APRIL 1961 VOL. 103 No. 2 THE ROYAL INSTITUTION OF NAVAL ARCHITECTS Quarterly Transactions Eated by CAPTAIN A. D. DUCKWORTH, R.N. ‘Secretary of The Institution PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT By R. N. NEWTON, R.GN.C,* (Member of Council) and H. P. Ravert Read in London at a Meeting of The Royal Institution of Naval Architects on October 6, 1960, Mr. L. Woollard, M.A. (Honorary Vice-President), in the Chair. Summary A limited series of twelve methodically varied three-bladed propellers for high-speed craft ‘was designed to cover pitch ratios from about 1-0 to 2-0 and blade area ratios from about 0-5 to 1-0. The parent propeller was based on the design for a particular high-speed craft. After initial tests of a model of the parent propeller, the leading edges of the blades were modified twice to eliminate face cavitation at the working condition. All model propellers of the series had a diameter of 10 inches and were tested in the Vosper Cavitation Tunnel each at nine cavitation numbers over a wide range of sli The results for the parent model are presented in graph form showing thrust coefficient, torque coefficient and efficiency as function of the rate of advance with cavitation number a8, parameter. The bulk of the results is presented in tabulated form giving thrust coefficients, torque coefficients and efficiencies for constant cavitation numbers at rate of advance intervals AT= 0-05, Representative photographs show the cavitation patterns at various working conditions. The discussion is restricted to a general assessment of what has been achieved with this limited methodical series In the optimum incidence range the drag-lit ratios for cavitating conditions approach the values for non-cavitating conditions at the same lift coeficient. Fully cavitating conditions with cavities emanating at the leading edges of the blades and extending beyond the trailing edges are obtained when the angle of incidence of the equivalent lade section corrected for loaded surface effect is approximately +2-5 deg. and the local cavitation number lower than 0-08 approximately. Under these conditions the drag-lift ratios are higher than for non-cavitating conditions at the same lift coefficients. They also depend on the thickness-chord ratio of the equivalent blade section. ‘A reduction of the drag-ift ratios for fully cavitating conditions may be possible by further modification of the leading edges of the propeller blades. How the data may be used for design purposes is shown in an Appendix. Introduction ‘The design of propellers for high-speed craft according to the now well-established vortex theory‘'~ is complicated by the fact that at the speeds in question the performance charac- teristics depend to a considerable degree on the cavitation numbers at which the propeller blade elements have to work. The theory of supercavitating hydrofoils in two-dimensional flow developed during recent years, has helped the develop- ‘ment, but does not supply answers to all the problems of the propeller designer who is faced with an essentially three- ‘dimensional problem and all its complications. In the past, methodical propeller series have proved invaluable to propeller designers when they adapted the propeller vortex theory originally developed for airscrews to the design of non- ccavitating marine propellers. Likewise, data obtained from cavitation tunnel tests of methodical propeller series should prove invaluable to propeller designers when they adapt the ‘existing theory for non-cavitating propellers to fully or super- ccavitating propellers. The series of thirty 16-in. model pro- pellers reported in Ref. 7 provides very useful information for the medium speed range up to about 40 knots, but does not cover the speed range of higher speed craft. Moreover, it was * Superintendent, Admiralty Experiment Works, Haslar, 4 Senior Hydrodjnamicist, Vosper Limited, Porismouth. aL 93 found from the analysis of tunnel test results for individually designed propellers with different types of blade sections that under cavitating conditions propellers with hollow-faced blade sections have better performance characteristics than propellers with flat-faced blade