Copyright
Arthur C Macurdy.

Exploratory investigation of nonwelded pressure hulls for hydrospace vehicles online

. (page 1 of 3)
Online LibraryArthur C MacurdyExploratory investigation of nonwelded pressure hulls for hydrospace vehicles → online text (page 1 of 3)
Font size
QR-code for this ebook


tF,



mz




DEPARTMENT OF THE NAVY



©TON



HYDROMECHANICS



EXPLORATORY INVESTIGATION OF NONWELDED
PRESSURE HULLS FOR HYDROSPACE VEHICLES



AERODYNAMICS



by



Arthur C. Macurdy



STRUCTURAL
MECHANICS



APPLIED
MATHEMATICS



w




&C TICS AND

/ NATION

,D3



_



648 (Rev 1-64)



STRUCTURAL MECHANICS LABORATORY
RESEARCH AND DEVELOPMENT REPORT



March 1964



Report 1762



EXPLORATORY INVESTIGATION OF NONWELDED
PRESSURE HULLS FOR HYDROSPACE VEHICLES



by



Arthur C. Macurdy



March 1964 Report 1762

S-ROll 01 01



TABLE OF CONTENTS

Page

ABSTRACT 1

INTRODUCTION 1

DESCRIPTION OF MODELS 2

TEST PROCEDURES AND RESULTS 3

DISCUSSION 4

SUMMARY 9

RECOMMENDATIONS 9

ACKNOWLEDGMENTS 9

REFERENCES 10

LIST OF FIGURES

Page

Figure 1 - Models Before Test 12

Figure 2 - Sketches of Models 14

Figure 3 -Measured Strain Sensitivities 15

Figure 4 - Models After Test 18

Figure 5 -Mechanical Joining Techniques 19



LIST OF TABLES



Page



Table 1 - Comparison of Experimental and Theoretical Strain

Sensitivities at Midbay 20

Table 2 - Ratios of Theoretical Collapse Pressures to Experimental

Collapse Pressures 20

Table 3 - Experimental and Prototype Collapse Pressures 21



ABSTRACT

The feasibility of fabricating a deep -submergence
pressure hull composed of rings joined by means other than
welding was explored in tests of three structural models
(two of aluminum and one of titanium) designed for a
collapse strength of 10,000 psi. These tests demonstrated
that hulls can be built without welding and that these
hulls can have collapse strengths comparable to monolithic
hulls. Longitudinal strength, watertight integrity, and
corrosion protection, which were not explored by these
tests, can be provided by any of several mechanical
techniques without compromising collapse strength.

INTRODUCTION

Growing interest in the ocean depths both in scientific and military
circles has emphasized the need for vehicles capable of operating at great
depths. It is recognized that the materials and the fabrication techniques
now in use place a severe limitation on the size and maximum operating
depth of positively buoyant vehicles. Much interest has been shown in
using light, high-strength materials such as aluminum, titanium, glass-
reinforced plastics, and high-strength steels. Solid glass is also being
investigated.

The most difficult problem in utilizing these new materials is that
very few of them can be welded and still retain their strength character-
istics. In addition, the thick sections required for larger diameter hulls
make welding extremely difficult and expensive even where it is possible.

Techniques have been proposed for eliminating or reducing the

welding problem where large rings are used to form the strength elements

of the cylindrical pressure hull. Reynolds proposed a hull of aluminum

rings held together with tension rods and covered with a thin coating of

pure aluminum for corrosion protection. The latest design of the ALUMINAUT

2
pressure hull cylinder is composed of rings held together by bolts. The

David Taylor Model Basin is currently investigating composite-type con-
struction in which the strength rings are secured and protected by a thin

3
outer jacket of weldable or otherwise easily fabricated material.



References are listed on page 10.



