LIBRARY

TECHNICAL REPOPTT SCOTtO*

NAVAL POSTGRADUATE SCHQOff
MONTERE*, CAUVOUtlA

fiQ 77* * V/

Prepared for
Office of Naval Research
Contract No. N00014-69-A-0200-4048

/O ^(T A
li *" t/ UCyVENG^414
/ FEBRUARY 1974

OPTIMUM DESIGN OF RING STIFFENED

CYLINDRICAL SHELLS , A.J. BRONOWICKI
R.B.NELSON
L.P. FELTON
L.A. SCHMIT. JR.

Reproduction in whole or in part is permitted

for any purpose of the United States Government.

UCLA SCHOOL OF ENGINEERING AND APPLIED SCIENCE

 TTMri.ASS-nrTF.T1
SECURITY CLASSIFICATION OF THIS PAGE (When Data Entered)

READ INSTRUCTIONS
REPORT DOCUMENTATION PAGE BEFORE COMPLETING FORM
1. REPORT NUMBER 2. GOVT ACCESSION NO 3. RECIPIENT'S CATALOG NUMBER

4. TITLE (and Subtitle) 5. TYPE OF REPORT & PERIOD COVERED

OPTIMUM DESIGN OF RING STIFFENED CYLINDRICAL

SHELLS 6. PERFORMING ORG. REPORT NUMBER
UCLA-ENG-7414
7. AUTHORfsj 8 CONTRACT OR GRANT NUMBERfsJ
Allen J. Bronowicki Lucien A. Schmit, Jr. Contract No.
Richard B. Nelson
Lewis P. Felton N00014-69-A-0200-4048
9. PERFORMING ORGANIZATION NAME AND ADDRESS T'lO. PROGRAM ELEMENT. PROJECT. TASK
! AREA A WORK UNIT NUMBERS
Mechanics and Structures Department j
School of Engineering and Applied Science j
University of California, Los Angeles, Cal. 90024
II. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Procuring Contracting Officer February 1974

Office of Naval Research 13. NUMBTR OF PAGES

Department of the Navy. Arlington, Va. 22217 _22

U. MONITORING AGENCY NAME ft ADDRESS-'/ dlttarent from Controlling Otllc.e) 15. SECURITY CLASS, (of this report)
Director, Office of Naval Research Branch Office-
Pasadena
1030 East Green Street 15 DECl. ASSIFiCATION DOWNGRADING
SOKC DI.LE
Pasadena, Calif. 91101
16. DISTRIBUTION STATEMENT (of this Report;

17. DISTRIBUTION STATEMENT (o> tho abstract entered In Ulcck 20. It different from Report)

18. SUPPLEMENTARY NOTES

19. KEY WORDS (Continue on reverse aide if necessary and Identity by block number)

Structural design, optimal; Shells, cylindrical; Vibration; Buckling

20. ABSTRACT (Continue on reverae aide it neceaaary and Identity by block number)

This report deals with the optimum structural design of circular cylindrical
shells reinforced with identical uniformly spaced T-ring stiffeners, and sub-
jected to external pressure loading. The optimization problems considered are
of three types: (1) minimum-weight designs, (2) design for maximum separation
of the lowest two natural frequencies, and (3) design for maximum separation
of the lowest two natural frequencies which have primarily axial content.
Gross buckling is precluded by specifying a minimum natural frequency, and

DD 1 jAN 73 1473 EDITION OF I NOV 65 IS OBSOLETE

S/N 0102-014-6601
SECURITY CLASSIFICATION OF THIS PAGE (When Dmtm Bntmrud)
-LUURITY CLASSIFICATION OF THIS PAGECHTjen Data Entered)

20. additional behavioral constraints preclude yielding or buckling of

panels, T-ring stiffeners, and web and flange instabilities within each
T-ring. Analysis is based on use of an equivalent orthotropic shell
model, and optimization is accomplished through use of a sequential
unconstrained minimization technique. Examples indicate that a small
increase in weight above optimum (minimum) values can result in rela-
tively large increases in frequency separation, and that maximum fre-
quency separation is obtained when second and third lowest frequencies
approach each other.

SECURITY CLASSIFICATION OF THIS PA0E(TW>n Data Entered)

 UCLA-ENG-7414
FEBRUARY 1974

OPTIMUM DESIGN OF RING STIFFENED CYLINDRICAL SHELLS

Allen J. Bronowicki
Richard B. Nelson
Lewis P. Felton
Lucien A. Schmit, Jr.

Prepared for
Office of Naval Research
Contract No. N00014-69-A-0200-4048

Mechanics and Structures Department

School of Engineering and Applied Science
University of California
Los Angeles, California 90024
 ACKNOWLEDGMENT

The research described in this report was performed under Office of

Naval Research Contract No. N00014-69-A-0200-4048.

ii
 ABSTRACT

This report deals with the optimum structural design of circular cylin-

drical shells reinforced with identical uniformly spaced T-ring stiffeners,

and subjected to external pressure loading. The optimization problems con-

sidered are of three types: (1) minimum-weight design, (2) design for maximum

separation of the lowest two natural frequencies, and (3) design for maximum

separation of the lowest two natural frequencies which have primarily axial

content. Gross buckling is precluded by specifying a minimum natural fre-

quency, and additional behavioral constraints preclude yielding or buckling of

panels, T-ring stiffeners, and web and flange instabilities within each T-ring.

Analysis is based on use of an equivalent orthotropic shell model, and optimiza-

tion is accomplished through use of a sequential unconstrained minimization

technique. Examples indicate that a small increase in weight above optimum

(minimum) values can result in relatively large increases in frequency

separation, and that maximum frequency separation is obtained when second

and third lowest frequencies approach each other.

iii
 SYMBOLS

t Thickness of shell skin (in)

s
t. Thickness of frame web (in)

t.c Thickness of frame flange (in)

d. Depth of frame web (in)

cL Width of frame flange (in)

I Spacing of frames (in)

W Normalized mass of structure

3
p Mass density of structural material (slug/in )

L Total length of structure (in)

R Radius to mid-surface of skin (in)

P Hydrostatic pressure (psi)

T Kinetic energy

x,<j),z Shell coordinate system

u,v,w Midsurface displacements

u,v,w Displacements of an arbitrary point

{A}J Eigenvector of ith natural frequency

0) Natural frequency

n,m Wave numbers

T Time (sec)

[K] Stiffness matrix

[KQ] Geometric stiffness matrix

[M]. Mass matrix

a Yield stress (psi)

a* Critical flange buckling stress (psi)

cr
aw Critical web buckling stress (psi)
cr
F Objective function

iv
X Vector of design variables

g(x) Design constraint

q Number of design variables

p Critical pressure for buckling of skin (psi) between frames

cr
Minimum allowable natural frequency of vibration in vacuo (Hz)
min
E Young's modulus

V Poisson's ratio

& Unsupported length of shell plating (in)

