FE LISA | Technical Report STRUCTURAL DESIGN OF CONICAL ACRYLIC VIEWPORTS June 1970 Sponsored by NAVAL FACILITIES ENGINEERING COMMAND NAVAL CIVIL ENGINEERING LABORATORY Port Hueneme, California This document has been approved for public release and sale; its distribution is unlimited. STRUCTURAL DESIGN OF CONICAL ACRYLIC VIEWPORTS Technical Report R-686 YF 38.535.005.01.005 by M. R. Snoey and M. G. Katona ABSTRACT The purpose of this report is to establish a rational engineering approach for the design of conical acrylic viewports. To achieve this goal, a time-dependent, yield-failure criterion was developed and utilized in the analysis of a variety of viewport configurations. Specifically, a range of thickness/minor diameter (t/d) ratios from 0.25 to 1.75 and included angles from 60° to 120° were analyzed by the finite element technique. Using the viewport structural analysis in conjunction with the yield-failure criterion for acrylic, time-dependent operating depths were determined as a function of viewport configuration. Paralleling the above, an experimental investigation was performed to validate the analytical results. Six full-scale viewports were tested for a year under simulated operational conditions that included simultaneous cycling of pressure and temperature, 0 to 4,000 psi and 70 to 35°F, respectively. Comparison of analytical and experimental results indicated excellent agreement for the physical location of viewport failure at specified loading histories. Design recommendations are presented in the form of design curves which enable the design of a conical acrylic viewport for a specified operating pressure and duration under load. To complete the recommendations, design information is given also on sealing with a conventional O-ring, as well as guidelines for elevating a viewport in its flange. This document has been approved for public release and sale; its distribution is unlimited. Copies available at the Clearinghouse for Federal Scientific & Technical Information (CFSTI), Sills Building, 5285 Port Royal Road, Springfield, Va. 22151 NN YAU UA UL 0 0301 0043844 2 CONTENTS page HINTICERO DG ON eee etc scmcnen ctl eie tern deere rr ck bicacntes Oar ert 1 @ pd] CCUIVEST= ARR arena: nn Rena na eRe TR ROT ARE SR ERB cn 1 FU OSCIMN Seat FaN.h Hendhy HPEME, Sebi ORteN ENR OCT EY RCNP Tem, OM chet cr ner eta 1 SCOPEORINVESTIGATIONT Mes fa oii: nee Anes on meat ene 3 Exfoe nite mtalieiaSGricnet es 1c auencas ee, os ee aR an, aera 3 PNGTEIWATICENPIMEISSE Vee cog Sina’ oc ae eel oa ee Ae oe 5 Gene tal Ree Sater Asher fre err meme kein, Ce cena 5 Concepteon tall UNE en eye tenn Meee Mae ee 5 Matinod Ov ANBIVSIS . o0ereescceaveooveudovsos 6 RESULTS ANID IDISGUSSION gape ccuce otro sees csscoonuee 8 EXPeti(Melita lM arama meter te vtec ie ores atid c mthe, Renata 8 POSE TES OWSSVEWIONS . .cccccccccecssauvcvc0ce 11 Displacements settee tn ter eee ee eee ee 14 PNT allWatil Cal ll eee seater eng ate at ey ns a te GR oe or 14 SI CSSCS aI ROS eee Pee IY Lee eee ae 14 BiSGIACE MENS pada cotin Oey een tee ppeeeiotech eer ep saeeees 16 Comparison of Experimental and Analytical Results ....... 16 DISCUSSION sveeeben eo es Pikes NTA ons Gees AIS ie a Senos 24 FaIIN DING Somer tan See rene aes ee ree ae NO he Rn aca loan, e ZAS) DESIGN RECOMMIENIDAIMIONS ccoccsscacasnacvnunausc0gges 27 BSI Ka LAM CRUISE aig ave, ks ace torent ere eee peer cner chi arenes meterire int crea 27 NATE TCT ot ts cae anaes rs cer aati tee Mg ape GRU a ae neue 27 S UCU CR peta ar ct ecleemepthe Petree akon yb rctys Rue cea are ag ee ia 28 E Xam Ole® Bice ein ie een MONEE Ea Abst Mahar olesiteNes grea ie aan een ens 30 AGN OWED. GINIEINT siess sunken ee twa eiaeis Giais = inva aon es erie 31 ee Be Vi | page APPENDIXES A—Investigation of a Viewport Seal Design ............... 32 B—Equipment and Procedure for Experimental Phase....... 41 C—Failure Criterion Applied to Conical Acrylic Viewports ... 47 REE ERENCES a ce tee os elo tits Seek aa AE pene Sra Bs 53 ELSTOIF SVIMIB OHSS ice iss Seam yh nam ane rar a eee 55 INTRODUCTION Objectives 1. Accumulate experimental test data on full-scale viewports under simulated operational conditions. 2. Develop a rational failure criterion for conical acrylic viewports. 3. Develop parametric design curves based on the failure criterion. Purpose The presence of viewports in practically every undersea vehicle is indicative of their importance for visibility in undersea research. Piccard! first introduced conical acrylic viewports in 1939 and, presently, 18 submer- sibles, or about half the total number, utilize Piccard’s 90° conical frustum design for a wide range of operating depths. Figure 1 shows full-scale, 90° conical viewports while Reference 2 contains a historical background on viewports along with a summary of viewport designs for about 40 vehicles. This study is intended to satisfy a need in the design of conical viewports that has existed and has seen no improvement for the past 30 years. The need was to develop parametric design curves based on the viewports functional use and to include time effects (creep) in the design curves. A thorough structural analysis of a conical viewport was not even available until just recently.2’4-® All three of these references provide good insight into the structural response of a conical viewport, however, they do not provide any information for the designer. Although acrylic viewports have accumulated a large amount of operational time, thereby generating some confidence in the material, trepi- dations still do exist. Design curves based on accurate experimental material tests, a rational failure theory, and complete structural analysis should relieve these fears and provide more confidence in viewport design. Design recommendations for conical acrylic viewports would not be complete without a section on sealing. Therefore, Appendix A contains information on methods of sealing and how they affect the viewport design economically, operationally, and structurally. Figure 1. Full-scale, 90° conical acrylic viewports. SCOPE OF INVESTIGATION The investigation was divided into two sections: (1) the experimental phase which will be discussed first followed by (2) the analytical phase. Experimental Phase The objective of the experimental tests was to accumulate data on full-scale, 90° conical viewports with a range of thickness to minor diameter ratios, t/d. Figure 2 shows a cross section of a typical viewport. The test conditions simulated actual operational conditions as nearly as possible so that the results could be used to provide insight on the type and location of failures occurring during operation. More importantly, the data would serve to test the validity of the analytical results. Tests were conducted on six viewports of three sizes as listed in Table 1. The viewports were all machined from a 4-inch-thick commercial cast sheet of Plexiglas G acrylic plastic. The minor diameters were approxi- mately 4, 6, and 8 inches which gave t/d ratios of 0.92, 0.61, and 0.45, " respectively. Table 1. Full-Scale Experimental Viewport Dimensions Included Major Minor : ; F Thickness, t Angle, & Diameter, D Diameter, d fea in. (+5') (in.) (in.) 16.025 8.431 Sy/iOM 0.451 A. | J VG 7 A, ‘/ tt T= rec Figure 2. Cross section of a typical viewport. The six viewports were tested simultaneously in three refrigerated pressure vessels. Displacements and post-test observations served as the recordable data. Details on the equipment and test procedure used are in Appendix B. Inasmuch as an undersea vehicle might make many dives, cycling of the pressure load was included as part of the test procedure. The simulation of actual dives required simultaneous changes in pressure and temperature. As the pressures were changed from O to 4,000 psi, the temperature simul- taneously was changed from 70 to 35°F. It was believed that the most severe test of the viewports would be to cycle the maximum creep strains occurring in each viewport under the 4,000 psi load. Therefore, the period of the cycle for all viewports was determined by loading the least conservative design, t/d = 0.45, and letting the viewport displace until the displacement- time curve was asymptotic within the accuracy of the measuring system. This resulted in a ‘‘load-on”’ time period of 500 hours with a “‘load-off”’ or relaxation period of 100 hours; the time necessary to completely relax all the strains so the viewport would return to zero displacement. The tests were conducted over a 1-year period which resulted in 13 total cycles. Analytical Phase General. The knowledge gained from the experimental tests was utilized as a basis for performing an analytical study on viewport designs. In particular, analytical investigation for a spectrum of viewport parameters, t/d and aranging from 0.25 to 1.75 and 60° to 120°, respectively, was ac- complished by means of the finite element technique. In order to develop rational design recommendations, it is necessary to define the capacity of the system, i.e., what constitutes failure. Since the failure mode definition and failure criteria influence the assumptions made in an analytical study, the concept of failure is discussed prior to the analy- tical method of attack. Concept of Failure. Failure must be considered from two viewpoints: (1) the