For this study three structural models, two of aluminum and one of
titanium were assembled from machined rings and were tested to determine
whether hulls built from separate rings could have collapse strengths
comparable to monolithic hulls. This type of hull, consisting of separate
rings, would be adaptable to a variety of techniques for attaining longi-
tudinal strength and watertight integrity. This report summarizes the
results of the tests of these three models .

DESCRIPTION OF MODELS

The three models were designated PJ-1S, PJ-1L, and PJ-2. Models
PJ-1S and PJ-1L were made from 7079-T6 aluminum alloy and Model PJ-2 from
6Al-4Va titanium alloy. Yield strengths of 62,000 and 150,000 psi were
used in the design calculations for aluminum and titanium, respectively.
Based on these strengths, the models were designed for a collapse strength
of 10,000 psi, A value of 10.5 x 10 psi was assumed for the Young's
modulus of the aluminum, and a value of 16.0 x 10 psi for the titanium. A
Poisson's ratio of 0.3 was assumed for all material. The models are shown
in Figure 1, and relevant dimensions are given in Figure 2.

The aluminum models (PJ-1S and PJ-1L) had rectangular inside frames
and heavy shell segments of uniform thickness. The typical bay weighed
67.9 percent of its displacement weight in sea water. The area per unit
length of the hull section was obtained by allowing the average two-
dimensional Hencky-Von Mises stress to reach about 62,000 psi at a pressure
of 10,000 psi. A highly stable shell structure is required to permit

utilization of the full -yield strength of the material. To this end, the

4
frames were designed for a general-instability pressure of about 30,000

psi. A short frame spacing was selected to minimize bending in the shell
due to hydrostatic loads and to provide a high elastic-shell buckling
strength. Model PJ-1S was 1.7 diameters long, representing a finite com-
partment length, and was closed at the ends with flat aluminum plates;
Model PJ-1L was four diameters long, approximating a semi -infinite cylinder,
and had hemispherical end closures. Models PJ-1S and PJ-1L were designed
to evaluate the effect of bulkhead spacing. Both models had identical
typical bays and had grooves at each frame. These grooves were filled with



an aluminum-impregnated, epoxy, which sealed the joint and provided limited
longitudinal strength.

The titanium model (Model PJ-2) had a double shell separated by
thin webs. The typical bay weighed 53.8 percent of its displacement weight
in sea water. In this model, area per unit length of the hull section was
obtained by allowing the average circumferential stress over the section to
reach 150,000 psi at a pressure of 10,000 psi. The depth of the web and
the average thickness of the shells were chosen to obtain an elastic
general -instability pressure for a semi-infinite cylinder of about
25,000 psi. The shell segments were designed to eliminate bending due
to the hydrostatic load. This was done by varying the thickness of the
shell in the longitudinal direction. Also, as in the first two models,
the end bays of Model PJ-2 were shorter than the typical bay.

Model PJ-2 was 1.6 diameters long and was closed at the ends with
flat steel plates. Grooves on the inside and outside at each web were
filled with an artificial rubber compound. The artificial rubber was
chosen for ease of application in the laboratory, rather than as a suggested
prototype material. It was also used to seal the closure plates to the
model .

TEST PROCEDURES AND RESULTS

The models were instrumented with foil-type resistance strain
gages. The arrangement of the gages is shown in Figure 3. Model PJ-1S was
tested to collapse in the 17-inch diameter, high-pressure test tank at
the Model Basin. Model PJ-1L was tested almost to collapse in the 17-
inch pressure tank and collapsed in the 9-inch-diameter, high-pressure test
tank. Model PJ-2 was tested to 7000 psi in the 13-inch diameter, high-
pressure test tank and collapsed in the 9-inch pressure tank. Strain data
were obtained only during tests in the 17 and 13-inch tanks. At least
three runs were made for each model to obtain strain data. The loads were
applied to Model PJ-2 at the same rate as the compression specimens.