Wmax Maximum allowable normalized structural mass

$(x) Composite objective function

e Constant for use in computing extended penalty function

r Positive scalar quantity

s Direction vector for uni-directional search

 INTRODUCTION

Although a wealth of literature exists for the static, dynamic and sta-

bility analyses of stiffened shells of revolution subjected to various applied

loads, with the majority of these studies devoted to cylindrical shells, the

work of Schmit and Morrow [1] serves as a pioneering effort toward the intro-

duction of structural optimization concepts into the design of stiffened cylin-

drical shells. In this reference a cylindrical shell, reinforced with longi-

tudinal and ring stiffeners, each with rectangular cross section, was designed

to carry a number of independently applied sets of static loads with minimum

structural weight. The shell was constrained against overall (system) buckling,

panel and stiffener buckling, and also against material yield. The mathematical

model which formed the basis for the stress and buckling analyses was an equiv-

alent homogeneous orthotropic shell; i.e., the discrete stiffeners and skin

stiffness properties were incorporated in the orthotropic elastic shell stiff-

ness properties. This theory, a 3rd order Flugge-Lur'e-Byrne type theory,

proved adequate provided the stiffener spacing and cross sectional dimensions

were sufficiently small to permit the smoothing operation inherent in the

orthotropic shell model.

In a more recent study, Pappas and Allentuch [2,3] investigated the mini-

mum-weight design of ring stiffened cylindrical shells, subjected to a number

of static applied load conditions. In this study, the ring stiffeners were

T-shaped rather than rectangular. The structural analyses of general instabil-

ity, localized panel instability, and stiffener instabilities were based on

buckling formulas contained in Ref. [4], along with the appropriate stress

limits.

The subject of this report is, in essence, a structural optimization

study which employs both a combination and an extension of the structural

models in Refs. [1-3]. In the present study, three somewhat different structural

optimization problems have been formulated. In the first, the minimum-weight

design of a ring stiffened cylindrical shell subjected to several applied static

loads is considered. The structure is constrained against all of the

buckling modes considered in Refs. [2,3] although for some of the modes, the

constraint equations differ. Further, the structure is constrained against

in vacuo natural vibrations below a specified frequency limit. In the second

formulation the shell is designed to maximize the separation between the lowest

two in vacuo natural frequencies, while being constrained against buckling and

yield behavior as specified in the first design problem, and while having a

weight less than a prescribed maximum. The third formulation is similar to the

second, with the distinction that the frequencies being separated are the two

lowest which have primarily axial content.

This optimization study, in either the weight minimization form or the

frequency separation forms, represents a considerable advance beyond previous

characterizations of the optimization problem. The structural model in this

study is based on the equivalent orthotropic shell model of Ret. [1], with

dynamic etfects added. This representation, although somewhat imprecise in its

ability to model a ring stiffened shell, has proved invaluable in providing an

economical and yet reasonably complete initial structural representation for

use in the optimization studies performed. A more detailed description of the

mathematical model, and a summary of the results of this research, are presented

in the following sections.

 ANALYSIS OF STIFFENED SHELLS

The investigation detailed in this report is concerned with the optimum

structural design of circular cylindrical shells with uniformly spaced T-ring

stiffeners (Fig. 1) and subjected to several different applied loading conditions

typical of submerged vessels, namely (1) specified external pressure (or vessel

depth), (2) specified axial compressive loadings, and (3) applied static loads

associated with vessel motion. Design variables, shown in Fig. 1, are skin

thickness (t ), stiffener web thickness (t,), web depth (d,), spacing ( ), flange
s <p cp X
width (df), and flange thickness (t ). Radius (R), length (L), and the material

properties are assumed to be preassigned parameters.

All of the research performed to date has employed the simplified ortho-

tropic shell model given in [1] in the calculation of the natural frequencies

of vibration and in the overall (system) buckling analysis. This idealization

has of.fered the advantage of mathematical simplicity and the disadvantage of a

somewhat limited modeling capability , but it has provided an economical initial

structural representation for use in the optimization studies.

In order to analyze the dynamic response of the cylinder in Fig. 1, it is

hypothesized (as in Ref. 1) that the frames and skin act as a unit accurding

to the Bernoulli-Euler deformation assumption as extended through the Flugge-

Lur'e-Byrne theory, and that the stiffness and inertia properties of the frames

are uniformly distributed over the length of the cylinder. It is then possible

to express the kinetic energy of the shell in the form

L 2TT

/ ffi
2L / / /(u2+52+42)dzJ4>dx
I = ^ / (1)
o o t

The model gives a very accurate representation of the structural behavior pro-
vided the characteristic wavelengths of Lhe modes of vibration (or of the static
displacements) are very long compared to both the ring spacing and ring cross-
sectional dimensions.
 WEB HULL PLATING

y
IV2

r i V

FLANGE r-^-l (b) FRAME AND SKIN DETAIL

'f

^i
JrTrx~2r~2r~2rTr~2T~3ririr'i

lLJT_5?_5L5L_K_JL_5LJ^__5L_5L_5r
(a) HULL SEGMENT CROSS-SECTION

Figure 1. Typical Hull Segment Cross Section (from Ref. 2).

where u, v and w are the displacements of an arbitrary point in the structure in

the x, <J>, and z directions, respectively (Fig. 2), and are given by L1J
9w ,, .
u = U-ZTS (2a)
dx
,. z. z 3w ,, .
v = (l--)v- - w (2b)

w = w (2c)

where u, v and w are the displacements of the shell's mid-surface. From the

kinetic energy the inertia terms in the appropriate equations of motion are

obtained by use of Hamilton's principle. These terms are appended to the

equations of static equilibrium for the stiffened shell in [1], which also con-

tain the influence of destabilizing forces.

Assuming that the external loads give rise to circumferential and longi-

tudinal compressive forces per unit length of magnitude PR and PR/2, respec-

tively, where P is hydrostatic pressure, then the combination of inertia, staticj

and destabilizing forces leads to the following three coupled partial differen-

tial equations of motion.

1 P VR
N* + fjN* - -(u**-Rw') - pi" = pr (3a)
xRcpxR 2 x

1 IIP PR
+ N
?*$ 'A " V*\ ~ -o^Z - f(v** + w*) - ^v" = p
R<px<pRx<|>2<f)R 2 *y (3b)
v/
is.
2
M" + hl*r + -k'* + ^rM** + h. + |"(v*-w**-Ru'- ~w") = p (3c)
v/
x R <px R x(p 2 <p R <p R 2 *z
R

where

Px = P / [ - zw'](l - |)dz (4a)

t

p = p [(1
y s/ - i)2* -(1 - i>tf>**i<i - f>dz <4b>

Pz PSJ tw + zu' + (1 - |)(f)v* - (z)V - (|) w**](l - |)dz (4c)

 N + aN
x0 < x0/9x) dx
N + 8l a0)d0
0x < V

Nx+(3Nx/8x) dx
N0+oiyd0)d0

M0+(3M0/80)d0 Mx +(3Mx/3x) dx

M
0x -K9M0x/30) d0 Mx0 -K3Mx0/3x) dx

Figure 2. Force Resultants and Coordinate System.

and where ( ) denotes -J^ , ( )' denotes -- , ( )* denotes -j^f- and J ( )dz

is the integral through the thickness of the shell and frame. The forces M and

N may be expressed in terms of the mid-surface displacements u, v and w (see

Appendix I) so that Eqs. (3) can be expressed in terms of displacements only.