structural level and (2) the material level. First, at the structural level, the investigator must define failure. In the past, structural viewport failure has been taken as the complete collapse of the system, often called the upper limit or ultimate strength, which is typified by large plastic flow and rupture. Another definition of failure is the lower limit capacity of a system which is defined as that load which causes initial plastic yielding in any local region of the system. The authors have chosen the lower limit or “yield criterion’’ as the definition for failure of acrylic viewports based on the following considera- tions. (1) The functional use of a viewport is to transmit undistorted and undiminished light to the viewer, however, earlier experimental results have shown that this function is seriously impaired when the viewport is loaded into the plastic range which results in crazing and distortion of the acrylic. (2) The pressure that causes yielding is much lower than the load causing ultimate collapse failure, consequently, a built-in safety factor for the hazardous environment is intrinsic in the lower limit design, in addition to the standard safety factor used for the functional aspect. The second major consideration in establishing a failure criterion belongs to the realm of the material itself independent of the particular structure configuration and functional use. Since, by definition, the structure fails when local yielding occurs, this dictates that a yield criterion for acrylic must be established which will adequately predict yield states under combined stresses. Fortunately, much research has been accomplished in this area and it has been shown that the Huber-von Mises-Hencky or distortion energy theory of failure® predicts combined-stress yield-states extraordinarily well for acrylic.”“® However, this flow law is complicated by the anisotropic response of yielding in tension and compression with tensile strength being lower than the compressive. A conservative approach was used to circumvent this problem. The tensile strength was applied to combined states of stress at any time when at least one principal stress was significantly in a tension zone, and the higher com- pressive strength was used strictly in the all-compression failure octant. Because acrylic is a typical thermoplastic, two additional variables significantly affect its material properties: temperature and time. In general, when temperatures are increased, acrylic material properties decrease in value. This feature of acrylic makes it an ideal material for its proposed use as a viewport for an undersea vehicle because the ocean provides a low-temperature environment which enhances properties over those measured at room temperature. The design recommendations set forth in this report are based upon material response at room temperature, thus allowing the additional strength of the acrylic due to lower temperatures at operational depths to increase the margin of safety. The second variable, time, is not as easily dealt with as temperature, but is certainly important due to the creep properties of acrylic. For this reason every effort has been made to rationally treat the effect of time on the stress distributions of the viewport and the yield strength of the acrylic. The details of this approach are given in Appendix C as a separate topic because this treatment is not limited to viewports; the approach may be used for any structure utilizing acrylic. Briefly, the approach is to choose the ‘‘worst”’ stress distribution and utilize this for any value of the parameter time, while the development of a yield-stress versus load-duration curve provides the factor of time in the design parameters. Appendix C deals with the concepts in a straightforward and rational manner. Method of Analysis. As discussed in Appendix C, the highest stress concentrations result from the viscoelastic solution when time is equal to zero. This is identical to elastic solutions; consequently, elastic solutions based on all the classical assumptions of linear, infinitesimal-strain, elastic theory were assumed and the solutions were obtained by the finite element program for an axisymmetric solid written by Wilson.2 References 4 and 5 verify the capability of using a finite element analysis on conical acrylic viewports. ea) The sequence of analysis was as follows: for each viewport configuration, the finite element program computed the local state of stress at every element and the corresponding effective stress defined as 05 = (log = Ale 2 (Ga 2 Gale oa Ole 2 where 0, = effective stress at an element per unit of applied pressure 01,097,903 = principal stresses at each element per unit of applied pressure Thus the relationship for yielding becomes ana 9, where o. = the maximum effective stress of all elements with all the principal stresses in compression KNOWN BY FINITE ELEMENT SOLUTION o, = the maximum effective stress of all elements with any tension in the principal stresses o, = material yield strength in compression for a particular loading duration KNOWN BY YIELD- STRESS VS TIME or CURVES en on : : o, = material yield strength in tension for a particular loading duration p = the maximum pressure that can be applied to the viewport which initiates yielding In any zone with all the principal stresses in DESIGN compression REQUIREMENT or p = the maximum pressure that can be applied to the viewport which initiates yielding in any zone with tension in the principal stresses The above format explicitly sets forth the procedure used to determine the capacity of the various viewport geometries. A further point of concern was in specifying the boundary condition between the viewport and the flange. In order to account for all practical contingencies, two extreme boundary conditions were applied to each view- port configuration as shown in Figure 3: (1) the boundary was fixed, repre- senting maximum friction between the acrylic viewport and its flange; and (2) the boundary was allowed to be frictionless, representing perfectly smooth, greased interfacing. Of these two boundary conditions, the one resulting in the controlling failure as outlined above, was chosen as the governing boundary condition. A structural response discovered during the course of the analysis was that an arbitrarily high stress concentration existed in the element at the corner of the low-pressure face for all viewport configurations and boundary conditions. This phenomenon correlated with experimental findings where it was discovered that the low-pressure face corner underwent deformation or ‘‘rounding”’ for even the smallest load magnitudes (see Figure 4). An intensive analytical investigation, outlined in Appendix C, revealed that the spiked stress concentration was relieved by deforming the corner of the mathematical model as was suggested by the experimental models. Thus, the viewport geometries considered in this report were all analyzed with a predeformed corner characterizing an actual viewport after its initial loading. The last analytical refinement considered was to investigate the response from an O-ring groove machined in the acrylic. The analytical results along with the experimental tests are presented in Appendix A. Input to Wilson’s code? required material properties for acrylic. Reference 5 describes experimental tests performed to determine these necessary values at room temperature as: modulus of elasticity, 444,000 psi; and Poisson’s ratio, 0.4. RESULTS AND DISCUSSION Experimental Post-test visual observations and displacements at the center of the viewport low-pressure face served as experimental data. Since the viewports were all exposed to the same pressure loading and environment, the cause of any post-test effects would be the differences in structural design (all other variables considered equal). (a) Fixed boundary. S 0 me © ee O be © Pe i ©) (b) Free boundary. Figure 3. Boundary conditions for finite element analysis. *SajDA9 U9aqI UY} Ja}je LOUMAIA Gy'O = P/} $O UO! PUOD jedISAUd 4Sa}-1SOg “py a4nbl4 0.75 : 1:28 0, = (contour number) (0.25) 1.75 Thickness/Minor Diameter Ratio, t/d Figure 12. Effective stress contour and surface plots for 90° angle and free boundary condition. 21 0.75 t 9, = (contour number) (0.25) 1.75 Thickness/Minor Diameter Ratio, t/d Figure 13. Effective stress contour and surface plots for 120° angle and free boundary condition. Operating Pressure, p (psi) Axial Displacement/Minor Diameter Ratio, 6,/d x 103 T p = 1,000 psi é, Angle, & Condition & 60° ——-—Free Oo 90° Fixed © 120° 0 0.4 0.8 eZ. 1.6 2.0 Thickness/Minor Diameter Ratio, t/d Figure 14. Instantaneous axial displacements at center of low-pressure face. 8,000 6,000 predicted failure region 4,000 predicted safe region 2,000 —O— Analytical (fixed) =| —-—O0-—-— Analytical (free) ——e®—- Experimental (failure) —o— Experimental (marginal) —o— Experimental (no failure) Se ee | ee) (0) 0.4 0.8 12 1.6 2.0 Thickness/Minor Diameter Ratio, t/d Figure 15. Comparison of experimental and analytical results. 