d
-
NATURAL FREQUENCY PARAMETER BASED ON DIAMETER
9

FIGURE 5. C o m p a r i s o n of F o r c e Response f o r Q u a r t e r - F u l l Tank

with E x p e r i m e n t s by A b r a m s o n , et a l . , (Ref. 15)
 67
16

14

12

10

0
-0.8 -0.4 0 0.4 0.8
LIQUID DEPTH 7
2,

FIGURE 6 . N a t u r a l F r e q u e n c i e s Given by R i l e y - T r e m b a t h (Ref. 17)