Under the assumption that the boundary conditions are of the simple support

type, then the solution to these equations takes the form

u = A, sin nty cox r~~ sin OOT

v = A2cos n(j> sin r~ sin OUT (5)

w = A~sin n<l> sin r sin OJT

where n = 0,1,2,... and m = 1,2,....

Substitution of Eqs. (5) into Eqs. (3) gives the algebraic eigenvalue problem

([K] - P[KG] - 0)2[M]){A> = {0} (6)

where the stiffness, "geometric" stiffness, and mass matrices are [K], [K ] and

[M], respectively, and are given in Appendix I.

It should be noted that the sine and cosine dependencies on the angle (f>,

and the similar dependencies on the axial variable x, could have been inter-

changed without influencing the matrices in Eq. (6) for n > 0. The n = 0 case

as given in Eq. (6) is actually a combination of the solution form in Eqs. (5)

(pure torsion) and the similar form with sine and cosine terms (with argument

zero) interchanged (torsionless motion).

Since the algebraic eigenvalue problem for given values of m, n and P is of

only rank three, its solution for the natural frequencies (eigenvalues) and

associated modes (eigenvectors) is accomplished without difficulty.

In the portion of the structural optimization research wherein the structure

was designed for the greatest separation of the lowest two axial-type vibratory

modes, the modes with A. = 1, and A2, A3 < 1 were termed "axial." In order to

prevent any general buckling from occurring, all the frequencies associated with
values n = 0 6, m = 1,...,6 were retained and forced to exceed a prescribed

minimum CO , .
min
In addition to this gross buckling, it is necessary to be able to detect

the occurrence of several additional modes of "local" failure. These local

failure modes consist of panel (inter-ring) yielding or buckling and yielding

or buckling of the webs and/or flanges of the stiffener rings.

Panel (skin) yielding will be precluded, according to the Von Mises

criterion, provided

/ 2
(a. - a.o +, a 2.1/2
) ' <^ a ,-..
(7)
<f> <J>x x y

where O = material yield stress in uniaxial loading, and a, and a are in-
plane stresses normal to the surfaces of the element in Fig. 2. From Ref. [4]

the maximum bending stresses in the panels due to external pressure are
PR
o. = -~[l + T(Hn + VHE)] (8a)
s
PR
crx = (1/2 + THE) (8b)
s
where T, H, and H are load factors defined in Appendix II.

A suitable approximate formula [4] for the critical external pressure

which will cause panel buckling is

Pcr = 2.42E(ts/D)5/2/ {(1-V2)V4[(/D) - 0.45(ts/D)1/2]} (9)

where % = & - t, is the unsupported length of a panel, and D = 2R.

Again following Ref. [4], the maximum compressive stress in the rings may
be taken as

ac = -PRQ/A (10)

where A = t t, + t,d. + t.d-, and Q is a load factor defined in Appendix II.

The magnitude of O must be less than the yield stress, a , and also less than

the critical values of compressive stresses at which the flange or web will
buckle. Assuming that the web and flange are infinitely long rectangular

plates, that the web is simply supported along all edges, and that the flange is

simply supported along three sides and free on one edge (all conservative assump-

tions) , then the critical stresses for buckling of the flange and web, respec-

tively, are [5]

0.5067T Ero ,(A .,2

Cf = =-[2tfr /(dfX -t.)]
Z 9
(11a)
cr 12(l-v )
2
4Tr E
a = [t./d.]2
9 (lib)
W
cr 12(1-V ) *
2

It should be noted that Eq. (10) neglects any effects of eccentricity in the

circularity of the cylinder and of lateral-torsional stiffener buckling on the

compressive stress in the rings.

 FORMULATION OF OPTIMIZATION PROBLEM

As indicated in the Introduction, the optimization problem takes the follow-

ing three alternative forms:

I. Find the minimum-weight design of the stiffened cylindrical shell subjected

to the applied loads described previously and constrained against overall

(system) buckling, panel (inter-ring) buckling, T web and/or flange buck-

ling, panel and/or ring yield. The shell is also required to possess a

lowest natural frequency (in vacuo) greater than a specified minimum.

II. Find the structural design which maximizes the separation between the low-

est two natural frequencies of vibration (in vacuo) for the stiffened cylin-

drical shell subjected to the applied loads described previously and con-

strained against overall (system) buckling, panel (inter-ring) buckling,

T web and/or flange buckling, panel and/or ring yield. The shell is also

required to possess a lowest natural frequency greater than a specified

minimum, and a structural weight less than a prescribed maximum.

III. Find the structural design which maximizes the separation between the

lowest two natural frequencies of vibration (in vacuo) which have primar-

ily axial content for the stiffened cylindrical shell subjected to the

applied loads described previously and constrained against overall (sys-

tem) buckling, panel (inter-ring) buckling, T web and/or flange buckling,

panel and/or ring yield. The shell is also required to possess a lowest

natural frequency greater than a specified minimum and a structural weight

less than a prescribed maximum.

Each of these problems has the form of an inequality-constrained optimiza-

tion problem, which may be solved by any of a number of mathematical programming

algorithms. The particular method of solution chosen for this work is the

Sequential Unconstrained Minimization Technique (SUM!) in which the optimum

10
structural design problem (in either form I, II, or III) is converted into a

sequence of unconstrained problems [6] by means of so-called "penalty functions."

In this technique, it is required to find a vector x = {x-x_...x....x } =

it t.d. d^t,,}, the components of which are the design variables, such that a
s <p (p x f f *
specified function of these variables, F(x), is extremized while satisfying a

set of constraints of form gk(x) ^ 0> k = l,...,s.

The "objective function," F(x), takes one of three forms, depending on

the optimization problem being considered. For Problem I, weight minimization,

2-rrp
F(x) = W = ^ {RLt + (L/ )[(R -ejt.d. + (R -ef)t df]} (12)
7T(R+t /2)2Lp S X 9 9 9 r r
s w

where parameters e, and ef are defined in Appendix I. In Eq. (12) F(x) has

been normalized by dividing shell mass by the mass of displaced water.

For Problem II, the case of separation of lowest natural frequencies, an

inverse formulation is used. The separation is maximized by minimizing the

inverse of the separation. For the modes being considered, an ordered list

is made giving 0) < )- < u>_, etc. The objective function is then

F
- j^r^J ()
where A is a normalization factor taken as the initial frequency separation.