23 Figure 16 compares the instantaneous displacements at the center of the low-pressure face for both the experimental and analytical work. The two analytical cases (fixed and free) successfully bracketed the experimental results. The smooth surface finish of 32 rms on the viewport flanges, in combination with the liberal use of silicone grease at the viewport-flange interface, resulted in the experimental displacements approaching the analy- tical free boundary condition. @ Experimental ——- — Analytical (free) —DO— Analytical (fixed) p = 1,000 psi Axial Displacement/Minor Diameter Ratio, 6,/d x 10% b 0.4 0.8 1.2 1.6 2.0 Thickness/Minor Diameter Ratio, t/d Figure 16. Comparison of experimental and analytical displacements. Discussion When comparing the design curves in this report with those in Reference 2, one notices a distinct difference. The curves in Reference 2, based on ultimate strength tests, indicate an exponential increase of strength 24 with respect to increases in the t/d ratio. This is understandable because the strength is based solely on physically forcing the material through the flange opening. Ultimate strength tests give an indication of the catastrophic load but do not indicate at what pressure the viewport fails to fulfill its function of providing visibility. The design curves in this report, however, are based on the visibility aspect of the viewports and show that the operating pressure or depth is limited by the yield strength of the material. As exhibited by Figures 8 through 13, the peak effective stress asymptotically approaches a finite maximum value as the t/d ratio is increased. No extensive creep data exist at temperatures other than room temperature. One can, however, get a comparison between 35 and TOM tests on viewports by comparing experimental displacement results of the t/d = 0.45 viewport in the main test, and the t/d = 0.45 viewport with O-ring in Appendix A. The displacement at 70°F with a time period of 500 hours was 7% greater than the displacement at 35°F for the same length of time, and after 4,800 hours the displacement was 20% greater. Figure 17 together with Table 3 compare the analytical results of this study with the present viewport designs in operational submersibles. A load duration time of 24 hours was chosen to include the emergency time of an average vehicle. From Figure 17 it appears that many of the present designs are conservative, which was probably due to lack of structural analysis at the time of design. FINDINGS Within the scope of this investigation, the following findings appear to be valid: 1. Based on both experimental and analytical results, the strength of a viewport increases with an increase in either the t/d ratio and/or the included angle for limits of 0.25 to 1.75 and 60° to 120°, respectively. 2. The maximum effective stress for a tensile region always occurs at the center of the low-pressure face with a fixed boundary condition. 3. The maximum effective stress for a compressive region always occurs at the corner of the low-pressure face with a free boundary condition. AS 8,000 nou od 6,000 predicted predicted failure safe region region 4,000 010 Operating Pressure, p (psi) 2,000 —D—__ Fixed boundary — -( — Free boundary [0] Actual designs (see table 3) ce) 0.4 0.8 1.2 1.6 2.0 Thickness/Minor Diameter Ratio, t/d Figure 17. Comparison of analytical results and operational designs. Table 3. Submersible Viewport Designs in Figure 17 (Extract from Reference 2.) Identification Number Submersible Name Operating Depth (ft) ALUMINAUT ALVIN ASHERAH BEN FRANKLIN DEERNEER DEEP QUEST DEEPSTAR 2000 DEEPSTAR 4000 DIVING SAUCER DOWB DSRV PAULO | PISCES IV ROUGHNECK SEA CLIFF STAR III SURV HU RGEE 1 2. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 26 4. In the actual viewport, a sharply spiked stress concentration exists at the low-pressure face corner which results in rounding of the sharp corner. 5. The transition to the dominance of flexural bending behavior occurs at about t/d = 0.6. 6. The stress distribution for conical frustums is independent of elastic modulus and only slightly dependent on Poisson's ratio, thus, the structural design curves are generally applicable to materials other than acrylic. 7. Aconventional O-ring design with the groove located in the viewport conical face satisfactorily fulfills the requirements for viewport sealing (Appendix A). DESIGN RECOMMENDATIONS The purpose of this section is to provide the designer with the necessary information to design a vehicle viewport accurately, quickly, and safely. The type of analysis used in this section was experimentally verified as previously described. In order to design a conical acrylic vehicle viewport, the designer must have available such information as: t/d ratio, included angle, elevation of viewport in flange, method of sealing, and flange surface finish. The design requirements would, of course, include the operating pressure, ambient temperature, and maximum length of dive. This section of the report deals with design curves for viewport dimensions and recommendations for elevation of viewport in flange. Design recommendations for the flange surface finish are in Reference 5, and an analysis of the viewport sealing problem is in Appendix A of this report. Annealing instructions for the viewport may be found in Appendix B or Reference 11. Design Curves Material. Figure 18 represents the material facet of the design or the relationship between the yield stress of acrylic and time. A more detailed discussion of the origin of these curves is presented in Appendix C. The figure contains both tensile and compressive curves. The yield strengths can be seen to reduce to about 50% of the instantaneous value after 500 hours. 27 These curves are only for room temperature, but a temperature of 35°F will increase the yield strength by about 20% according to References 4 and 10. The strength is, of course, much lower with a higher temperature and, therefore, temperatures considerably above room temperature should be avoided. Figure 18 is used concomitantly with Figure 19 for calculating design strength. Structure. Figure 19 represents the structural facet of the design. The curves include the parameters: t/d ratio, included angle, safety factor, yield stress, and operating pressure. The t/d ratios range from 0.25 to 1.75 while the included angles range from 60° to 120°. The curves are based on a finite element stress analysis. With four t/d ratios, three included angles, and two boundary conditions (fixed and free), there were a total of 24 computer runs. Figure 19a represents the maximum effective stress in the compression-compression-compression octant and the worst of the two boundary conditions, the free case. This point always occurred at the low- pressure face corner for the complete range of t/d ratios and angles. This is attributed to the plugging action creating a higher effective stress at this point. 10,000 Plexiglas G Temp = 70°F 8,000 compressive 6,000 4,000 0.2% Offset Yield Stress, Oy (psi) 2,000 0.1 1 10 100 1,000 10,000 Time of Load Duration, T (hr) Figure 18. Yield stress-time curves for acrylic plastic. 28 (p)(S.F.) (Pressure) (Safety Factor)/Yield Stress Ratio, (p)(S.F.) (Pressure) (Safety Factor)/Yield Stress Ratio, 1.0 o for) ° o>) © iS ° ne) 8.0 an ° I) fo) failure region —— safe region peanana = 0.4 1.2 iaiure \Z I /' zone fal NS N Thickness/Minor Diameter Ratio, t/d (a) Compressive region. 1.6 2.0 failure zone —— (aa failure region —— safe region 0.4 ed Thickness/Minor Diameter Ratio, t/d (b) Tensile region. 1.6 2.0 Figure 19. Design curves for conical acrylic viewports. 29 Figure 19b represents the maximum effective stress in the tension-tension quadrant due to the worst boundary condition, the fixed case. This point always occurred at the center of the low-pressure face for the complete range of the variables. The fixed boundary case was the worst due to the increase in flexural bending. Example To indicate how the design curves may be used, a hypothetical case is considered. Suppose the design requirements are as follows: Maximum operating pressure = 2,000 psi Maximum temperature = 700F Safety factor = 1.5 Length of dive* = 100 hours Included angle = 90° Minor diameter = large as possible To properly design this viewport, the first step is to find out what t/d ratio is necessary. From Figure 18 with an abscissa value of 100 hours, the follow- ing results are obtained: compressive yield stress = 6,000 tensile yield stress 3,000 Proceeding to Figures 19a and 19b,and calculating the ordinate value for each curve compressive boda tee) = NEAIOO) DES = 0.5 : oy 6,000 (p) (S.F.) _ (2,000) (1.5) _ , , 0 3,000 y tensile Using the above ordinate values, the following t/d ratios were taken off the abscissa: from Figure 19a t/d = 0.58 from Figure 19b = t/d = 0.34 * This assumes that there is adequate relaxation time between dives. 