sections. This finding is in agreement with the theory of super-cavitating hydrofoils in two-dimensional flow. “However, empirical values, in particular of the dragelift ratios, differ considerably from theoretical values. The discrepancies are due to the low effective aspect ratio of the propeller blade elements, neglect in the theory of viscosity effects and resistance increments due to the finite Jeading edge thickness of the blade sections. It can therefore be expected that for the perfection of the design methods for high-speed marine propellers @ great deal can be learned from @ ‘methodical series of propellers based on a parent design of a ‘s00d high-speed propeller and tested over a wide range of slip and cavitation numbers. This was the basis of an Admiralty contract for a limited ‘methodical series of tests of 10-in, model propellers placed with Vosper Limited, Portsmouth. The principal features. of the propellers were ‘arranged with the Superintendent, Admiralty Experiment Works, Haslar, with whom close liaison was ‘maintained during the progress of the work. ‘The series consists of twelve 10-in. diameter, three bladed, methodically varied, geometrically similar model propellersPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT based on a parent design for given design conditions. The leading edges of the parent model propeller were modified twi in order to achieve freedom from face cavitation and hi efficiency for the operating conditions. The total number of ‘models tested therefore was fourteen. ‘The working section of the Vosper cavitation tunnel is 19-7 x 19-7 x 86-5 in. with rounded corners, the cross-sectional area being 383 sq. in. ‘The maximum water speed obtainable in the ‘working section is approximately 41 ft/sec. Model Propellers The series consists of twelve propellers based on a parent design with the following main characteristics. Optimum distribution of circulation.* Face pitch ratio, P/D = 1-2 = constant from root to tip. Blade area ratio, BLA.R. ~ 0-750. Blade thickness ratio, B.T-R, = 0-06 with non-linear thickness distribution. Number of blades: Three, ‘Type of section: Shape of face according to NACA mean line @ = 1-0 with quasi-lliptic thickness form super- imposed. Tt was arranged that the propellers of the series should have the same camber ratio distribution as the parent propeller. The combination of face pitch ratios and blade area ratios originally decided upon were as tabulated below. ih ao a | ve | 2 72] 16 | 20 i 20 (Parent propeller in double framed window) Tt was expected that owing to lack of data and experience with this type of propeller it would be necessary to reconsider the section shape after the tests had commenced, PROPELLERS OF SERIES / /) esgraencane {3 Saas | [eee lee) le Feo. Following initial tests of the parent model propeller the leading edges of the blades were modified twice to achieve freedom from face cavitation and a high efficiency at the design ‘operating, conditions, Details of the modifications are shown, in Fig. 1. As at that time drawing office work and pattern making for the other model propellers had progressed so far that adherence to the originally agreed blade area ratios and, * This means that a propeller for given thrust or power has mini rum loss of energy and highest efictency..-The optimum condition [Satisfied when the vortex sheet, after an initial ited deformation, rove auialy Backward a ign screw surface pitch ratios would have meant repetition of work and con siderable delay it was decided to apply these modifications to all the propellers of the series as originally designed with the result that the blade area ratios became smaller and the face pitch ratios slightly higher. The face pitch ratios for the blade sections at 0-7 radius and the blade area ratios of the propellers as tested are tabulated below: AES == 0-48 fer ree “Ts | 1-66 0:95 108 re | 208 (Parent propeller in doubleframed window) ‘The propellers are designated by a letter and three numbers, The letter indicates the type of propeller, the first number indicates the number of blades, the second number the blade area ratio and the third number the pitch ratio. ‘Thus the parent propeller is given the designation