To separate frequencies having primarily a longitudinal deformation content

(Problem III) it is necessary to examine and order frequencies in each primarily

axial mode, i.e., the ones having both A and A smaller than A . The objec-

tive function is then given by

L L
2 l
where the subscript L has been added to denote the longitudinal character of

the shell vibration.

11
 The number of design variables, q, is a maximum of six in this study, but

may be less if certain of the design variables are fixed. Also upper and lower

limits U. and L., respectively, are assumed specified for each variable x..

These upper and lower limit constraints, respectively, are written in the normal-

ized form

g(x) = (U - x^/OJ. - L) * 0 (15)

1-1, ...,q
8
q+i(- = (X
i " Li)/(Ui " Li} ~ (16)

It is also necessary to include a geometric admissability constraint which

serves to keep the frame flanges from overlapping. This is expressed in the

normalized form
1
82q+lW - ' V*x * <17>
The behavioral constraints may also be normalized. The panel yield con-

straint is expressed as
1
2 2 1/2 /Q
g2q+2<5> - - <a<|> " GxacJ> + 0 )
x y * <18>

The frame yield, flange buckling, web buckling and skin buckling constraints,

respectively, are written as

g2q+3 - 1 - K>y * 0 (19)

S2q+4(x) = 1 - |0e|/OfM > 0 (20)

g2q+5W - 1 - kcl/%cr > 0 (21)

g2q+6W - 1 - WP > 0 (22)

Finally, to prevent gross buckling the lowest frequency, U)- , should be

greater than a specified minimum frequency, 0) . . This is stated in the form

82q+7<*> - ->! - %n)/i2 * 0 (23)

where A0 is the initial value of ), - U) M .

< 1 min

12
 For the minimum-weight design of the shell these are all the constraints

required, but in order to separate frequencies two other constraints must be

included. It is conceivable that large separations could be achieved, but

possibly only at the expense of large increases in weight. It is therefore

necessary to establish an upper bound to the mass of the structure, W

max
This constraint is expressed by

S2q+8Ci> " 1 - W/Wmax >- 0. (24)

The final constraint is not required in the definition of the frequency

separation problems, but is convenient for numerical solution when penalty

functions are utilized in the Sequential Unconstrained Minimization Technique.

The most efficient unconstrained minimization methods require the computation

of gradients of the objective function in order to operate. These gradients

should be smooth and continuous, but experience shows that, as ), and )~ are

separated, w and w approach a common value. Eventually OJ and )~ will

switch, with the result that the mode which previously represented co now

represents w and vice versa. This causes discontinuity in the gradient of

the objective function and subsequent difficulties with the numerical algorithm.

This difficulty can be overcome simply by requiring that the second and third

frequencies never be equal. This requirement is expressed as a "singularity

avoidance" constraint in the two cases of frequency separation by

S2q+9(^ = W
3 " W
2 >
(25
>

or

82q+9C*> = \3 " \2 > 0 (26)

In order to apply the methods of unconstrained nonlinear programming, the

basic inequality-constrained problem is recast in the form of an interior

13
penalty function [6], the solution of which requires finding that jx which

minimizes
s
$(x) = F(x) + r XX (x) (27a)
j-1 J
where
l/gj(x) if gj(x) > e
Pj(x) - 2 (27b)
l2e - gjtoj/e if gj(x) < E

P.(x) is the so-called extended penalty function [7] which allows the use of

infeasible designs in the search for an optimal feasible design. The quanti-

ties E and r are small positive scalars and s is the number of constraints.

Solving this unconstrained minimization problem for successively smaller values

of r gives a series of designs which converge to a local or global optimum of

F(x). This procedure is referred to as the Sequential Unconstrained Minimi-

zation Technique (SUMT). Each unconstrained minimization is performed by the

Davidon-Fletcher-Powell method [6]. In order to apply this algorithm it is

necessary to compute the gradients of the objective function and the con-

straints with respect to each design variable. The mathematical complexity of

this problem makes analytic calculation of partial derivatives impractical

and it is found convenient to use a forward difference technique to find the

gradients numerically. The success of the unconstrained minimization hinges

on the ability to perform an accurate unidirectional minimization. A special

hyperbolic interpolation formula was developed [6] for use with the SUMT

method. A test for the minimum was developed and incorporated into the minimi-

zation algorithm which requires that a measure of normality between the direc-

tion vector s_, and the gradient, V$, be less than 0.001, i.e.,

(s V$)/(|sJ |7$|) < 0.001 . (28)

14
 EXAMPLES

Design examples given in [3] were re-evaluated in this study for optimal

performance for situations I, II or III as detailed above. In all of these

designs the preassigned parameters were given the following values:

R = 198 in., L = 594 in., E = 30 x 106 lb/in.2, O = 60 x 103 lb/in.2,

-4 3
p = 7.75 x 10 slug/in. , V = 0.33 and specified pressure associated with

depths of water of 1000, 2000 and 3000 ft. The prescribed minimum natural

(in vacuo) frequency was taken as 12.0 Hz. Except for the dynamic effects

these problems are quite similar to those in [3], The structure was designed

initially for minimum weight (problem I) for the three different operating

depths. The results, given in Table 1, indicate normalized minimum weights

of 0.13317, 0.22295, and 0.31922 for operating depths of 1000, 2000 and

3000 ft., respectively, values somewhat lower than reported in [3], This

occurrence is due to the fact that in this study the ring spacing was

represented by a continuous variable and web and flange thicknesses were

included as independent design variables, while in [3] the frame spacing was

a discrete parameter and the web and flange thicknesses were linked and re-

quired to be not less than 1/18 of the web and flange depths. In the designs

presented herein the frame webs are very thin and are critically stressed,

i.e., on the verge of buckling. Consideration of lateral-torsionl buckling

and the effects of hull eccentricity may alter this condition, although these

effects were excluded in the present work.

One important benefit in obtaining the minimum-weight designs in Table

1, is that they serve as initial, feasible designs for design problems II and

III, provided the same minimum frequency constraint is employed and the maximum

allowable structural mass is greater than that of the minimum-weight (mass)

design for the static load conditions.

15
 Table 1. Design Problem I - Weight Minimization

(R = 198 in., L = 594 in., E = 30 x 106 psi, 0 = 60,000 psi)

Depth 1000 ft. 2000 ft. 3000 ft.

t 7, in. 1.2108 2.4856 3.5156

s
C
A in
0.3765 0.4207 0.4543

19.589 19.284 19.915

<P
, in. 33.602 51.528 36.195

df, in. 17.664 14.999 16.991

t , in. 0.4705 0.4984 0.5453

00- , Hz. 28.12 26.05 26.64

CJ2, Hz. 49.39 30.04 35.88

a), Hz. 52.31 51.98 55.11

Weighta 0.13317 0.22295 0.31922

(Normalized)

Maximum (upper bound) weights (normalized) are 0.15, 0.25

and 0.35, respectively.