30 To safely withstand the operating requirements, a t/d ratio of 0.58 or larger would be used. Since the minor diameter is to be as large as possible and commercial acrylic sheets come in thicknesses up to 4 inches t/d = 0.58 O = ib = Bos = i .d ES SOE 7 6.9 inches The largest minor diameter would be 6.9 inches. The elevation, h, (see Figure 2) of the viewport in the flange is another important aspect of the total viewport design. Using a t/d ratio of 0.58 on the abscissa of Figure 14, a value for the ordinate is obtained from Figure 14 Opie : F lines beumdery) Aah 0.0021 inch per 1,000 psi It is recommended by the authors that the viewport be elevated five times the analytical (free case) displacement i) = B., (6) (002) (6.9) (2,000) = 0.145 inch For safe operation of the viewport, a minimum elevation of 0.15 inch should be used in this particular design to insure adequate distance to the end of the conical surface, again recognizing the fact that any significant elevation above this achieves nothing (stress distribution remains the same) and, in fact, only adds additional flange material (i.e., weight) to the vehicle. Figure 19 provides for the use of a safety factor in designing conical viewports. Safety factors are utilized when a designer cannot control all of the many extraneous variables during fabrication and service that might affect his original design. In view of the in-depth research for both the structure and the material properties contained in this report, and the reproducibility of commercial Plexiglas, the authors feel that a safety factor of 1.5 is more than adequate, if all other design recommendations contained herein are followed. ACKNOWLEDGMENT Mr. John McKay conducted the experimental tests and reduced the experimental data used in this study. 31 Appendix A INVESTIGATION OF A VIEWPORT SEAL DESIGN INTRODUCTION A necessary, although oftentimes neglected, function of a viewport design is that of maintaining waterproof integrity at all times. Even very small leaks would be detrimental in a habitat (long term submergence) because of the large accumulation over a long length of time, not to mention the psychological effects on the occupants. Waterproof integrity is accom- plished by a viewport seal system which must function in two environments: (1) at the surface both before and after dives (low pressure), and (2) at the operating depth (high pressure). The objective of this appendix is to (1) review and evaluate existing methods of sealing submersible (short term submergence) viewports, but with regard to the additional habitat functional requirements, and (2) investigate both experimentally and analytically the most promising of the designs. BACKGROUND To provide waterproof integrity, the first viewport seal designs relied on a lapped-joint seal between the viewport and flange as shown in Figure A-1a. Achievement of the seal is by 80 to 90% contact surface area, which requires surface finishes in the 8 to 32 rms range. This design main- tained a seal at the lapped joint for both low-pressure and high-pressure sealing. Sometimes a flat gasket is used for padding between the retainer ring and viewport. The next design was the gasket seal as shown In Figure A-1b. This particular design relied on a gasket located at the perimeter of the high- pressure face for low-pressure sealing and on the lapped joint for high-pressure sealing. A similar design, as shown in Figure A-1c, had anO-ring located at the corner of the viewport high-pressure face. Here again the low-pressure sealing was accomplished by the O-ring and the high-pressure sealing by the lapped joint. The design shown in Figure A-1d is currently being used for submersibles. Here both high-pressure and low-pressure sealing is accom- plished by the O-ring seal. ; o2 (a) Lapped-joint seal. (b) Gasket seal. (c) O-ring seal no. 1. (d) O-ring seal no. 2. Figure A-1. Current viewport seal designs. 33 Submersible designs are not necessarily acceptable for habitat designs because of the more stringent functional requirement due to creep. It would, therefore, be worthwhile investigating a new seal design which would meet the requirements and still reduce cost without compromising the structural integrity of the viewport. A new design or variation of the seal in Figure A-1d is shown in Figure A-2. It isa conventional circular O-ring in a rectangular groove located on the conical face of the viewport. Note that all five designs discussed require retainer rings. O-ring seal no. 3 O-ring - Parker 2 - 457, N-183-9 nitrile (Buna N) 90 durometer Note - After machining, anneal at 175°F for 22 hours; cool at 5°F per hour. Figure A-2. Recommended viewport seal design. COMPARISON OF DESIGNS The five seal designs described may be compared in three different areas: (1) operation, (2) economy, and (3) structure. Operation Functional requirements for viewport seal systems include sealing under all conditions. As previously observed in the main body of this report, creep occurs in the viewport after a long term submergence. Even after returning to the surface, the viewport does not immediately return to its 34 Original position. Therefore, the first three seal designs, a, b, and c, in Figure A-1, would find it difficult to efficiently seal after the habitat surfaces. Neither the retainer ring, gasket, nor O-ring seal no. 1, respectively, would be able to follow the creep displacement of the viewport; the necessary compres- sion of the components for sealing would be lost. These designs function perfectly for submersibles because their dive times are short, thus minimizing the effects of creep. O-ring seal no. 2 and O-ring seal no. 3 are both able to follow the displacement of the viewport and to maintain the seal. A compressible gasket might be employed at the retainer ring-viewport interface, not as a seal, but as a spring to maintain pressure on the viewport for low-pressure sealing. Operationally, an O-ring that is stretched around the viewport (O-ring seal no. 3) is much easier to work with than the reverse (O-ring seal no. 2). The design in Figure A-2 appears to exhibit operational advantages over the other four designs. The success of such a seal system (Figure A-2) has already been proven under long term tests in the ocean. Jenkins and Reinhart’? tested such a design, a conventional O-ring seal in an angular flange, for 189 days at a depth of 5,900 feet in the Pacific Ocean. The O-ring gland materials were carbon steel, aluminum, and clad steel. The O-ring material was nitrile (Buna N) with a hardness of 90 Shore A durometer. The lubrication was petroleum based, although a silicone grease is recommended to preclude any deleterious effect on the acrylic. Results of the tests indicated: (1) no deterioration of the O-ring material, (2) no significant change in O-ring hard- ness or resilience, (3) no significant amount of water absorption by the O-ring, and (4) no leakage past the seal system. Economy The total cost of a viewport assembly could be reduced if an O-ring seal was used rather than a lapped-joint seal. Lapped-joint seals are extremely expensive due to: (1) manual lapping of viewport, (2) tight tolerances on dimensions, (3) additional machining for smooth surface finishes, and (4) absence of interchangeability of viewports. Even for low-pressure sealing, seal systems a, b, and c in Figure A-1 require tight tolerances to insure that the viewport seats at the correct level in the flange. O-ring seals no. 2 and no. 3 on the other hand, can seal anywhere on the faying surfaces so the tolerances need not be so tight. Between the two designs, no. 2 and no. 3, the latter appears to be the more economical. In initial fabrication, it would be more economical to cut the groove in the easily machined acrylic. Also, if the flange is to be clad for corrosion pur- poses, the process is more economical if it is not necessary to contend with 35 an O-ring groove. Since the flange is usually welded to the steel pressure hull, it would be cheaper to replace a damaged O-ring groove in the viewport as opposed to a damaged one in the flange. ‘ Structure With the design in Figure A-2 exhibiting advantages both operationally and economically, it remains only a question of whether the viewport can structurally withstand the stress concentration due to the presence of the O-ring groove. The earlier seal designs in Figure A-1 did not structurally affect the viewport. Designers probably did not use the design in Figure A-2 because of the unknown effects of the stress concentration; there was no thorough stress analysis until Reference 5 was completed. The possibility does exist with O-ring seal no. 2 in Figure A-1, that the softer acrylic will creep down into the O-ring groove in the flange and cause an extremely high stress concentration that could be catastrophic with long term loading. The O-ring groove in the viewport seal design of Figure A-2 was located near the high-pressure face because the results discussed in the main text of this report indicated that the stress distribution was nearly hydro- static in this area. The rest of this appendix is concerned only with the design in Figure A-2 and the structural response of the viewport to the presence of the O-ring groove. SCOPE OF INVESTIGATION To thoroughly determine the structural effect of the groove on the viewport, both an experimental and analytical phase were conducted on the design shown in Figure A-2. Experimental Phase The viewport used to test the O-ring design was very similar to the t/d = 0.45 viewports used in the main text and is shown in Figure A-3. The temperature during the test was maintained between 65 and 75°F because this would impose a more severe test than a lower temperature. While the pressure was held at 4,000 psi, the displacements at the center of the low- pressure face were recorded until the displacements became asymptotic within the measurement accuracy. The machining procedure, pressure vessel type, and testing procedures are all discussed in Appendix B. 36 ‘sAep QOZ Jae Bull Z 4a1j4e Buls-CE Y1IM YOdMaIA Gp’ = P/} JO UONIPUOD |edISAYd jsa}-3SOq “E-W a4nbl4 sano ls SH Analytical Phase The viewport was simulated using the finite element method to obtain effective stress values at critical points for later analysis. The mesh contained 664 nodal points and 610 elements. Both the free and fixed boundary con- ditions were used in this analysis. RESULTS Experimental The viewport with the O-ring was loaded at 4,000 psi for a total time of 200 days before the creep leveled off. Post-test observations, as shown in Figure A-3, indicated rounding of the corner and presence of cracks. Inspection of the O-ring groove, however, found no crazing or cracks whatso- ever. Figure A-4 presents the absolute displacement at the center of the low-pressure face versus time in days. Creep accounted for 35% of the total displacement at 200 days. Even with this much displacement the O-ring did seal when the dive was completed and the viewport was relaxing. 