A3/71/125 ‘The blade outlines for the propellers of the different blade area ratios are shown in Fig. 2. Representative blade sections of the propellers with blade area ratio 0-71 are shown in Fig. 3. ‘The blade section offsets measured from the original face pitch lines as indicated for the section at 0-7 radius in Fig. 3 are tabulated in Tables A, B and C for the blade area ratios 0-48, (0-71 and 0-95 respectively. All model propellers were manufactured by Vospers, employ- ing the point drilling method to the following tolerances: Mean pitch + 4 per cent Local pitch 2 4 per cent ‘Thickness = 10/1,000 in. on inner radi = 5/1,600 in. on outer radii the transition being at about r ~ 0-7R, ‘The material used for the models was manganese bronze. ‘The surfaces were polished to a high quality smooth finish, Test Procedure All tests were carried out with propeller shaft and dynamo- eter on the upstream side of the tunnel working section. This arrangement is physically more realist than the usually adopted downstream arrangement. "Each propeller was tested at nine cavitation numbers @ = 0:25, 0:30, 0°40, 0°50, 0°60, 0°75, 1°00, 2:50 and the cavitation number which corresponds to atmospheric pressure in the dame of the tunnel, ¢ = 5:5. For the majority of tests the undisturbed water speed in the tunnel working section was 19 ft/sec. A few tests were repeated at 15 tse. to extend the slip range covered, the upper limit ofthe slip being governed by the manima of thrust and torque ‘which can be measured by the dynamometer, With a water speed of 19 f/se, the Reynolds numbers for the equivalent blade sections at 0-7 radius ranged from about 7-1 > 108 for the narrow-bladed propeller A3/48/208 at J = 2-0, 10 about 4:5 » 10° for the wide-bladed propeller A3/95/104 at, J= 0-50. “The corresponding. limiting, values for the water speed of 15 ffs. are about 5-6 % 10° and 3-6 x 108. The effective Reynolds numbers are higher because of the inherent turbulence inthe water stream. The exact amount by which the quoted Reynolds numbers have to be multiplied, however, is rot known because the turbulence factor ofthe cavitation tunnel hhas not as yet been determined, The effective values are, how orPERFORMANCE DATA OF PROPELLERS FOR 2010" ee Fic, 3.—BLADE SECTIONS OF MODEL PROPELLERS WITH BAR. 0°71 fever, well above the generally accepted critical value for tran sition from laminar to turbulent flow on propeller blades. ‘The air content of the tunnel water was kept constant through- ‘out the series of tests at about 50 per cent of saturation Gla, + 0-50). In each test thrust and torque measurements were taken at 20 to 50 speeds of rotation of the shaft. Some tests were carried ‘out two or thee times to check the accuracy with which the results were repeatable. Measurements of idle thrust and torque were taken with bladeless bosses in place of the model propellers covering the same range of test conditions as required for the propellers. ‘These measurements, too, were repeated several times to check the accuracy of the results Cavitation patterns on the propellers were photographed at about 30 to 35 different test conditions for each model. ‘The rates of advance at which face cavitation disappears were ascertained by visual observation. Analysis of Results After correcting for idles, the net thrust and torque values ‘were reduced to non-dimensional coefficients in the usual way, HIGH-SPEED CRAFT 095 BAR. 7 Fig. 2.—OUTLING oF MoDsL PROPELLERS i.e, to thrust coefficient Ky and torque coefficient T pepe ‘The coefficients thus obtained from the measured values were then plotted as function of the rate of advance pe =D with the cavitation number Pye by as parameter. ‘The efficiencies were calculated from the relation using the Ky and Kg values from the curves faired through the ‘measured values. ‘The results have not been corrected for tunnel wall inter- ference. It is felt that these corrections are not as yet known, with sufficient accuracy to apply them to the results. It is con sidered that these corrections are best left to individual users who, in any case, may have their own method of correlating the results of tunnel tests with the performance of the corresponding full-scale