16
 The results of design problem II, optimization for maximum frequency

separation are given in Table 2 for the cases of the same three preassigned

operating depths. The maximum allowable normalized weight was taken as

approximately 10% greater than that for the minimum-weight design for the

static load condition, the values being 0.15, 0.25 and 0.35, respectively,

for the operating depths of 1000, 2000 and 3000 ft. For the three depth

requirements the frequency separation was increased from the minimum-weight

design values of 21.27, 3.993 and 9.2418 Hz to 23.59, 24.28 and 25.33 Hz,

Of interest is the fact that the first and third solutions (for 1000 ft.

and 3000 ft. operating depths) gave frequencies io2 = 0)_, i.e., nearly identi-

cal second and third frequencies, and that in these two cases the maximum

weight constraint was less than critical. It is thus apparent that the

major portion of the frequency separation is obtained by bringing the second

frequency up to the third frequency, and that little or nothing is to be gained

by adding more material to the structure after this is accomplished. In the

example with 2000 ft. depth the second and third frequencies were unable to

completely approach each other before violating the maximum weight constraint,

which became critical in this case. It may be noted that for these designs

the frame webs are very thin and the frame flanges are relatively thick. It

thus seems that the frequency separation has been achieved by making the

moment of inertia of the frames as large as possible.

Three problems of category III were designed to find the maximum fre-

quency separation in primarily longitudinal modes of vibration for the three

different operating depths. Results are given in Table 3. For the cases

which have operating depths of 1000 ft, and 2000 ft. the algorithm became

entrapped in a singularity in the design space. The problem run at 3000 ft.

depth was unable to reach the singularity because the maximum weight

17
 Table 2. Design Problem II - Frequency Separation

(R = 198 in., L = 594 in., E = 30 x 10 psi, C = 60,000 psi)

Depth 1000 ft. 2000 ft. 3000 ft.

t , in. 1.2216 2.4717 3.3986

s
in
V - 0.3950 0.3884 0.5168

in 20.722 19.674 23.764

V -
A 7, in. 33.853 51.417 36.065
X

d , in. 17.551 15.075 16.864

tf, in. 0.4653 1.8638 0.9485

Cx)1 , Hz. 28.3711 29.4905 29.7225

w2, Hz. 51.9638 53.7674 55.0508

W-, Hz. 51.9640 54.1466 55.0512

Weight3 0.1351 0.25 0.32928

(Normalized)
(u^-tj^) Hz. 23.5928 24.2769 25.3283

Maximum (upper bound) weights (normalized) are 0.15, 0.25

and 0.35, respectively.

18
Table 3. Design Problem III - Longitudinal Frequency Separation

(R = 198 in., L = 594 in., E = 30 x 106 psi, a = 60,000 psi)

Depth 1000 ft. 2000 ft. 3000 ft.

t , in. 1.5975 2.6720 3.9377

s*
fc
A in. 0.2500 0.4171 0.42116
<P
<L in. 11.603 17.384 19.817

I , in. 38.672 52.232 36.143

X*
d , in. 9.1454 13.232 16.907

t , in. 0.3826 0.5248 0.4927

(x)., Hz. 163.0006 163.0640 138.2134

w2, Hz. 192.2375 192.7389 162.4559

co3, Hz. 222.6656 222.7126 221.6170

Weight3 0.14215 0.23559 0.34907

(Normalized)

(OL -00 ) Hz. 29.2369 29.6749 24.2425

L L
2 l

nyiaximum (upper bound) weights (normalized) are 0.15, 0.25,

and 0.35, respectively.

19
constraint had become active. A singularity avoidance constraint could have

been developed to enable the algorithm to proceed, however this was not done

because in order for the optimization problem to have a significant physical

purpose, a better definition of what actually constitutes a longitudinal mode

is needed. As may be seen in Table 3 for the 3000 ft. case, the mode having

one longitudinal wave and no circumference waves (m = 1, n = 0) has a "longi-

tudinal" frequency of 138.21 Hz, because the v-component of the associated

eigenvector is zero and the w-component is less than 1.0 (0.977). As the

algorithm separates the frequencies, the eigenvector associated with this

frequency changes form with the w-component increasing to a value slightly

greater than 1.0. As this occurs, a second eigenvector of this mode's

vibratory class (m = 1, n = 0) becomes longitudinal in form, according to

the definition currently in use. The result is the situation encountered

for the 1000 ft. and 2000 ft. cases, where the m = 1, n = 0 mode has the

second lowest frequency. The algorithm becomes entrapped at a frequency

separation of 29 Hz as two eigenvectors for the m = 1, n = 0 mode switch

back and forth, having w-components both approximately equal to 1.0. Since

the m = 1, n = 0 mode has two frequencies of vibration with almost the same

mode shape it is not realistic to call either one the unique longitudinal

frequency for that mode.

This effect was only recently encountered and must be given additional

study. When successfully resolved, the result will be a capability for

generating optimum solutions to design problems I, II and III. For the

future, research should be directed toward incorporating more sophisticated

structural models, with more design flexibility, in anticipation of obtaining

even more efficient structural configurations from both the standpoint of

minimum-weight and maximum frequency separation.

20
 REFERENCES

1. Morrow, W.M., and Schmit, L.A,, "Structural Synthesis of a Stiffened

Cylinder," NASA CR-1217, December 1968.

2. Pappas, M., and Allentuch, A., "Structural Synthesis of Frame Reinforced

Submersible, Circular, Cylindrical Hulls," NCE Report No. NV5, May 1972.

3. Pappas, M., and Allentuch, A., "Optimal Design of Submersible Frame

Stiffened, Circular Cylindrical Hulls," NCE Report No. NV6 (Revised),
July 1972.

4. Principles of Naval Architecture, Society of Naval Architects and Marine

Engineers, New York, 1967.

5. Timoshenko, S.P., and Gere, J.M., Theory of Elastic Stability, 2nd Ed.,
McGraw-Hill, 1961.

6. Fox, R.L., Optimization Methods for Engineering Design, Addison-Wesley,

1971.

7. Kavlie, D., and Moe, J., "Automated Design of Frame Structures," Jour, of
the Structural Division, ASCE, Vol. 97, No. ST1, Jan. 1971, pp. 33-62.