4,000 psi _ 65° to 75°F SS 0.08 0.06 50 psi/min 0.04 t/d= 0.45 Q@ =90° d =84 in. hy =10!24hin: 0.02 Axial Displacement at Center of Low-Pressure Face, 6, (in.) 0 40 80 120 160 200 Time, T (days) Figure A-4. Displacement response of viewport with recommended seal design. 38 Analytical Figure A-5 shows the resulting contour plot of effective stresses for the viewport with the O-ring. Two stress concentrations are clearly found: (1) at the low-pressure face corner, and (2) at the O-ring groove. Table A-1 summarizes the data at the critical locations and indicates that the maximum effective stress value around the O-ring groove is only about one-half that at the low-pressure face corner. The boundary condition essentially created no difference in the stresses around the O-ring groove. t/d=0.45 a@ =90° boundary = free Gy= (contour number) (0.25) Figure A-5. Effective stress contour plot for 90° angle and free boundary condition with O-ring groove. Table A-1. Effective Stress, 0,, for Viewport With O-Ring Groove Boundary Condition Maximum at O-ring groove (compressive) Center of low-pressure face (max tension) Low-pressure face corner (max compressive) 39 DISCUSSION The experimental results and the analytical results had excellent agreement. Even when the viewport failed at the low-pressure face corner, there was no failure at the O-ring groove. The design recommendations (Figures 18 and 19a) of this report predicted failure for this viewport and, as shown in Figure A-3, failure occurred. It might also be interesting to note the flexural behavior of this viewport as shown in Table A-1. For the free boundary condition, no tensile stresses appeared at the center of the low-pressure face as they did for the fixed boundary case. Although minimum research time negated any investigation of mesh-size influence on the O-ring groove, the results were in agreement with the experimental results. Also, the analytical simulation of the groove was a worst case solution. This means there was no rounding of the corners in the simulated model while in the experimental case the corners had a 1/8-inch radius. FINDING A conventional O-ring design with the groove located in the conical face of the viewport (Figure A-2) satisfactorily fulfills the requirements for viewport sealing in three ways: operationally, economically, and structurally. 40 Appendix B EQUIPMENT AND PROCEDURE FOR EXPERIMENTAL PHASE INTRODUCTION Information relative to the design and operation of equipment for the simulated operational tests on full-scale viewports is contained in this appendix. The testing requirements were as follows: Test at least three sizes of full-scale viewports Pressure range, O to 4,000 psi Temperature range, 35° to 70°F Test time, 1 year 5. Cycling to simulate dives of an undersea vehicle ga ©) I> Due to the size of the viewports and the length of time of the tests, a decision was made to design pressure vessels and a refrigeration system specifically for these tests. As a result, pressure vessels with other capabili- ties would not be tied up for a year. HIGH PRESSURE SYSTEM Pressure Vessels The pressure vessel design was as shown in Figure B-1. This particular design yielded the following advantages: 1. Efficiency—each vessel tested two viewports at one time. 2. Safety—low volume of fluid to negate ‘‘bomb” danger. 3. Economy—materials and fabrication costs were low. A _ Ease of operation—simplicity and compactness of design allows ease in pressure and temperature cycling. 5. True simulation—low-pressure face exposed to constant room temperature while high-pressure face exposed to simulated cycling of temperature and pressure. 41 There were two possible disadvantages to the design: (1) opening and closing of a vessel was time consuming and (2) the O-ring extruded when operated to any reasonable pressure and required replacement after each test. These disadvantages are not severe because the system is for long term operation, not day-to-day operation. A typical pressure vessel consisted of two 32-inch- diameter steel flanges as shown in Figure B-1. The conical surfaces of the flanges were machined to a 32 rms finish. The three vessels had through-hole nominal diameters of 4, 6, and 8 inches. Figure B-1. Typical pressure vessel flanges with an 8-inch minor diameter viewport. High-Pressure Pump The high-pressure pump was an air-driven double-acting piston type rated for 20,000 psi. Tubing made of 1/4-inch-diameter, 316 stainless steel connected the pump to the accumulator. Accumulator The air/water accumulator provided a means of storing energy in the high-pressure system to maintain a constant pressure in the vessels. The vessels were all connected to a common manifold, thus, pressure cycling was identical for all three vessels. 42 REFRIGERATION SYSTEM A refrigeration system was necessary in this study to aid in simulating actual environmental conditions. Located on the pressure vessels were Plate- Coil heat transfer units as shown in Figure B-2a. The single, embossed units were rolled to the vessel diameter to insure adequate heat transfer between the vessels and the units. Two units which were connected together by a jumper hose were necessary for one vessel. A heat transfer cement, Thermon T-3, was used between the vessel and heat transfer units to increase the over- all heat transfer coefficient. Sheets of an elastomeric insulation, Armaflex, were used on top of the heat transfer units as shown in Figure B-2b. Glycol brine was circulated through the refrigeration system. Cooling was provided by a 5-ton-capacity air-cooled chiller. Temperature measurements were taken at both the inlet and outlet of the chiller. INSTRUMENTATION SYSTEM Dial indicators, as shown in Figure B-3, were utilized on each of the six viewport faces to measure absolute displacement of the center of the face with respect to the pressure vessel. Lufkin dial indicators with gradua- tions of 0.001 inch were used. The averages of the displacement readings for the three pairs of viewports in the pressure vessels were calculated and are presented in Figure 7 of this report. Pressure gages with a range of 5,000 psi were the high-pressure measuring system, while a 50-psi gage was used on the no-load part of the cycle. Here it was necessary to maintain a 1 to 5-psi load to keep the view- port in position. VIEWPORTS The seven viewports in this study (including one used in the O-ring test) were machined from a 4-foot by 5-foot by 4-inch cast sheet of Plexiglas G acrylic plastic. The viewports were rough machined and then annealed in accordance with the manufacturer's recommendations’ ! as follows: Section Annealing : Cooling ; Time Thickness Temp (hr) Rate (in.) (Ce (OF/hr) 4 AS 22 5 43 (a) Heat transfer units installed on pressure vessel. | DANGER VESSEL UNDER PRESSURE TEST IN PROGRESS DO NOT DISTURB | (b) Pressure vessel with installed insulation. Figure B-2. Typical pressure vessel with refrigeration system. 