Model PJ-1S collapsed at 12,500 psi, Model PJ-1L at 11,900 psi, and
Model PJ-2 at 10,100 psi. Strain sensitivities devised from the initial
slopes of the pressure-strain plots are given in Figure 3. Figure 4 shows
Models PJ-1L and PJ-2 after test.



DISCUSSION



The measured strains, presented in Figure 3, are compared with
theoretical strains in Table 1. The agreement with theory is very good,
indicating that the elastic strains were not affected by the structural
discontinuities at the frames. The calculations are from the theory of

r 7

Salerno and Pulos as presented by Lunchick and Short.

All of the models failed by inelastic general instability in the
n=2 mode. This is indicated by the appearance of the models after test
(Figure 4) as well as by the theoretical calculations. The computed
collapse pressures for Models PJ-1S and PJ-1L indicate a very high degree
of stability in the elastic shell buckling modes and a margin of at least
20 percent in the inelastic shell buckling modes; see Table 2.

Because of the highly stable shell design, the theoretical in-
elastic shell buckling pressures ' ' correspond to strain levels in excess
of those measured in tests of material compression specimens. The ratios
reported in Table 2 correspond to average strain levels of 1.5 percent. No
theory is available to compute the shell-buckle pressures for nonuniform
shells such as those of Model PJ-2. All of the theories presented in
Table 2 assume the models to be of monolithic construction.

The experimental collapse pressures and the scaled collapse strengths
are given in Table 3. To compare the ring models with monolithic hulls,
data from two other models are also included in Table 3. Model DSRV-P 3
a small machined-aluminum model, similar to Model PJ-1S was 1.4 diameters
long and had a modified-Bryant critical buckling pressure of 3.55 times its
collapse pressure. Since it is impossible to machine a monolithic sandwich
hull such as Model PJ-2, a similar two-piece hull is included in this
discussion for comparison. Model OV-4 was made by inserting a cylinder
with outside rectangular frames into a closely fitted jacket, which formed
the outer shell. Model OV-4 11 had nearly the same semi-infinite, elastic,
general-instability collapse pressure as Model PJ-2 and was 4 diameters
long.



Generally, the collapse strength of models that fail inelastically
is proportional to the yield strength of the material and to the weight-
to-displacement ratios. This permits comparison of collapse strengths
among similar models. The scaled collapse pressures for Models PJ-1S and
PJ-1L were obtained by scaling the model yield strengths to 62,000 psi.
The scaled collapse pressure for Model PJ-2 represents a semi-infinite
hull of the same typical bay geometry and with a yield strength of 150,000
psi. The collapse strength of Model DSRV-P was scaled to 62,000 psi yield
strength and a weight-to-displacement ratio of 67.9 percent. The collapse
strength of Model OV-4 was scaled to a yield strength of 150,000 psi and a
weight-to-displacement ratio of 53.8 percent.

The comparisons shown in Table 3 illustrate that ring construction
need not result in any sacrifice in collapse strength compared to monolithic
hulls. Since the distortion and weakening effect of welding stresses are
not present, ring construction may permit some increase in collapse strength
relative to welded hulls.

A submarine pressure hull is designed principally to resist external
hydrostatic pressure. However, in addition to the hydrostatic loads, a
submarine or other structure is subjected to overall bending moments. To
resist these bending moments, the structure must possess longitudinal
tensile strength in addition to its hydrostatic collapse strength. This
tensile strength is required only when operating on or near the surface.
At deeper depths, the longitudinal, hydrostatic compressive load exceeds
the tensile load of the bending moments and less longitudinal tensile
strength is required. For example, an oceanographic research station and
bottom-based vehicles could be assembled on site with a minimum of tensile
bending.

For vehicles which must operate on the surface or at high speed,
some form of mechanical joining is required. The fundamental considerations
in designing a joining device are weight, volume, and stress concentration
in the pressure hull. For a given level of longitudinal strength, it is
desirable that the least excess weight be added to the structure by the
joints. In addition, it is desirable to consume as little as possible of
the valuable interior space. Any joining procedure which induces stress
concentration in the pressure hull may lower the collapse strength or



induce a fatigue problem. Many mechanical joining techniques have been
proposed for this application. A few are considered here. The following
discussion presents several typical ideas and some of their strengths and
weaknesses .