propellers. In order to enable a quick estimate of the ‘order of magnitude of the tunnel wall interference, corrections for 10-in. diameter model propellers in the Vosper cavitation tunnel have been calculated according to the method developed by R. McK. Wood and R. G. Harris by application of the axial ‘momentum theory.” The ratios Vp/Vz shown in Fig. 4 as funetion of Kz/J* are calculated according to an approximate formula derived by H. Glauert® from the general equations of Wood and Harris, Throughout the range covered by the graph (Fig. 4) Glauert’s approximate formula gives values of VelVr Which differ by less than 1 per cent from the values obtained by the more exact general equations. 9sPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT om| \seenor waren naw eae Se se BST hen e a bo: TS 3 i] o te Or ry - Of Os SCRE OF lt? Flo, 4 TUNEL Watt comECTION FoR 105N. DAMETER MODEL In assessing the effectiveness of a cavitating propeller, ‘overall performance characteristics are usually compared those of a non-cavitating propeller producing the same thrust ‘or absorbing the same power under otherwise identical con- ditions. The propeller designer who has to compute the blade sections which will produce the lift required with a minimum of resistance is more interested in the performance of individual blade elements. He can Jean a great deal from a further analysis of the overall performance characteristics which will give him the dragulift ratios and the lift coefficients of the equivalent blade sections at say 0-7 radius, as well as the angles of incidence and local cavitation numbers at which these values are obtained, Reduced to this form the data can be compared with the theoretical results for cavitating hydrofoils in two- dimensional flow. ‘The corrections for the three-dimensional propeller flow corresponding to those which are already known for non-cavitating propellers, can be used to assess the design of fully or “supercavitating” propellers. The method used for this part of the analysis was derived by Professor H. W. Lerbs from similar methods used for calculating 28 the performance characteristics of airscrews from the charac- teristics of equivalent blade sections." ‘The blade section at = 0-7 R is chosen as the equivalent section because the mass of water passing through a propeller dise of area zr R? is halved by the circle of radius r= ®. = 0-707 8 assuming uniform inflow velocity. v2 Tt ean also be argued that the section at 0:7 R eorresponds to the centre section of a fully submerged hydrofoil of finite aspect ratio because the maximum circulation occurs in this region. Application of the method to cavitating propellers involves the assumption that the ideal efficiency of a cavitating propeller in non-viscous fluid is the same as that of a non-cavitating pro- peller with the same distribution of eireulation having the same Giameter and number of blades and producing the same thrust at the same rate of advance. It was found from a preliminary lanalsis that this assumption is justified. This was confirmed by Professor Lerbs(") and by Tachmingji and Morgan. "Yhe values of thrust coeficent and efficiency used for the numerical analysis were read from the curves faired through the measured values. Results of Tests Graphic presentation of the results, although by far the most instructive, would require a great number of relatively large diagrams to be of value to users who wish to read off numerical values of the characteristics with a degree of accuracy acceptable for engineering applications. Tt was therefore decided that the results should be tabulated for rate of advance intervals AJ = 0-05. This form of presentation combines accuracy with ‘compactness and leaves the user the choice of the independent variable whenever he requires a graphic presentation. Photographic reductions of sample plots for the parent pro- peller showing thrust and torque coefficients as function of the 026) Siecle (cal coal a a7 0808 18 RATE OF ADANCE — 3 Fio, $.