44 Upon completion of initial anneal, the viewports were machined to final dimensions and annealed again. The viewports were polished to an optical finish before testing to facilitate post-test observations of possible cracks. Figure B-3. Experimental test setup. All minor diameters of the viewports were designed larger than the penetrations in the pressure vessel to elevate the viewport in its seat in order to avoid the stress concentration at the break of the conical surface. Because the test occurred over such a length of time, it was necessary to insure accuracy of results. Therefore, two viewports of the same dimen- sions were tested at the same time to provide redundancy to the test. TEST PROCEDURE The conical surfaces of the viewports were first coated with Dow Corning No. 4 silicone grease, a lubricant which will not attack the acrylic plastic. After the six viewports were placed in the three vessels, the vessels were filled with tap water. Initially, the pressure was raised to 2,000 psi and quickly returned to O psi. This procedure seated the viewports, squeezed out a majority of the excess grease, and allowed the dial indicators to be zeroed. Subsequently, the pressure and temperature were changed simultaneously 45 from 0 to 4,000 psi and 70 to 35°F, respectively. This portion of the cycle represented the dive of the vehicle. Upon completing the desired loading time of 500 hours, the pressure and temperature were simultaneously changed from 4,000 to O psi and 35 to 70°F, respectively. The displacements at the center of the low-pressure face were taken twice a day except for weekends and holidays. The morning and the after- noon were chosen for the displacement recordings because the minimum and maximum pressures for each working day corresponded with the minimum and maximum temperature fluctuations for the 24 hours of the day. 46 Appendix C FAILURE CRITERION APPLIED TO CONICAL ACRYLIC VIEWPORTS A FAILURE CRITERION This appendix outlines the development of a rational yield-failure theory utilized in the design guidelines of this report. In keeping with the yield-failure definition set forth in this report, two entities must be completely described. First, the location and relative magnitude of the maximum effective stress, o,, must be evaluated; and second, the material yield strength, 0, must be determined. Unfortunately, acrylic exhibits time dependency, consequently, both of the above entities are functions of time, i.e., loading history. Because of the infinite variety of possible loading histories, the general design procedure would be to choose the most unfavorable loading history to account for all other less severe cases. For the special case of acrylic viewports, the most detrimental loading history would be represented by a step loading raised immediately to maximum operational depth and held there for the given duration. Moreover, if acrylic was purely viscoelastic, the stress distribution in the viewport at time T = On is identical to the distribution resulting from an elastic analysis and represents the most unfavorable distribution in terms of the high magnitudes and stress concentrations. As time progresses, stress relaxation occurs relieving and redistributing high stress concentrations. Thus, in terms of the viscoelastic behavior, the elastic solution (T = 0*) may be conservatively used to deter- mine the maximum stress concentrations and assume these magnitudes constant with time rather than decreasing. Actually, the conservative assump- tion of constant viewport stresses is quite accurate due to the phenomenon that the elastic stresses of conical frustums are independent of Young's modulus and only slightly dependent on Poisson's ratio. Hence, in view of the correspondence principle, which relates elastic and viscoelastic solutions, time would have very little effect on the viscoelastic stresses. In light of the above, one can rationally accept the stresses resulting from an elastic analysis (approximated by the finite element technique) as the governing stresses throughout the loading duration, providing there is adequate recovery time between loadings. 47 The effect of time on the material yield stress, Oy, which has been observed to decrease as loading time progresses,is next considered. A graphical relationship of this is presented in Figure 18 in the Design Recom- mendation portion of this report. At this point, a qualitative outline is presented, illustrating the origination of the graph. Consider the typical creep data shown in a of Figure C-1 where each curve represents the strain-time relation for a given level of stress. If a parti- cular time, T,,is selected and corresponding values of stress and strain are plotted as shown in b of Figure C-1, an isochronous stress-strain relationship results. Because time has been held constant, any curvature in the stress- strain relation is due to either yielding or nonlinearity of the material. Curvature arising solely from nonlinearity does not limit the yield-strength of the material as previously defined, however, existing creep data for acrylic does not contain unloading histories and permanent set records, thus render- ing it impossible to separate the effect of nonlinearity from yielding. Therefore, it was conservatively assumed that all curvature was due to yielding and that the yield strength at time T, was defined as the 0.2% offset strain intersect of the isochronous stress-strain relation. Following the above procedure for several different times of interest, a curve can be generated as in c of Figure C-1 which establishes the yield-stress versus load-time relationship. The graph of this relationship for acrylic, displayed in Figure 18, was derived from creep data obtained from three sources. '9.13.14 For some loading durations, creep data were nonexistent for Plexiglas G, but were approximated by engineering judgments of creep data on Plexiglas I-A. A STRESS SINGULARITY AND THE ANALYTICAL TOOL (FINITE ELEMENT) In modeling the viewport by the finite element technique, a preliminary investigation was conducted to determine an adequate number of quadrilateral elements to represent the system. It can be shown by variational theorems that the finite element solution approaches the exact solution as the number of elements approach infinity. Of practical importance, the convergence of the finite element model can be determined by comparing stresses obtained from a coarse mesh with those obtained froma finer mesh. In so doing, it was discovered that about 400 elements was a sufficient number, such that, further increase in the number of elements did not appreciably alter the numerical values of stresses at a given point. This held true throughout the entire cross section except for one very notable 48 point, namely, a minute region in the neighborhood of the low-pressure face corner. At this particular location all stresses were becoming exceedingly higher as the number of elements were increased. Further subdivision in this corner revealed that these stresses were becoming arbitrarily high, suggesting that the corner point stresses would approach infinitely high values if the subdividing process continued. In witness of this phenomenon, Figure C-2 portrays the computed effective stress as a surface over the cross section of a typical viewport in which the corner has been finely subdivided. This representation readily depicts the singular nature of this stress concen- tration as a Dirac spike. It was hypothesized that this spike has physical significance and that it is not simply an error emanating from the numerical technique involved. In support of this contention, it was observed that all experimental viewports exhibited a minute permanent plastic deformation at this corner even for moderate loads. Taking the cue from experimental observations, the corner of the analytical model was ‘‘rounded”’ in an attempt to model the plastic deform- ation occurring in the actual viewport. Furthermore, care was taken to determine what effect various amounts of “rounding” had on the resulting stresses. It was discovered that the percentage of area removed did not alter stress values significantly providing the percentage of deformation (percentage of area removed) was greater than about 0.02%. Figure C-3 illustrates this relationship where it is seen that the relative maximum effective stress stabilizes for deformation percentages above 0.02%, independent of the mesh size. |n view of this, it was hypothesized that in a physical case the viewport would deform approximately 0.02% to relieve the Dirac spike stress concen- tration. This contention was supported by the percentage of deformation measured in the viewport shown in Figure 4 of this report. In summary, the philosophy to be adopted in comparing the analytical and experimental results is that, under initial loading, the actual viewport will undergo minute local yielding which results in deformation at the low-pressure face corner, to relieve itself of the spiked stress concentra- tion. This results in a ‘‘modified”’ structure which, when analyzed, does not display the Dirac spike stress concentration, but rather a realistic stress concentration with some finite value. Thus, from the analytical standpoint, the initial yielding problem is dismissed by beginning the analysis with the modified structure (deformed corner). Accordingly, the failure definition applies to the modified structure for both fixed and free boundary condition. 49 Strain,€ ——$> (a) Creep curves. Stress, 0 ——> at Strain, € ———