The rings could be held in place by tension wires or rods secured
at the bulkheads (Figure 5a) . This method has the advantage of low
weight since very high-tensile-strength steel may be used. One problem
involves the location of the rods. They cannot pierce the frames or webs
since, this would introduce severe stress concentrations particularly for
nonductile materials. If they are inboard of the frames, they are less
efficient structurally and consume valuable interior space. Ideally, they
would be placed outside the hull, but there they are subject to mechanical
damage and to corrosion.

Bolts might be used to secure the rings (Figure 5b). This technique
is being considered for ALUMINAUT. An important problem here is the
stress concentration in the bolt holes. Large, round taper pins might
overcome the stress-concentration problem of the holes (Figure 5c), but
they would be loaded in shear and would have a stress-concentration problem
under longitudinal loading.

Clamps and similar devices carry no hydrostatic load and may be
prohibitively heavy for joining individual shell and frame rings. However,
sections consisting of several frame and shell sections could be joined
by a system as shown in Figure 5d. The sections could be assembled by one
of the other techniques considered here or, for small diameter hulls,
could be machined. This procedure is being considered for deep-running
torpedoes . The clamping ring may be formed in several segments which are
joined by bolts.

In the technique of composite construction, the rings are inserted
into a jacket of weldable or otherwise easily fabricated material (Figure
5e). The jacket yields during the initial submergence, holds the rings in
place, and provides watertight integrity and corrosion protection. The
main difficulty with this technique is in the repair of the jacket. Some
materials, particularly aluminum and glass -re in forced plastics, are heat
sensitive, and heat applied in welding the jacket might seriously weaken
the inner rings. This problem could be overcome by using an epoxy or



glass-reinforced plastic jacket or, possibly, an intermediate heat shield.

The rings could also be designed to be self-locking. This could
be done by allowing for positive interference and shrink-fitting the
rings together (Figure 5f ) . Aside from the difficulties in fabrication,
this technique precludes later separation of the rings and may be a source
of stress concentration if close tolerances are not maintained. The rings
could also be made with threads and screwed together (Figure 5g); here the
main problem is the stress concentration at the root of the threads.

One of the most interesting joining and sealing techniques available
is the use of adhesives . Adhesives have the significant advantages of low
weight and economy of application. They can be used either alone or in
combination with other mechanical bonds.

The availability of epoxies or similar materials with the necessary
bonding properties and tensile strengths has not been established. It is
felt that the rapidly expanding adhesive technology will be able to provide
a suitable material if the need is established. In addition, adhesives
have two critical defects which impair their usefulness in deep-submergence
applications; one is a problem in joint design and the other a problem in
materials .

An epoxy adhesive was used on the deep-submergence Krupp sphere of
the bathyscaphe TRIESTE to hold the three segments of the sphere together „
The joint was a simple glue-line with a layer of epoxy between the two
metal surfaces. When the Krupp sphere surfaced after the first deep dive
(Dive 61), the two joints were parted. Fortunately, there was no immediate
danger to the occupants of the bathyscaphe since the segments were held

together by the pressure of the water; for further dives, the sphere was

12
held together by a system of rings and bands.

The failure of the TRIESTE bond was probably caused by the dete-
rioration of the epoxy under high compressive loads . Epoxy resins and
other adhesives have relatively low Young's moduli and compressive yield
strengths. This means, of course, that they may experience excessive
deformation when subjected to the same stress as a metal. Conversely, if
the adhesive is subjected to the same deformation as a metal, it will
carry a very much smaller stress. The models in this report illustrate
one type of joint where the adhesive may act as a bond and seal, without



being subjected to high compressive stresses. Much work is required to
evaluate this, and other, improved designs for adhesive joints. It is
important that the presence of the groove for the adhesive joint did not
seem to impair the collapse strength of the models.