-RESULTS OF CAVITATION TUNNEL TESTS OF MODEL PROPELLER A3/7I/125 96PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT LA aan > che 1 aie] bt gears i fs ist | t= Sass { 2 $28 | toe 2 8333 | agar Onn ra OB EEE OO are LOLEEE aeaeeaai: LeraeeeLl att OF ADVANCE J RATE OF ADVANCE ~ Fig, 7—RISULTS OF CAVITATION TUNNEL TESTS OF MODEL PROPELLER A3/TI/125, ”PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT Fig, 8—CaviTaTiOn PATTERNS ON MODEL PROPELLER A3/TI/125PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT Tate of advance with cavitation number as parameter are repro- ‘duced in Figs. 5 and 6 respectively, These graphs give an indi cation of the amount of scatter of the test points. The efficiency values of the parent propeller calculated from the faired curves, are shown in Fig. 7. ‘The results are compiled in Tables T to XI ‘The values of the thrust and torque coefficients in these tables were read from the original curves faired through the test points which had been plotted on graphs large enough to read numerical ‘values to four significant figures. The efficiencies were calculated. ‘A few representative photographs are reproduced in Fig. 8. ‘The left-hand column shows the cavitation patterns on the ‘model of the parent propeller A3/71/125 at constant rate of advance J = 0-8 for cavitation numbers fram 0-25 to 0-75. ‘The centre column and the right-hand column of Fig. 8 show cavitation patterns for the same model at constant rates of advance J=0-9 and J= 1-0 covering the same range of cavitation numbers. The rates of advance at which face cavitation disappears for the model propellers with blade area ratio 0-71 are shown in second modification to the leading edges of the blades are compared with the design values required for the full-scale parent propeller. Considering that the Reynolds number of the full-scale pro- peller is about sixteen times that of the model, the agreement between measured and design values is considered reasonable, This confirms that the propeller vortex theory as developed for non-cavitating propellers is applicable to cavitating propellers. The performance of the parent model propeller at other working conditions can be assessed from the thrust and torque coefficient curves in Figs. 5 and 6 and in particular from the most striking Feature, in contrast to propellers with Jess cambered or flat- faced blade sections, is that, with increasing slip, the efficiencies for cavitating conditions approach the values for non-cavitating. propellers. This applies to the efficiency curves of all the propellers in the series. As mentioned before, it can be assumed that the ideal efficiency ‘of cavitating propellers is given by the same function as for non- cavitating propellers provided the radial distribution of cit culation, the number of blades, the rate of advance and the thrust Toading coefficient are ‘the same. This means that T T dferences in performance between cavitating and non-cavitating | propellers with the same ideal efficiency and characteristics as ol t oe mentioned are due only to differences in the drag-lift ratios, the products of lift coefficient times chord length being the same. 20) ‘Therefore the performance of the methodical series propellers | | can be assessed in a more general way by comparing the drag 1 peta | Tift ratios for various cavitating conditions with those for non- » —— | cavilating conditions, the independent variable being the lift $ T coeficient of the equivalent blade section at 0-7 radius. Biol tet | | . : fread) acct | 20, ic |__| —_—j I | : an saminos Saal \ \» asrmas ea | sar766 tol oie | ola! onl a © 16 20 30 35 a CANTATION NUMBER = O° i Fig, 9.—ONSERVED DISAPPEARANCE OF, FACE CAVIFATION FOR MODEL, @ PROPULLIRS OF BAUR OTE Eon] Interpretation of Results 2 ocd Owing to the large number of variables involved, detailed 3 interpretation of the result is perhaps best left tothe user. The following discussion is therefore restricted to a more general assesoment of what has been achieved with this limited methodical series. in the folowing table the performance data measured at the working condition of the parent model A3/71/125 after the ot 7 ° 5 0 a0 ‘Yog means ot | Deon aan ug for uae eT COMMISETT OF MANE SECTION A TTS eat ENT OE mane Kaos | ot Fie ii BARS OA AS TUNCTON OF DADE LESAN FE =o — | 0-007 ‘Comments 70K 0°7 Rabie 1066 | 0-708 In Fig. 10 minimum values of drag-ift ratios forthe propellers eoper = 0-104 « extrapolated to full scale (allowing for difference in frictional losses) = 0-09 99 with blade area ratio 0-71 at various local cavitation numbers are compared with values for non-cavitating conditions of the same propellers using the lift coefficient of the equivalent blade section as basis of comparison. The dragiift ratio values forPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT en aE ° INN - = | | 0 See) 296 | ag ng _| a L 4 aS a dasa Sarai —a57 oF -GE OF Dea, CATON NUMBER AT 07 RADIUS Fic. 11.