(b) Isochronous curves. 0.2% Yield Stress, Oy => Time, log T —— > (c) Yield stress-time curve. Figure C-1. Material property curves for acrylic plastic. 50 t/d= 0.75 @ = 90° boundary = free SS Mol e aT ra i — —— mS \\ LT | STL] “apse Lhe ALATA his EAE Ir ary, Sea LEE, ETN Na ASS ie S Oe aN nce Baw wine toe }—" am | Benes pis a QOSSs : OME EEE ADF oo AA EEE +} ALLA MLL a Tet = VIAL SASS Soe Figure C-2. Effective stress surface plot for sharp corner 51 Effective Stress, D5 t/d = 0.75 a@ =90° boundary = free 0.04 0.08 0.12 0.16 Deformed Area/Total Area Ratio, Ap/Ay7(%) Figure C-3. Effective stress variation at low-pressure face corner. 52 0.20 REFERENCES 1. A. Piccard. Earth, sky and sea. New York, Oxford University Press, 1956. 2. M.R. Snoey and J. D. Stachiw. ‘Windows and transparent hulls for man in hydrospace,”’ in A critical look at marine technology; transactions of the 4th annual MTS conference & exhibit, Washington, D.C., July 8-10, 1968. Washington, D.C., Marine Technology Society, 1968, pp. 419-463. 3. Allied Research Associates, Inc. Technical Report ARA 350-3: Tests of acrylic deep submergence windows under simulated operational conditions, by R. Winter and J. Pozerycki. Concord, Mass., Sept. 1968. (Contract no. N00024-67-C-5351) (AD 848287) 4. Naval Ship Research and Development Center. Technical Report 2944: Evaluation of full-scale DSRV-1 acrylic windows under external pressure, by M. A. Krenzke, M. C. Breiter, and L. N. Gifford. Washington, D.C., Jan. 1969. (AD 849354) 5. Naval Civil Engineering Laboratory. Technical Report R-675: Stress analysis of a conical acrylic viewport, by M. R. Snoey and J. E. Crawford. Port Hueneme, Calif., Apr. 1970. 6. A. Nadai. Theory of flow and fracture of solids, vol. 1. New York, McGraw-Hill, 1950. 7. General Electric Company. Advanced Technology Laboratories. Report No. 61GL181: The effects of biaxial stresses on the deformation and fracture of polymethy!methacrylate, by R. L. Thorkildsen and W. V. Olszewski, Schenectady, N.Y., Apr. 1962. 8. Army Missile Command. Physical Sciences Laboratory. Report No. RR-TR-64-15: Combined stress properties for acrylic tube specimens, by R. E. Ely. Redstone Arsenal, Ala., Aug. 1964. (AD 450662) 9 £E.L.Wilson. ‘Structural analysis of axisymmetric solids,’’ American Institute of Aeronautics and Astronautics, Journal, vol. 3, no. 12, Dec. 1965, pp. 2269-2274. 10. Armed Forces Supply Support Center. Military Handbook MIL-HDBK-17: Plastics for flight vehicles, pt. 2. Transparent glazing materials. Washington, D.C., Aug. 1961. 11. Rohm and Haas Company. Publication no. PL-28n: PLEXIGLAS design and fabrication data, Rev. ed. Philadelphia, Pa., Sept. 1964. 53 12. Naval Civil Engineering Laboratory. Technical Note N-1072: Seal systems in hydrospace, Phase III: Effects of long-term hydrospace exposure on seal system integrity. 189 days at 5900 feet, by J. F. Jenkins dnd F. M. Reinhart. Port Hueneme, Calif., Jan. 1970. 13. J. Marin and Y. Pao. ‘On the accuracy of extrapolated creep-test relations for plexiglas subjected to various stresses,’ American Society of Mechanical Engineers, Transactions, vol. 74, Oct. 1952, pp. 1231-1240. 14. Rohm and Haas Company. PLEXIGLAS handbook for aircraft engineers, 2nd ed. Philadelphia, Pa., 1952, pp. 15-16. 54 LIST OF SYMBOLS Ap. AT Deformed and total areas of viewports, in.2 Diameter of high-pressure face, in. Diameter of low-pressure face, in. Elevation of viewport in flange, in. Pressure applied to viewport, psi Radial coordinate, in. Safety factor Time, indicated units Temperature, °F Thickness of viewport, In. Axial coordinate, in. Included angle of conical viewport, degrees Axial displacement of viewport, in. Strain, in./in. Tangential coordinate, degrees Principal stresses per unit of applied pressure, psi/psi Effective stress per unit of applied pressure, psi/psi Yield stress, psi 55 ay i, Sy iF : = \ 3! aff i ' i i c vs s 5 ! i = S Ve ? f 1 i rt ki { if fi 3 a) it i = is 1 i , x 32h ate ie ' : : i ; i Asta d eras a) ai qe DISTRIBUTION LIST No. of Total Activities Copies 1 20 Defense Documentation Center 1 10 Naval Facilities Engineering Command 10 10 NAVFAC Engineering Field Divisions 9 9 Public Works Centers 1 1 Public Work Center 11 11 RDT&E Liaison Officers at NAVFAC Engineering Field Divisions and Construction Battalion Centers 310 310 NCEL Special Distribution List No. 9 for persons and activities interested in reports on Deep Ocean Studies Sy py eA Ria Oh i Vie etd . ‘ i “ aos i iy Si ules i 1 eel \ 5 ete Dy ped ey I thy . ahi tit , fies A i) | / i" AL i REE yh (Ge H rt 1 gener I I * vn y i an ON) wf ti wes 1 Uap inte Ges ty Vi nant) li ines \ “ AHA ie 1 i I J WA Unt “aBue}} S21 ul odmaiA e Huljenaja 104 sautjapinb se ||aM Se ‘Bull-CG |BUO!UAAUOD e Y1IM Buljeas uO osje UaAIB Ss! UOIWeLUIOJUI UBISap ‘SUOIepUaLU -WO09a1 84} ajajdwWod O| “peo| JapuN UO!leinp pue ainssaid Hulyesado paijyioads e 104 14OdMaIA aijAuoe jeo1u09 e yo uBisap ay} ajqeua YdIyM Savino UBisap yO LW41O} ay} Ul Pa}Uasaid aie sUOI]ep -uawwodas UbIsaq “Ajanijoadsas ’4,G€ 01 OZ Pue ISd OOO’ 0} O ‘a4n}e1adWwWa} pue ainssaid jo BuljaA0 snoaue}jNUWIs Ppapnjou! JY} SUO!}IPUOD jeUO!JeJadO Paye;nNWis JapuN 4JeaA e 10} paisa} d4aM S}1OdMaiA ajeds-||N} XIS “S}|NSad jed!1VAjeue ay} ajepijeA O} PawW4ojsad sem uoljebijsaAu! je}UaWadxa ue ‘anoge ay} Guljayjeueg “UOIeINBIyUOD JuodmaiA yo UO!JOUNY e Se PaulWa}ap 313M syidap Buljeiado yUapuadap-awi} ‘d1jA49e 10} UOIIAIIN9 BiNylej-PyalA ay YIM UONDUN[UODd UI sisAjeue je4njon4js wodmaiA ay} Huls—.) “anbiuyda} uawaja ayiuly ayy Aq pazAjeue a1aM OOZL OF 909 Wo. sajbue papnjoul pue G/*| 03 GZ Woy solje4 (P/}) Ja}aWeIP JOUILU/ssauxd14} $O abued e ‘AjyjeadiyioadS “suoijzeinbijuo0d yuodmaiA yo Ajaien e yo sisAjeue ay} Ul pazijiin pue padojanap SEM UO119}149 ainjiej-PjalA ‘JUapUadap-alW!} e ‘jeEOH siy} aAaIyDe O| “sJOdMaiA 31;As9e jed!1U09 yo uBisap ay} 40} yoeoidde Buliaauibua jeuoljes e Ysijqeisa O} Si 140da4 Si4y} yo asodind ayy S00 LO'SOO'SES’8E 4A || adeys |e1u0D ‘7 Siuodmain a1ysejd a1j)Asoy * 1 OL6L eunr snji-d 2g 989-Y1 euoje> “5 “\ Pue Aaous “y “WW Aq ‘(jeu 4) SLYOdM3IA DITAYOV TVOINOD 4O NOISAG IWHNLONYLS Asoyesoge7 Bursaauibu gg jiaig jeaey palyissejouy ‘abue|s Si! Ul JsOdmain e Buljenaja 404 Sauljapinb se |JaM se ‘Buld-C |BUO!UaAUOD e YyIM Buljeas uO osje UaAIb s} UOI}eWIOJU! UBIsap ‘suo!}epuaw -WO9a1 a4} ajajdwod O| ‘peo| sapuN uO!jeinp pue ainssaid Buljesado paljioads e 104 Jodmain a1jAude jeo1U09 e 4o uBisap ay} ajqeua YdIYM saAino UBisap 4O W4O} au} Ul Pa}Uasaid ase suOI}ep -uaWWosas UbIsaq “Ajaniqoadsa1 ‘4 GE 01 OL pue isd QOO'r 0} O ‘asniesadiua} pue ainssaid jo Bulj9A9 SnoaueljnUWIS Papnjdu! Jey} SUOIIPUOD |eUdIe1adO Paj}e;NWis JapuN sea e 404 pa}say G19M S}JOdMaIA ajeds-||N} XIS “S}jNSas jed1yAjeue ay} azepljeA O} pawW4ojiad sem uoljebiysaAul je.UalWiadxa Ue ‘aAoge ay} BHuljajjeseg “UOI}eINHIyUOS YAOdMaIA JO UO!OUNY e se Paul Wwa}ap 31am syidap Buljeiado Uapuadap-aw)} ‘o1;Au9e JO} UO!IA}IND AINjley-pjal!A ayi YIM UONOUN[UOD Ul sisAjeue je4Nj9N43s JIOdMaIA ay} Buis~, “anbiuysai yUaWala aj!uly ay} Aq pazAjeue aiam 0 0ZL 909 Woy sajbue papnjoul pue G/*| 0} GZ’Q WO4} Soije4 (P/}) JayaWe!p JOU!LU/ssauxd!4i yo abued 2 ‘AjediyinadS “suoieisnbiyuod wOdmain jo Ajaien e jo sisAjeue ay) Ul pazijizn pue padojanap SEM UO!JA} 19 ainjlej-pjalA ‘jUapuadap-auui} e ‘jeOb siyy aAaiyse Of “SJ4OdMaIA d1|A19e jeED!1UOD jo uBisap ay} 10} yoeoidde Bulsaauibua jeuoijes e ysijqeisa 0} si 140da4 siy} Jo asodind ay S00 L0°SOO'SEG’8E 4A 1 adeys jes1u0D “Z syodmatn aiysejd a1Asoy “1 OZ6L 9unr sniti-d 2g 989-Y1 euoie>| ‘5 “WW pue Asous ‘y “WW Aq ‘(jeu 4) SLYOdM3IA DITAYDV TVOINOD 4O NOISSG IVWYNLONYLS Asoyeioge 7 Bulsaauibu gz jinig jeaen palsissejouq ‘aBuels S11 Ul aodmain e Buljenaja 104 sauijapinb se |jam se ‘Buld-E |EUO!UBAUOD e YiIM Buljeas uo osje uaAIB si UOIZeLWWIOJUI UBISap ‘suolZepUuaW -WO9a1 a4} a}a|dWOd O| “peo| JapUN UO!}eINp pue ainssaid Buljesado paljioads e 10} LIodmain o1)Au9e jed1U09 e 40 UBIsap ay} ajqeua YO!YM Savino UBIsap yO WO} ay} Ul Pa}Uasaid aie suO!}ep -uawwooas uBIsaqg “Ajanijoadsas “4oGE 02 OL Pue isd OOO’ 02 CO ‘a4ne1ad Wa} pue ainssaid jo BuljoAo SnoaueyjnWwis papnjou! Jey} SUOI}IPUOD |eUO!}e4adO PaYe;NWis JapuN JeaA e 40} paqsal 818M S}JOdMaiA ajeds-||N} XIS “S}|NSa4 jed!1yAjeue ay} ajyepljeA O} PawsOjsad sem uolebijsanul JejUaWUadxa Ue ‘aAoge ay} Buljajjesey “UO!e1NHiyUOd YOdMaAIA 4O UO!}OUNY e Se PauIWajlap a4aM Syidap Buljesado jUuapuadap-aut} ‘d1;A19e 104 UOIJa}I49 auNpiej-PjalA ayy YIM UO!OUN[UOD Ul sisAjeue jeinjoni3s JJOdmain ayy Buis~, ‘anbiuyoa} Juawaja aziuly a4} Aq pazAjeue aiam 00Z1 0} 909 Woy sajbue papnjou! pue G/*| 01 GZ WO so}ze4 (P/}) Ja}aWeIP 4OUILW/ssauxd14} Jo aBued e ‘Ajjeaiyioads ‘suoieinBiyuoo yodmain yo Ajalse e yo sisAjeue au} Ul pazijizn pue padojanap SEM UO!Ja}149 Binjlesy-pjalA ‘yUapUadap-ali} e ‘jeOH siyi aAaiyoe O| “suodmain d1)As9e |ed1U09 jo uBisap au} 10} yoeoidde BHulsaauibua jeuoljes e ysijqeisa 0} Si J40da4 siyi yO asodind ayt G0O0°'LO'SOO'SES'8E 4A || adeys |e91Uu0D °Z syodmain aisejd a1jAsoy “1 OZ61 aunr sniid 2g 989-H1 euojey *5 ‘pj pue Aaous "y “py Aq ‘(jeul4) SLYOdMAIA DITAYOV TVOINOD 40 NOISAG TWHNLONYLS Asoyesoge 7 Bulsaauibu gz jIAig jeaen paryissejour ‘aBue}} sy ul uodmain e Buljenaja 104 sauijapinb se |]aM se ‘Buls-E JEUO!ZUBAUOD e YIIM Buljeas UO Osje UAAIB Ss! UOIeWAOJUI UBISap ‘sUONepUaLy -WO9aJ 384} a}a;|dWO09 O| “peo| JapuUN UO!}eINp pue ainssaid Huljesado paljioads e 40} TIOdMaIA d1,As0e [eD!1U09 e yO Ubisap ay} ajqeua YdIYM SaAino UBisap yO W4O} ay} Ul PajUasaid ase SUO!ep -ualuwooa4 UBIsag “Ajaniyoadsa “4AgGE 01 OL pue isd QOO’r 0} CO ‘a4n}e19dWa} Pue ainssaid jo Bulj9A9 Snoaue}jNWis papnjoul Jey} SUO!IPUOD |eUOIIeJadO Paje;NWIS JapUN JeaA e 10} paisa} 948M SJJOdMaIA ajeds-||N} XIS “S}]NSa1 je917Ajeue ay} azepljeA O} PatWdOjsad sem uoljebijsanu! je}UatWisadxa Ue ‘aAoge ay} Buljajjesey “UO!esNHIyU0d YOdMaIA JO UO!OUNY e se PaulWwa}ap ajaM syuidap Buljesado jUapuadap-aui} ‘9!;As9e 10} UOIJa}ID auNypiey-PjalA ay} YIM UOIDUN[UOD UI sisAjeue j24njoN4ys WOdmaiA ay Buis~) “anbiuysay Juawaja aqui} ayi Aq pazAjeue a1aM QZ | O} 909 Wo1y sajbue papnjou! pue G/*| 01 GZ’ W044 SOl}e4 (P/}) JayaWeIP JOUILW/ssau%9!4} yo abued 2 ‘Ajjeaiyioads “suoieinbiyuoo yOdmain jo Aqalsen e jo sisAjeue ay} Ul Pazi}izn pue padojanap SEM UO!JA]II9 aunjiej-PjalA ‘Juapuadap-au)} e ‘jeOb siy} aAaIyoe O| “SsJsOdmMaI!A 91|As9e |BDIUOD JO uBisap ay} 10} yoeoidde BulsaauiBua jeuoljes e ysijqeisa 0} S! J40daJ siy) yO asodund ay, S00°L0°SO0'SES'8E 4A || adeys |e1u0D °Z suodmain a1sejd a1)Aioy “1 OZ6L aunt snjji-d 2g 989-Y1 euoje> °D “| pue Aasous -y “W Aq ‘(Jeu! 4) SLYOdM3AIA DITAYDV TWWOINOD 4O NOISAG IWYNLONYLS Asojesoge 7 BursaauiBugz jing jeaen pais!ssejou | fea ’ 8 iiegecr ai!) fh i ie ee oe ries E 4 Vi ra ‘ i me Gi cy i “aBuels Sy! ul Todmain e Buljenaja 404 sauljapinb se |JaM se ‘Buls-CG |BUO!UBAUOD e Y1IM Buljeas uO Osje UAAIB Si UOIZeLUIOJU! UBISap ‘SuO!epUaLY -W0981 84} aJaj}dWOd O| “peo; Japun uOlleinp pue ainssaid Buljesado paiyioads e 10} YOdmain o1jAsoe jed1U09 e 40 UBisap ay} ajqeua YydIYM SaAino UBisap yO W404 ay} Ul Pa}Uasaid aie suOI}eEp -uawuwosas ubisagq “Ajaaijoadsa4 “4QGE 02 OL Pue sd OOO’ 03 GO ‘a4n}e1adwWa} pue ainssaid jo Bulj9A9 snoaue}nWis papnjoul 1eYy} SUOIIPUOD jeUO!