In addition to deterioration due to loading, adhesive materials are
subject to eventual deterioration due to exposure to salt water and to
biological fouling. These material problems must be solved by developing
new adhesives and material -protection techniques. Until they are solved,
adhesives will be limited to short-term applications or will require
periodic replacement.

The primary advantage of ring construction is that it permits the
use of many light, high-strength nonweldable materials for deep-submergence
pressure hulls; but there are a number of other features which may
produce considerable savings . The combining of rings into a pressure hull
by mechanical means is essentially a faster and less expensive operation
than welding and requires less time spent in the shipway. This potential
saving is balanced by the greater machining costs associated with ring
construction. In addition, it is often necessary to open a hull to make
repairs or to replace machinery. It is not possible to open a hull by
cutting or burning a ring since it cannot ordinarily be rewelded; but it
may be practical to separate two rings at a joint, and then reassemble the
section after the repair work is completed.

Present forging capacity limits the maximum size of nonweldable
metals to a diameter of about 12 to 15 feet. Aluminum must be forged to
obtain its highest yield strength, but titanium and steel may be welded and
then heat treated to higher strengths. It is felt that the primary
applications of ring construction will be in structures of smaller
diameter. The possibilities of ring construction in onsite assembly of
undersea laboratories and in oceanographic research vehicles have already
been mentioned. In addition, the technique can find military applications
for deep-running-torpedo housings or for submerged missile-silos.



SUMMARY

1. A submarine pressure hull built of separate rings may be so designed as
to have a hydrostatic collapse strength comparable to that of a monolithic
hull.

2. The technique of ring construction permits an increase in static
collapse strength compared to current steel hulls of the same weight,
through the use of light, high-strength, nonweldable materials. This
technique may also result in savings in fabrication and assembly costs,
which would make it practical for use with weldable materials .

3. A wide variety of mechanical techniques is available to provide the
necessary longitudinal strength, watertight integrity, and corrosion
protection for hulls of ring construction. The expanding technology of
epoxy plastics may provide materials suitable for these applications.



RECOMMENDATIONS

Future research should include studies of the effects of hull
bending moments and longitudinal tensile loads on various types of
mechanical joints.



ACKNOWLEDGMENTS

The author wishes to acknowledge the assistance of Mr. Martin A.
Krenzke who initiated the project and designed Models PJ-1S and PJ-1L.



REFERENCES

1. U. S. Patent 3,029,966 (April 7, 1962), Reynolds, J. L., "Sub-
mersible Pressure Vessel."

2. Walsh, J. B., "Strength of the ALUMINAUT Hull," Woods Hole
Oceanographic Institute, Ref. No. 62-32 (Apr 1962).

3. Krenzke, M. A. and Kiernan, T. J., "Structural Development of a
Titanium Oceanographic Vehicle of Operating Depths of 15,000 to 20,000
Feet," David Taylor Model Basin Report 1677.

4. Pulos, John G., "Structural Analysis and Design Considerations
for Cylindrical Pressure Hulls," David Taylor Model Basin Report 1639
(Apr 1963).

5. Short, R. D. Jr., "Membrane Design for Stiffened Cylindrical Shells
under Uniform Pressure," David Taylor Model Basin Report (in preparation).

6. Salerno, V. L. and Pulos, J. G., "Stress Distribution in a
Circular Cylindrical Shell under Hydrostatic Pressure," Polytechnic
Institute of Brooklyn Aeronautical Laboratory Report No. 171-A (1951).

7. Lunchick, M. E. and Short, R. D., Jr., "Behavior of Cylinders
with Initial Shell Deflection," David Taylor Model Basin Report 1150
(July 1957).