—Mrsiaint DRAG-LIFE_ RATIOS OF METHODICAL” SERIES. PRO. ‘peviers Wii BARR. 0:71 As FUNCTION OF LOCAL CANITATION Nos eo prac ur RATIO“€ oat 8 8 LUFT COEFFICIENTS OF EQUNALENT BLADE SECTIONS, ° 5 the © (ap.2) curves for the various cavitation numbers at which the propellers of blade area ratio 0-71 were tested are plotted in Fig. 11. This graph shows an even more pronounced trend of draglift ratio with pitch ratio. These differences in the + 2:5 deg. the cavitation patterns on the models are ‘completely stable, in particular if the cavities extend well beyond the trailing edge, a condition which nowadays is referred to as ‘supercavitation, but which is perhaps more correctly described a fully cavitating. ‘The lift coefficients and drag-lift ratios of the equivalent blade sections at a9.7 = + 2:5 deg. for all propellers of the series are plotted in Fig. 12 as function of the local cavitation number 9.1. Fully cavitating conditions with cavities extending beyond the trailing edges of the blades are obtained when 9.7 < 008 approximately. The scatter of the lift coefficient values due to small differences in face camber ratio caused by the leading edge modifications, inaccuracies of model manufacture and experimental errors is such that one can draw a reasonably fitting mean curve. This proves the theoretical prediction that, for fully cavitating conditions, the lift coefficient of a blade section depends only on cavitation number, angle of incidence and shape of the face and not on the basic thickness form of the section. The values of the dragclift ratios, however, show a definite trend with the thickness-chord ratio ofthe blade sections. This isin conflict with theory and may be due to the fact that it is physically impossible to use infinitely thin leading edges as stipulated in theory. ‘The possibilities of fully or supercavitating propellers have |. Rees oan |- BERRI wn, BAROMAT 5254 oo _| \ ‘VALUES FOR, — NoNSewrrarinG PROPELLERS fe on be H | ben,| ie ~ a 3TF t O.0a]LARGE BLADE AREA RAG LFT RATIOS ta ao Grp, ox 1° a7 9 ar LOCAL CAVITATION NUMBERS AT.O7 RADIUS Flo. 12.-—Lirr COHFICIENTS AND DRAG-LIFT RATIO8 OF EQUIVALENT [BLADE SECTIONS (X'— 0:7) At p77 =" 2-8" FOR PROPELLERS OF me = ob the cavitating propeller with the Jowest pitch ratio are only slightly higher than the values for the non-cavitating propellers at the same lift coefficients. It appears, however, that the drag- lift ratios for cavitating propellers increase with increasing pitch ratio. To investigate this trend further, the faired envelopes of ° on O5, 330 LUFT COEFFICIENTS OF BLADE SECTIONS AT ‘GP RASS" Co, Fig. 13.~DRAG-LIFT RATIOS OF FULLY CAVITATING PROPELLERS AS FUNCTION OF BIADE FLEMENT LIFT COEFRICIENTS FOR O°7 RADIUS ~ 100PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT been given considerable publicity recently, eg. in Refs. 11 and 12, and an attempt has been made to ascertain how this series of propellers under fully cavitating conditions compares with non- cavitating propellers and with other fully cavitating propellers for which data are available. To this end, the dragclift ratios of the methodical series propellers with blade area ratio 0-71 ‘at ap.7 = + 2:5 deg. and fully cavitating conditions are plotted. in Fig. 13 as a function of the lift coefficient of the equivalent biade section. Also plotted are values for the supercavitating D.T.MB, propeller No, 3509 with modified Tulin sections at the design rate of advance J = 1-125 and cavitation numbers ‘@ = 030, 0-35 and 0-40. The values are obtained by applying the method of analysis used for the reduction of data in this Paper, to the published performance data of this propeller! ‘The curve