21
 APPENDIX I: DYNAMIC ANALYSIS

The forces on the cylindrical shell are as shown in Fig. 2. The forces

are expressed in terms of displacements as:

N = Hu' + (Hv/R)v* - (HV/R)w + (D/R)w"

N = HVu' + [(H+HC+HF)/R]v* - [H+HC(l+e,/R) + HF(l+ef/R)]w/R

- [D/R+HC(e +p2/R) + HF(ef+p2/R) ]^j w**

R
N , = (S/R)u* + Svf + (K/R2)w!*

N = (S/R)u* + Sv - (K/R2)w'*

M = - (D/R)uf - (Dv/R2)v* - Dw" - (Dv/R2)w**

M = (HCe./R + HFef/R)v* - [D/R + HC(e +p2/R) + HF(ef+p2/R)]w/R

- Dvw" - [D + 2
HC(P -KAR) + HF(p2+a2/R)w**/R

Mx(J) = - (2K/R)(v,+w1*)

M. = (K/R2)u* - (K/R)v' - (2K+Q)wf*/R

K K }
' 3(|> ' 3x

The matrices [K], [K ] and [M] of equation (6) are:

Ml 12 M3
[K] = M2 v
22 '23

M3 '23 '33

k H 2
ll = " ^ - (S/R2)n2

k12 =
" (Hv+s)An/R

[HVA+DX3 - (K/R2)An2]/R
M3
k = .2^2 2 ,2
22 " (S+2K/R M - [H+HC(l-e./R) + HF(l-ef/R)]n /R

22
k23 = - (3K+Dv)X2n/R2

+ [D/R2-H+HC(p2/R2-l) + HF(p2/R2-l)]n/R2

- [HC(a^/R2-e(j)) + HF(a3/R2-ef)]n3/R3

k33 = - [H+HCU+e^/R) + HF(l+ef/R)]/Ri

+ 2[D/R+HC(e(|+p2/R) + HF(ef+p2/R)]n2/R3

- DA4 - [2Dv+4K+Q](An/R)2

- [I>fHC(p2+a3/R) + HF(p2-ta3/R)](n/R)4

_kGll k
G13
[KG] = 0 k
G22 k
G23
k k k
_ G13 G23 G33 _

k = - (n2+X2/2)/R k = n/R
Gll G23

kG13 = -> k = - (n2+A2/2)/R

G33

k =
G22 " (n2+x2/2)/R

m m
11 13
M = m m
22 23
m m m
13 23 33

m. = t + A(l-e,/R) + B(l-e./R)
11 s <p f

m [t3/(12R) + AC-e^/R) + B(-ef+p2/R)]A

13

m t (1 + -- ) + A(l-3e, /R+3pK2/R2-a3/R3)
22 s 4R2 <$> <(> <f>

+ B(l-3ef/R+3p2/R2-a3/R3)

23
m = [t /(6R2) +
23 s A(-e(()/R+2p2/R2-aJ/R3)

+ B(-ef/R+2p2/R2-a3/R3)]n

m = t +
33 s (x2+n2/R2)tg/12 + A[l-e,/R + (A2+n2/R2)(p?-aJ/R)]

+ B[l-ef/R+(A2+n2/R2)(p2-a3/R)]

The section properties are defined as follows:

H = Et /(1-V2) D = Et3/[12(1-V2)]
s s

A = t,d./,
(J) $ x B = t.dji
f f x

HC = EA HF - EB
G = E/[2(1+V)] t = t + 2d
*

S = Gt K = Gt3/12
s s

QH = G(J,+J.)
<p f :

J 3 C = 285e -fo.49 ^) +0 316

4, - yy* CO - - V V -
3 -[0.49 )
J
f " Cfdftf cf = - 0.285 e \ t
f/ + 0.316

e
% - TVV f - l<df-rtt)
P . = -zd , + -rd. t + Tt
q> 3 (}) 2 <J) s 4 s

p = + +
f ^f 2*f \ 4fct

3 = 1,3 ld232 13
<p 4 (p 2s(p 8 s (p 8s
3
1,3 _,_ 1 ,2 _,_ 3 2, _,_ 13
a =
f ^f + 2fctdf + Vf + 8*8

24
 APPENDIX II: STATIC STRENGTH ANALYSIS

The critical compressive stresses in the skin are assumed to occur on the

outer surface at mid panel. The stresses there are

0(, = - (PR/ts)[l+r(Hn+vHE)]

x= - (PR/
V [I + E]
where the various parameters are given as:

-PR/t = hoop stress

R = radius to mid plane of shell

r = [l-v/2-B]/(l+)

B = ratio of shell area under frame web to total frame area = t t,/A
s <p
A = t t, + t,d, + trd.
s <p 99 f f

3 = 2N{l/[3(l-V2)]}1/4(Rt3)1/2/A
s
N = (cosh9-cos9)/(sinhe+sin0)

9 = S,[3(l-v2)/(Rts)2]1/4

% = unsupported length of shell = - t,

H = - 2[sinh(e/2)cos(e/2)+cosh(6/2)sin(e/2)]/(sinh9+sine)

H^ = - 2[3/(l-V2)]1/2[sinh(e/2)cos(9/2)-cosh(e/2)sin(e/2)]/(sinhe+sin6)

The hoop compressive stress in the frame is

O = PVR/A

where V is a load factor given as:

V = t<J)[l+(l-V/2)3/B]/(l+B)

25
Distribution List

Chief of Naval Research Watervliet Arsenal

Department of the Navy MAGGS Research Center
Arlington, Virginia 22217 Watervliet, New York 12189
Attn: Code 439 (2) Attn: Director of Research
471

Director Redstone Scientific Info. Center

ONR Branch Office Chief, Document Section
495 Summer Street U.S. Army Missile Command
Boston, Massachusetts 02210 Redstone Arsenal, Alabama 35809

Director Army R&D Center

ONR Branch Office Fort Belvoir, Virginia 22060
219 S. Dearborn Street
Chicago, Illinois 60604 Technical Library
Aberdeen Proving Ground
Director Aberdeen, Maryland 21005
Naval Research Laboratory
Attn: Library, Code 2029 (ONRL) Navy
Washington, D.C. 20390 (6)
Commanding Officer and Director
Commanding Officer Naval Ship Research & Development
ONR Branch Office Center
207 West 24th Street Washington, D.C. 20007
New York, New York 10011 Attn: Code 042 (Tech. Lib. Br.)
700 (Struc. Mech. Lab.)
Director 720
ONR Branch Office 725
1030 E. Green Street 727
Pasadena, California 91101 012.2 (dr. W.J. Sette)

U. S. Naval Research Laboratory Naval Weapons Laboratory

Attn: Technical Information Div. Dahlgren, Virginia 22448
Washington, D.C. 20390 (6)
Naval Research Laboratory
Defense Documentation Center Washington, D.C. 20390
Cameron Station Attn: Code 8400
Alexandria, Virginia 22314 (20) 8410
8430
Army 8440
6300
Commanding Officer 6305
U.S. Army Research Office Durham 6380
Attn: Mr. J.J. Murray
CRD-AA-IP Undersea Explosion Research Div.
Box CM, Duke Station Naval Ship R&D Center
Durham, North Carolina 27706 Norfolk Naval Shipyard
Portsmouth, Virginia 23709
Commanding Officer Attn: Dr. Schauer
AMXMR-ATL Code 780
Attn: Mr. J. Bluhm
U.S. Army Materials Res. Agcy.
Watertown, Massachusetts 02172
26
Naval Ship R&D Center Director, Aero Mechanics
Annapolis Division Naval Air Development Center
Annapolis, Maryland 21402 Johnsville
Attn: Code A800, Mr. W. L. Williams Warminster, Pennsylvania 18974