8. Reynolds, T. E., "Inelastic Lobar Buckling of Cylindrical Shells
under External Hydrostatic Pressure," David Taylor Model Basin Report 1392
(Aug 1960).

9. Lunchick, M. E., "Plastic Axisymmetric Buckling of Ring-Stiffened
Cylindrical Shells Fabricated from Strain-Hardening Materials and Subjected
to External Hydrostatic Pressure," David Taylor Model Basin Report 1393
(Jan 1961).

10. Lunchick, M. E., "Graphical Methods for Determining the
Plastic Shell Buckling Pressures of Ring-Stiffened Cylinders Subjected to
External Hydrostatic Pressure," David Taylor Model Basin Report 1437
(Mar 1961).



10



11. Horn, K. and Blumenberg, W. F., "Hydrostatic Tests' of Structural
Models for Preliminary Design of a Web-Stiffened Sandwich Pressure Hull,"
David Taylor Model Basin Report 1763, September 1963.

12. Picard, J. and Dietz, R. S., "Seven Miles Down," Putnam, 1961.



11



Figure 1 - Models before Test




Figure la - Model PJ-1S




gPSD 3084:
Figure lb - Model PJ-1L



12




,PSD 310594
Figure lc - Model PJ-2




PSD M0^=>
Figure Id - Model PJ-2 (Showing Rings)



13




Decrees


90 Degrees


€ <f>


{ x


V


e x




-0.28




-0.28


-0.29


-0.29


-0.29


-0.28


-0.28


-0.28


-0.29





-0.25


-0.12


-0.26





-0.45


-0.47


-0.47


-0.37


-0.50


-0.37


-0.49


-0.38



d


^





-0.51


30


-0.52


60


-0.50


90


-0.50


120


-0.49


150


-0.51


180






Figure 3b - Model PJ-1L

6 Circumferential location of gages

£ i Circumferential strain sensitivity in fiin/ia. per psi

£ Longitudinal strain sensitivity in fiia/in per psi



16



Inside


d


<<!>


{ x





-0.62


-0.76


30


-0.63




60


-0.63




90


-0.60




120


-0.60


-0.58


150


-0.63




180







210


-0.56




240





-0.56


270


-0.62




300


-0.57




330


-0.64






Outside


6


^


e x





-0.68


-0.62


30


-0.66




60


-0.60




90


-0.62




120


-0.62


-0.54


150


-0.60




180


-0.65




210


-0.65




240


-0.70


-0.64


270


-0.61




300


-0.61




330


-0.65





Figure 3c - Model PJ-2

Circumferential location of gages

( i Circumferential strain sensitivity in ^tin/in per psi

a Longitudinal strain sensitivity in /zin/in per psi



17




Figure 4a - Model PJ-1L




Figure 4b - Model PJ-2
Figure 4 - Models After Test



18




TENSION RODS
OR WIRES
SECURED TO
BULKHEADS




Figure 5a - Tension Wires or Rods



PIERCING FRAMES



fHf

Figure 5b - Bolts




Figure 5c - Taper Pins



hi



RING SEAL



\v\\\\\\\VW\V?




THREE SEGMENT CLAMPING RING



IMHtt



Figure 5d - Clamping Ring Joining Sections of Several Frames





Figure 5e - Composite Construction



Figure 5f - Self -Locking




Figure 5g - Threads
Figure 5 - Mechanical Joining Techniques



19



TABLE 1

Comparison of Experimental
and Theoretical Strain Sensitivities'* at Midbay



Location


Theoretical


Average PJ-1S


Average PJ-1L


Circumferential
Inside
Outside


0.56
0.50


0.58
0.49


0.50


Longitudinal
Inside
Outside


0.15
0.34


0.16
0.34


t
0.38


"Strain sensitivities in M-in/in/psi.

"Theory of Salerno and Pulos.


1 3

Online LibraryArthur C MacurdyExploratory investigation of nonwelded pressure hulls for hydrospace vehicles → online text (page 1 of 3)