for non-cavitating propellers shown in this graph is the same as the one shown in Fig. 10. The draguift ratios for fully cavitating conditions are higher than the optimum values for cavitating propellers in Fig. 10 because the angles of incidence required to obtain sheet cavitation emanating from the leading ‘edge are higher than is consistent with the requirements for minimum drag. This applies in particular to fully cavitating propellers with narrow blades which require higher face camber ratios and, therefore, higher angles of incidence than fully cavitating propellers with medium width or wide blades. Conclusions ‘The following conclusions can be drawn from the results of the methodical series tests. (1) In contrast to propellers with flat-faced or Jess-cambered blade sections, the efficiency curves for the methodical series propellers under cavitating conditions approach the efficiency curve for the respective non-cavitating, propeller as slip increases 2) In the optimum incidence range the drag-lift ratios for ‘eavitating conditions approach the values for _non- ‘eavitating conditions at the same lift coefficient. For a given local cavitation number o.7 the propeller with the lowest pitch ratio has the lowest drag-lft ratio. (3) Under fully cavitating conditions (cavities emanating at the leading edges of the blades and extending beyond, the trailing edges) which for the methodical series pro- pellers are obtained when ag 7= +2:5° and ¢9.7<0°08, the lift coefficient of the equivalent blade section depends. only on angle of incidence, shape of face and local cavitation number, but not on the basic thickness form of the blade section. This sin agreement withthe theory of supercavitating hydrofoils. (4) The draglift ratios for fully cavitating conditions show a definite trend with the thickness-chord ratio of the equivalent blade section. This conflicts with theory ‘and may be due to the fact that itis physically impossible to use infinitely thin leading edges as stipulated in theory. (5) Compared with the drag-lft ratios of non-cavitating pro- peller blade sections, the dragslift ratios of fully cavitating propeller blade sections are higher. The “differences are small for low lift coefficients, but increase ‘with increasing lift coefficient. This may be due to the fact that, with the type of section used, the high face ‘camber ratios required to produce the lift coefficients necessitate high angles of incidence to achieve sheet cavitation emanating from the leading edge. These angles are higher than is consistent with the require ‘ments for minimum drag. (© A reduction of the angle of incidence required to obtain supercavitating conditions and hence a reduction of the {=| SSSSSSSSESESERINITSS | = |, |,| SESS88SSSeEggegeeseng | 7 |*| SSSSScSSSSSeesssesees ees 55555555 555s 5S 5S 55555 2 |, BESSSSSSSSSLSRISTHS. @ ||| SS85SSSSS2SSSSsese8s, |" | $8838353332523555353 & |, |"| sssscsssessseseseees i |.) SESeEZE89%.e2gezgrsge E |i |.) gegpgrasen neeargegea BEETEeeE osc esosecee ns | 5 AGEN HSOS Sor OE AAA ; |x| SSSSSS28888R082959538 i |. [=| S2SeSS2e088888292g595 7 |[]8$88s05555555555e5s58 E [3 || ssccccssssssssssessss | i|,| EBSESEESESSESSERSSEE. af |i).| SERgS288s3¢es2s222¢3. S ea ee ee arora eee 3 8 4 SESCCCSSSSSOSSSSSSSS _ 1 | ~| BERSSRESSS SBSRSESR EE Be |i BERSESSZAR RASSSES SES |" | "| SEPEESSses°sssesesete &5 Agana n=S68 So0n 2 oaaNe | «| S8SSESSSSRERSRSESASS 2 |: ZSSZRERRSRERSRSSE ASSES { SesgaSEESSEDREEERGS | ga | Seesecenaaeaanesessa. a 22i2dseeesecesceeee” @ aly. SESS TEST TES EESSS EES. | Laaatassees | bceceenne g peeeeoness Seon nn naan ~ |»| SSSRSRS 288228 SRARRE z |. |,| SESRSRSSSSESSSS ESS SSS 3 |) seeeeeadegeasrseneia, & a Wee beeseeseseeeereae. 2|;| SS8SS8SSSSSS9esegsas. |.) S2aS8BsEssseseage ges 2 -jeesesssesccssessesse” 2 sscslecsesscessesese 1 | «| SS8RBSRRES SERAERRSES 1 |«| SBSSSSSSSs SS EsSR5 Eas 2228655055 Sa005G0r=5 | peer eee Sos Soon ssnnk Senenaces gsenan Base Q giguanaresosenanengsg & peeebebess SSs505ce05 S | escceessss sasccesces 105PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE C Orrsets FoR BLADE SecTIONS OF MODEL PROPELLERS WITit B.A.R. 0:95 09 Radon s S=05 Radin 2s ] x= 07 Radios 503 Rava bs 35 “| g52828 00 e888889808858 SeSEEEUSOSESRE2E=253. 