Technical Library Naval Air Systems Command

Naval Underwater Weapons Center Department of the Navy
Pasadena Annex Washington, D.C. 20360
3202 East Foothill Blvd. Attn: NAIR 320 Aero & Structures
Pasadena, California 91107 5320 Structures
604 Tech. Library
U.S. Naval Weapons Center 52031F Materials
China Lake, California 93557
Attn: Code 4520 Mr. Ken Bischel Naval Facilities Engineering Command
Department of the Navy
U.S. Naval Ordnance Laboratory Washington, D.C. 20360
Mechanics Division Attn: NFAC 03 Res. & Development
RFD 1, White Oak 04 Engineering & Design
Silver Spring, Maryland 20910 14114 Tech. Library

U.S. Naval Ordnance Laboratory Naval Ship Systems Command

Attn: Mr. H. A. Perry, Jr. Department of the Navy
Non-Metallic Materials Division Washington, D.C. 20360
Silver Spring, Maryland 20910 Attn: NSHIP 031 Ch. Scientists for
R&D
Technical Director 0342 Ship Mats. & Structs.
U.S. Naval Undersea R&D Center 2052 Tech. Library
San Diego, California 92132
Naval Ship Engineering Center
Supervisor of Shipbuilding Prince George Plaza
U.S. Navy Hyattsville, Maryland 20782
Newport News, Virginia 23607 Attn: NSEC 6100 Ship Sys. Engr. & Des.
6102C Computer-Aided Ship Des.
Technical Director 6105 Ship Protection
Mare Island Naval Shipyard 6110 Ship Concept Design
Vallejo, California 94594 6120 Hull Div.
6120D Hull Div.
U.S. Naval Ordnance Station 6128 Surface Ship Struct.
Attn: Mr. Garet Bornstein 6129 Submarine Struct.
Research & Development Division
Indian Head, Maryland 20640 Naval Ordnance Systems Command
Department of the Navy
Chief of Naval Operations Washington, D.C. 20360
Department of the Navy Attn: NORD 03 Res. & Technology
Washington, D.C. 20350 035 Weapons Dynamics
Attn: Code OP-07T 9132 Tech. Library

Deep Submergence Systems Engineering Department

Naval Ship Systems Command U.S. Naval Academy
Code 39522 Annapolis, Maryland 21402
Department of the Navy
Washington, D.C. 20360
Attn: Chief Scientist

27
Air Force Director
National Bureau of Standards
Commander WADD Washington, D.C. 20234
Wright-Patterson Air Force Base Attn: Mr. B.L. Wilson, EM 219
Dayton, Ohio 45433
Attn: Code WWRMDD National Science Foundation
AFFDL (FDDS) Engineering Division
Structures Division Washington, D.C. 20550
AFLC (MCEEA)
Code WWRC Director
AFML (MAAM) STBS
Defense Atomic Support Agency
Commander Washington, D.C. 20350
Chief, Applied Mechanics Group
U.S. Air Force Inst. of Tech. Commander Field Command
Wright-Patterson Air Force Base Defense Atomic Support Agency
Dayton, Ohio 45433 Sandia Base
Albuquerque, New Mexico 87115
Chief, Civil Engineering Branch
WLRC, Research Division Chief, Defense Atomic Support Agcy.
Air Force Weapons Laboratory Blast & Shock Division
Kirtland AFB, New Mexico 87117 The Pentagon
Washington, D.C. 20301
Air Force Office of Scientific Res.
1400 Wilson Blvd. Director Defense Research & Engr.
Arlington, Virginia 22209 Technical Library
Attn: Mechs. Div. Room 3C-128
The Pentagon
NASA Washington, D.C. 20301

Structures Research Division Chief, Airframe & Equipment

National Aeronautics and Space Admin. FS-120
Langley Research Center Office of Flight Standards
Langley Station Federal Aviation Agency
Hampton, Virginia 23365 Washington, D.C. 20553
Attn: Mr. R.R. Heldenfels, Chief
Chief of Research and Development
National Aeronautics & Space Admin. Maritime Administration
Associate Administrator for Advanced Washington, D.C. 20235
Research & Technology
Washington, D.C. 20546 Mr. Milton Shaw, Director
Div. of Reactor Develop. & Technology
Scientific & Tech. Info. Facility Atomic Energy Commission
NASA Representative (S-AK/DL) Germantown, Maryland 20767
P.O. Box 5700
Bethesda, Maryland 20014 Ship Hull Research Committee
National Research Council
Other Government Activities National Academy of Sciences
2101 Constitution Avenue
Technical Director Washington, D.C. 20418
Marine Corps Development & Educ. Command Attn: Mr. A.R. Lytle
Quantico, Virginia 22134

28
PART 2 - CONTRACTORS AND OTHER TECHNICAL Professor Paul M. Naghdi
COLLABORATORS Div. of Applied Mechanics
Etcheverry Hall
Universities University of California
Berkeley, California 94720
Professor J.R. Rice
Division of Engineering Professor W. Nachbar
Brown University University of California
Providence, Rhode Island 02912 Dept. of Aerospace & Mech. Engrg.
La Jolla, California 92037
Dr. J. Tinsley Oden
Dept. of Engr. Mechs. Professor J. Baltrukonis
University of Alabama Mechanics Division
Huntsville, Alabama 35804 The Catholic Univ. of America
Washington, D.C. 20017
Professor R.S. Rivlin
Center for the Application of Mathematics Professor A.J. Durelli
Lehigh University Mechanics Division
Bethlehem, Pennsylvania 18015 The Catholic Univ. of America
Washington, D.C. 20017
Professor Julius Miklowitz
Div. of Engr. & Applied Sciences Professor H.H. Bleich
California Institute of Technology Dept. of Civil Engineering
Pasadena, California 91109 Columbia University
Amsterdam & 120th Street
Professor George Sih New York, New York 10027
Department of Mechanics
Lehigh University Professor R.D. Mindlin
Bethlehem, Pennsylvania 18015 Dept. of Civil Engineering
Columbia University
Dr. Harold Liebowitz, Dean S.W. Mudd Building
School of Engrg. & Applied Science New York, New York 10027
George Washington University
725 23rd Street Professor A.M. Freudenthal
Washington, D.C. 20006 George Washington University
School of Engrg. & Applied Science
Professor Eli Sternberg Washington, D.C. 20006
Div. of Engrg. & Applied Sciences
California Institute of Technology Professor B.A. Boley
Pasadena, California 91109 Dept. of Theoretical & Applied
Mechanics
Professor Burt Paul Cornell University
University of Pennsylvania Ithaca, New York 14850
Towne School of Civil & Mech. Engrg.
Room 113 Towne Building Professor P.G. Hodge
220 So. 33rd Street Department of Mechanics
Philadelphia, Pennsylvania 19104 Illinois Institute of Technology
Chicago, Illinois 60616
Professor S.B. Dong
University of California Dr. D.C. Drucker
Department of Mechanics Dean of Engineering
Los Angeles, California 90024 University of Illinois
Urbana, Illinois 61801