}e1adO paje|NWwis JapuUN 4eaA e 404 Pajsay 248M S1OdMaIA ajeds-||N} XIS “Si|/NSa1 jed!yAjeue ay} ayepijen 0} PawWOjsad sem uo!}ebISaAU! jelUawiiadxea ue ‘anoge au} Buljayjeseg “UO!eINBIyUOD YuOdMaIA Jo UO!OUNY e se PaullW4aj}ap a4amM suidap Buljesado Juapuadap-awi} ‘d1;A19e 40} UOIIA8IIN9 auNyley-PjalA ayi YIM UO!OUN[UOD UI sisAjeue je41njons3s aodmaln ay} Buls-, “anbiuyda} JUaWala ajiuly ayy Aq pazAjeue aiam o0Z1L 0} 909 Wo} sajbue papnjsu! pue G/*| 0} GZ’O WO} Solqes (p/}) Ja}aWeIp 1OUIW/ssau}914) Jo abued e ‘AyjeaiyinadS “suoieinbijuod yodmaia jo Ajaien e yo sisAjeue ay} ul paziji3n pue padojanap SEM UO119}149 ainjiej-pjalA ‘JUapuadap-auui e ‘jeOH siy} aAaIYyde OJ “syOdMain d1jAsde |ed!1U0d JO uBisap ay} 10} yoeoidde BulsaauiBbua jeuoljzes e ysijqeisa 0} SI 140da4 siy} yo asodind ayt S00'LO'SO00'SES'SE 4A |1 adeys |e1U0D °Z syuodmain ai1sejd a1Asoy * | OL6L aunr sniji-d 2g 989-41 euoie> “5 “Wy pue Asous -y “W Aq ‘(jeu! 4) SLYOdM3IA DITAYOV TVOINOD 4O NDISSG IWYHNLONYLS Asoyesoge Bulsaauibu J |IAig jeaen palyissejoun “aBuelj Si! Ul Odmain e Buljenaja 10} sauljapinb se |JaM se ‘Buld-C JeEUO!UaAUOD e YIM Buljeas UO Ose Uanib si UOI}eWWIOJU! UBISap ‘sUOIZepUaL -WO9a1 84} aJajdWO9 O| “peo| JapuN uoljeinp pue ainssaid Buljyesado paijigads e 404 WAOdMaIA 21|As9e jed1UO0D e 40 UBISap ay} ajqeua YIM Saninod UBisap yo W404 ay} U! pa}Uasaid a4e sUO!ep -uawiwodal ubisaq “Ajaniioadsas “4oGE 01 OL pue sd QOD’ 0} C ‘a4n}esadwa} pue ainssaid jo Bulj9A9 Snoauej|NUWIs Ppapnjou! 1e4y} SUO!}!|PUOD jeUO!}e1ad0 paye|NWiISs JapUN 4eaA e 10} Pasar 348M SUOUMAIA ajeds-|jN} XIS “S}{NSa1 jedI7Ajeue ay} alepljeA 0} Paw sOjsJad sem uol}ebisanu! Je} Uawisadxa ue ‘anoge ay} Buljajjeueg “UO!eINBIyUOd JAOdMaIA JO UO!OUNY e se paulWalap a4am syidap Buljesado juapuadap-awi} ‘91jAs9e 40} UOIJa}1N9 BiNjley-PyalA ayi YIM UOIIOUN[UOD Ul sisAjeue [e1njon43s OdMaiA ay} Bulsy, ‘anbiuyda} yUaWa}a ayiuly a4} AG pazAjeue asam o0ZL 09 Woy sajbue papnjou! pue G/*| 01 GZ‘O WO} Sones (P/}) JalaWeIP JOUILW/ssaUx914i yo abues @ ‘Ajjeaij!9adS “suoieinBiyu0d yiodma~ain jo Ajaiien e yo sisAjeue ay} U! pazijizn pue padojanap SEM UO!1a}149 ainjlej-pjalA ‘juapuadap-atwi} e ‘jeob siyi analyse O| “SJsOdm~aiA 91jA49e jeD1U0D jo uBisap ay} 40} yoeoidde Bulsaauibua jeuoljes e Ysijqeisa 0} SI 140da1 siy} yo asodind ay] G00'L0'S00'SES'8E JA || adeys jes1Uu0D “Z stuodmatn a138e|d a1)Asoy *1 Palsissejour) OL6L aunt snii d Lg 989-Y1 euoye> *5 “J Pue Asous “y “| Aq ‘(jeul4) SLYOdMAIA DITAYDV TVOINOD 4O NOISAG IWYNLONYLS Asoyesoge7 Bulsaauibuz jiaig jeaeyy | paisissejoun a4aM syidap Buljeiado juspuadap-aw!} ‘91|As9e 40} UOIIA1II9 ain|ie}-PyalA aud YIM UONIUN[UOD UI | | | | | | | | | | a4aM syidap Buljyesado Juapuadap-auui} ‘91|As9e 10} UOII9}119 asN|ie4-PjalA ayi YIM UONDUN{UOD UI | Paljissejour) | ‘aBuely S11 ul YuodmaiA e Burjenaja 10} sauljapinb se ||aM se ‘Buls-Q J2UONUaAUOD e YIM Buljeas uO osje UaAIB SI UOI}eW OU! UBISap ‘suOIepUaW -WO9a1 ay} 3}3;dWOd O}] peo) JapuN UO!einp pue ainssaid Buljesado paijioads e 4104 LIodmaIA 21|As9e Jed1U09 e 4O UBisap ay? ajqeua Yd!IYM Sanino UBISaP JO WO} ay} UI Pa}Uasaid a4e suoNep -uawWwodas ubisaq “Ajaaijoadsas “4oGE 01 OL pue isd OOO'P 01 C ‘asniesadiua} pue ainssaid jo BuljoAd Snoaue}jnWis papnjoul Jey} SUO!1IPUOD jeUOIeJadO Paze|NUWIs JapUN JeaA e JO} paisar 318M S}JOdmain ajeds-||N} xXIS “S}]NSad jed!1yAj}eue ay} ajepljen 0} Pawsojiad sem uoljeBbijsanu! je}uawWiiadxe ue ‘anoge ay} Buljajjeseq “uo!einb1yUu0d YuodmaiA jo uO!}OUNY e se Paullwiayap 009 Woy sajbue papnjou! pue G7") 01 GZ’ WO44 Sones (P/) Ja}aWeIP 10UIW/ssaux914) yo abued e *Ajjediji9adS “suoljzeinBiyuoo ysodmain jo Ajalsen e yo sisAjeue ay} Ul Pazi|iin pue padojanap SEM UOlJ9}149 ainjlej-pjaiA ‘juapuadap-awi} e ‘jeOBb siy} analyse O| “SysOdnain d1|As9e |eD1U09 jo | uBlsap ay} 104 yoeoidde Bulsaauibua jeuoljed e ysijqeysa 0} si 140da1 si} JO asodind ayy | adeys jeo1u0D “7 | S00 LO'GOO0'SES'8E 4A || syuodmaia oijsejd aijAsoy “1 OZ6L eunr snil! “d 2g 989-H1 Bude “5 “Aj Pue Avous *y “WW Aq ‘(jeul4) SLYOdMAIA DITAYOV TIVOINOD 3O NOISSG IWHYNLONYLS Asoyesoge7 Bulsaauibuz jiaig jeaen | ‘aBue}} Si! ul uodmain e Bunenayja 104 sautjapin6 se |JamM se ‘Buls-E |EUO!}UaAUOD e YIM Buljeas uO Osje UAAIB Ss! UON}eWWIO}UI UBIsap ‘suonepuaw -W0981 84} 8}aj|dWOD O}| “peo| JapuN uOlleinp pue ainssaid Buljesado paijioads e 410} OdMaIA 91)As9e [BD1UOD e 4O UBISap ay} ajqeua YO!UM Saino UBlsap jo WO} ay} Ul PalUasaid ase SUOIJep -uawiWwodas uBisaqg “Ajaaijoadsas “4oGE 01 OL pue isd 000‘p 01 0 ‘a4njesadWa} pue ainssaid yo Bulj9A0 Snoaueijnwis papnjou! Jey} SUOI}!PUOD |eUO!e1adO Paye;NUWIs JapUN JeaA e 104 Pa}saz 343M SJJOdMaIA ajeds-|[N} XIS “S}|NSad jed!7Ajeue ay} aJepljen 0} Pawsojiad sem uo!ebiysanu! Je}UuaWIdadxa ue ‘anoge ay2 Buljajjeseg “UO!eINbiyuod WOdMaIA yO UO!OUNY e se paulWwsayap sisAjeue jesnjoniys y4Odmain ay} BuisA, ‘anbiuyoa} JUaWaja ajiuly ayy Aq pazAjeue aiam 90Z1 0} 009 Woy sajbue papnjou! pue G/*| 01 GZO WoO4J Soles (P/i) Ja}aWeIP JOUILW/ssaUyD14} Jo abued e ‘Ayjeaiyi9ads “suoijzeinBiyuo0o yOdmaiA yo Ajalien e 4O SisAjeue au} U! Pazi|izn pue padojanap SEM UO!I9}149 ainjlej-PjalA ‘Juapuadap-awi} e ‘jeoH siyi analyse OF “s}IOdMaIA dI}As9e |eD1U09 JO uBisap ay} 40} yoeoidde Bulsaauibua jeuoljes e ysijqe}sa 0} Ss! 140da4 siyi yo asodind ay, GOO'LO’SOO'SES’8E SA |} adeys |eo1u0D “7 sysodmaia oijsejd aAsoy “1 OZL6L eunr snl -d 2g 989-Y1 Buoje>| “5 “i Pue Aaous “y ‘Wj Aq ‘(jeu 4) SLYOdM3IA DITAYDV TVOINOD 40 NOISSG IWYNLONYLS Asojesoge 7 Buissauibuz jiaig jeaen J bi ee ite ; ; ' it 1 Ne 1 ‘ Seri yo teas ag ‘ ve —. ay i 5] F s i > 5 \ ie ‘ ' os Kia Unclassified Security Classification DOCUMENT CONTROL DATA- R&D (Security classification of title, body of abstract and indexing annotation niust be entered when the overall report is classified) ii ee ae : : Unclassified Naval Civil Engineering Laboratory SIRGROUE Port Hueneme, California 93041 3. REPORT TITLE STRUCTURAL DESIGN OF CONICAL ACRYLIC VIEWPORTS 4. DESCRIPTIVE NOTES (Type of report and Inclusive dates) Final; April 1967—April 1969 8. AUTHOR(S) (Firat name, middle initial, laat name) M. R. Snoey and M. G. Katona 6. REPORT DATE 7a. TOTAL NO. OF PAGES 7b. NO. OF REFS June 1970 57 14 Ba. CONTRACT OR GRANT NO. 94. ORIGINATOR’S REPORT NUMBER(S) -prosectno. YF 38.535.005.01.005 TR-686 9b. OTHER REPORT NO(S) (Any other numbers that may be assigned this report) DISTRIBUTION STATEMENT This document has been approved for public release and sale; its distribution is unlimited. - SUPPLEMENTARY NOTES 12. SPONSORING MILITARY ACTIVITY Naval Facilities Engineering Command Washington, D. C. 20390 ABSTRACT The purpose of this report is to establish a rational engineering approach for the design of conical acrylic viewports. To achieve this goal, a time-dependent, yield-failure criterion was devel- oped and utilized in the analysis of a variety of viewport configurations. Specifically, a range of thickness/minor diameter (t/d) ratios from 0.25 to 1.75 and included angles from 60° to 120° were analyzed by the finite element technique. Using the viewport structural analysis in con- junction with the yield-failure criterion for acrylic, time-dependent operating depths were determined as a function of viewport configuration. Paralleling the above, an experimental investigation was performed to validate the analytical results. Six full-scale viewports were tested for a year under simulated operational conditions that included simultaneous cycling of pressure and temperature, 0 to 4,000 psi and 70 to 35°F, respec- tively. Comparison of analytical and experimental results indicated excellent agreement for the physical location of viewport failure at specified loading histories. Design recommendations are presented in the form of design curves which enable the design of a conical acrylic viewport for a specified operating pressure and duration under load. To com- plete the recommendations, design information is given also on sealing with a conventional O-ring, as well as guidelines for elevating a viewport in its flange. DD Wala S Wats 1 Unclassified S/N 0101- 807-6801 Security Classification Unclassified Security Classification KEY WOROS Acrylic plastic Plexiglas Undersea vehicles Design curves Habitats Viscoelastic Submersibles O-ring seal Finite element Pressure cycling Deep submergence Stress analysis Yield failure criterion Windows Failure theory Hydrostatic pressure tests Pressure vessels Conical viewports DD (2"..1473 (Back) Unclassified (PAGE 2) Security Classification er - “al =