29
Professor N.M. Newmark Professor A.C. Eringen
Dept. of Civil Engineering Dept. of Aerospace & Mech. Sciences
University of Illinois Princeton University
Urbana, Illinois 61801 Princeton, New Jersey 08540

Professor James Mar Dr. S.L. Koh

Massachusetts Institute of Technology School of Aero., and Engrg. Sei.
Room 33-318 Purdue University
Dept. of Aerospace & Astro. Lafayette, Indiana 47907
77 Massachusetts Avenue
Cambridge, Massachusetts 02139 Professor R.A. Schapery
Civil Engineering Department
Library (Code 0384) Texas A & M University
U.S. Naval Postgraduate School College Station, Texas 77840
Monterey, California 93940
Professor E.H. Lee
Dr. Francis Cozzarelli Div. of Engrg. Mechanics
Div. of Interdisciplinary Studies & Stanford University
Research Stanford, California 94305
School of Engineering
State University of New York Dr. Nicholas J. Hoff
Buffalo, New York 14214 Dept. of Aero. & Astro.
Stanford University
Professor R.A. Douglas Stanford, California 94305
Dept. of Engrg. Mechs.
North Carolina State University Professor Max Anliker
Raleigh, North Carolina 27607 Dept. of Aero & Astro.
Stanford University
Dr. George Herrmann Stanford,California 94305
Stanford University
Dept. of Applied Mechanics Professor Chi-Chang Chao
Stanford, California 94305 Div. of Engrg. Mechanics
Stanford University
Professor J.D. Achenbach Stanford, California 94305
Technological Institute
Northwestern University Professor H.W. Liu
Evanston, Illinois 60201 Dept. of Chemical Engrg. and Metal,
Syracuse University
Director, Ordnance Research Lab. Syracuse, New York 13210
Pennsylvania State University
P.O. Box 30 Professor S. Bodner
State College, Pennsylvania 16801 Technion R&D Foundation
Haifa, Israel
Professor J. Kempner
Dept. of Aero Engrg. & Applied Mech. Dr. S. Dhawan, Director
Polytechnic Institute of Brooklyn Indian Institute of Science
333 Jay Street Bangalore, India
Brooklyn, New York 11201
Professor Tsuyoshi Hayashi
Professor J. Klosner Department of Aeronautics
Polytechnic Institute of Brooklyn Faculty of Engineering
333 Jay Street University of Tokyo
Brooklyn, New York 11201 BUNKYO-KU
Tokyo, Japan

30
Professor J.E. Fitzgerald, Ch. Professor R.B. Testa
Dept. of Civil Engineering Dept. of Civil Engrg.
University of Utah Columbia University
Salt Lake City, Utah 84112 S.W. Mudd Bldg.
New York, New York 10027
Professor R.J.H. Bollard
Chairman, Aeronautical Engrg. Dept. Dr. Y. Weitsman
207 Guggenheim Hall Dept. of Engrg. Sciences
University of Washington Tel-Aviv University
Seattle, Washington 98105 Ramat-Aviv
Tel-Aviv, Israel
Professor Albert S. Kobayashi
Dept. of Mechanical Engrg. Professor W.D. Pilkey
University of Washington Dept. of Aerospace Engrg.
Seattle, Washington 98105 University of Virginia
Charlottesville, Virginia 22903
Professor G.R. Irwin
Dept. of Mech. Engrg. Professor W. Prager
Lehigh University Division of Engineering
Bethlehem, Pennsylvania 18015 Brown University
Providence, Rhode Island 02912
Dr. Daniel Frederick
Dept. of Engrg. Mechs. Industry and Research Institutes
Virginia Polytechnic Inst.
Blacksburg,Virginia 24061 Mr. Carl E. Hartbower
Dept. 4620, Bldg. 2019 A2
Professor Lambert Tall Aerojet-General Corporation
Lehigh University P.O. Box 1947
Department of Civil Engrg. Sacramento, California 95809
Bethlehem, Pennsylvania 18015
Library Services Department
Professor M.P. Wnuk Report Section, Bldg. 14-14
South Dakota State University Argonne National Lab.
Department of Mechanical Engineering 9400 S. Cass Avenue
Brookings, South Dakota 57006 Argonne, Illinois 60440

Professor Norman Jones Dr. F. R. Schwarzl

Massachusetts Institute of Technology Central Laboratory T.N.O.
Dept. of Naval Architecture & Marine Engrg. Schoemakerstraat 97
Cambridge, Massachusetts 02139 Delft, The Netherlands

Professor Pedro V. Marcal Dr. Wendt

Brown University Valley Forge Space Technology Cen.
Division of Engineering General Electric Co.
Providence, Rhode Island 02912 Valley Forge, Pennsylvania 10481

Professor Werner Goldsmith Library Newport News Shipbuilding

Department of Mechanical Engineering & Dry Dock Company
Division of Applied Mechanics Newport News, Virginia 23607
University of California
Berkeley, California 94720

31
Director Mr. D. Wilson
Ship Research Institute Litton Systems, Inc.
Ministry of Transportation AMTD, Dept. 400
700, SHINKAWA El Segundo
Mitaka 9920 W. Jefferson Blvd.
Tokyo, Japan Culver City, California 90230
Dr. H.N Abramson Dr. Kevin J. Forsberg, Head
Southwest Research Institute Solid Mechanics
8500 Culebra Road Orgn 52-20, Bldg. 205
San Antonio, Texas 78206 Lockheed Palo Alto Research Lab.
Palo Alto, California 94302
Dr. R.C. DeHart
Southwest Research Institute Dr. E.M.Q. Roren, Head
8500 Culebra Road Research Department
San Antonio, Texas 78206 Det Norske Veritas
Post Box 6060
Mr, Roger Weiss Oslo, Norway
High Temp. Structures & Materials
Applied Physics Lab. Dr. Andrew F. Conn
8621 Georgia Avenue Hydronautics, Incorporated
Silver Spring, Maryland 20910 Pindell School Road, Howard County
Laurel, Maryland 20810
Mr. E.C. Francis, Head
Mech. Props. Eval.
United Technology Center
Sunnyvale, California 94088

Mr. C.N. Robinson

Atlantic Research Corp.
Shirley Highway at Edsall Road
Alexandria, Virginia 22314

Mr. P.C. Durup

Aeromechanics Dept., 74-43
Lockheed-California Co.
Burbank, California 91503

32