0. 0. 0-0 0 0. 0. 0. 0. 0. 0. O 0. 0. 0. 0. 0. lo 0 0. le SSESS0SSER BER eeSzEae E2Q2SSSRRURRARERSRERE Sbacasscascoasesecese BSSSRG See ee eek seezeg SS S3ShSGER FERS RSZERE SSARERRSRRSR SS RZASSE isebaccesececaees sess E2SSSNFTITSTTAS SES sesscassssssesaeece® SERRCZEROE RRSERSRSE 98e 2 vansses see eee eee ESE ERIS EE22 106PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE I ‘Mopet. Propenter A3/48/105 2eg8063- || SeEeeegeeee2- : geozas: z| S883; 22235 zg 2/7) S32588 = 7) SEESEEEEESE) a|? *\a| #BEREGE. *|.| S$Bgzpegguae. Ta stenatal |-| saepeneaezest - eee |_| _ ae $35) SARRSRES Sz ,|4) Hee! |*) SEsezeseae | ‘el |) | gene. eeest (+) dadeeeesene’ - | #988. . | $egagggens. -| sgazegseeacey. {=| 8: , |+| SSSegg888. | SEEB8B8889228 * | EB. | sa 2) SERIE ~ | 38888 || sepersagss .| snsserengneset 107PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE 11 Mopet Prore.ter A3/48/126 ono SRRSERERY SESSEEEESEEEEE | ageSEnaaeaaegs. gerseansseranas 108 ereggggserad RESALES: | + | S3888a2. «| SERBSESSSEE58. | S2ER2EESS. | 8835833 | 285dseeeseess | SSesseeese | | | BeEeeeg ,| SSSSRSERES a eegoe ee 2) s55555' | S8e8g88esses: —\2 SES855 3 eeecees | oe ieeoass SESSEE g Sgr e see es | a 7 | & «| BREESE2. a) SEGRSSS525832., t|s| BRSSBREREE. pggeses’ . | seesaneseran’ | ee i eau B8898—, | S885SS2esas. 7 | gees: | SERSSSSSS22 : 27) SH2! 2 |*| SSB5BEEEss! =|? a nae ji : gegs 2 |_| gesegmmegas $ z| 53828. «| BRSSREES288. |PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE Mt Monet. Prowse A3/48/167 SRSSETETES. EEESDREZERENEE | BASSSLRRASELESE. SSQRLARFPRISS' S38SS8SS28. e075 Ka geaasseraasnes: TNS. Ke Kr ~ | sxenenaeced ~ | apepeeeene a ~ | seuszromaeuens? ggeeess i eeteatichiay BE conan * | seeps. *|2| gusnsesennsangeys, |*) .| eappyeasanyaezeyya. aageans gggesesaaangeansses gggsenansageangs 109 2PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE IV Mopet. Proventen A3/48/208 - | PRESS SSSRERRANE SSSA. pgvagagmonnenganess SESReCESeRERzEa IEE | |Ssessssssssessesie5 gnengeansgerszages]e= - |SSSSSZBSR2988. aoe EN 8 EESESEERSURESEIED EOSSSEREEE || eeseueeeeeees | « /*] SESSSSESBSCRESECEESUESERE | 2 |u| STSESRSRREESESEEIERS a|" |Ssescsscszess a |" |Sssssssssesecsccccssosces |f|"| sssessessscsosoasess *) g| ESRABSSSS282. * |p] SSSESERESRSSSRARUENZRDORS. |; | BRERESERARERROREEITS. |escesssaascee Soasssesaccuccsssessscess |°| | Ssccessssoossesaaaas ~ | eeangserggassé . |ss2zansneyaugeerszanssecad gngeguseerensnesesaak 8 8 gegneserssasss RSSSBRSPRLBBRRSSLIE enananeyaneserseanesecaadPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE V Mopet, Prope.ter A3/71/105 S2R2S390828 | BRSBERRESER. SSRREBSEES. : z| 32282885828, : z| 38238992952. : 2525289828, ~ | Segegeeesees : .| gengseccaad : e| 58888288. : 22898828828. : ESESR2893. + | SS85R88- POSUGEEUERES- «| SReeBeea8— *) | sasgeas. *) «| sxemmeaszees. "| gpeseenne. erasassé BgeRegeassses eeeggeses5 aPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE VI Mopet, Proveuter A3/71/125 ono ] | ‘| ESSSESENeSESs. » | G8828REES5S. \ 3888 S5ebseseses | SESEREE222528 _ |<} SesgSS3eeee35 | #)*| EeSa222255 | | g| | seeseeseeeces z Sosscaccccs | Seeeessesageg. |" |. eangegezsense, =|’ |, saseseneeys. | ebinett ees | | Sesesscesss | || Ssegsgseranen’ |. | sesgerananeé | | seedencsssties |” | SszzesSeeere | | : fe Jeet Saee + | S888S8ES2RES8s. -| S82ERZ8Sa9E. | scSSSSS5ES5222 | 33sseceeee3 | ee] 2] | Sescces «| SESUUERREEES "| «| gepeseszags. egggeecad ~| S®eeegagseranss | gagseranengy |-—-_-—- —_—----— poo ——— —- - —— | | segezga. “| 2) SRESSESSESS225. "ls SRREEESE SS | Seeesee $3355SSe5e550 S86S555s50° | eaggsend ~| Sexggaggseranss gggsesenasa 12PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE VIL Mops. Proretter A3/71/166 RRSESZREES B85. : seagengpangyenss : B2e8889S5358. : : “SERENE x | REERSESESRE288. -| stemranerons ||.) gguseennmeoncnsent | |. oe N, eg082055. * | mpasapegeengye. ‘ oe oo oe 3PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE VII Moet, Proprvter A3/71/206 SERRSESSESELARAREESR. eeseeeeetuenTEG eeananeyanesseresassse® ePOgeHeG E0228 a SESRSERER EASE ABER EEBESS. cegteancesent || someeuresneaeregoe EE! -gagenasggernenenenee- : : + | ERSSA SSR GESESSSESROESE., : x | SESBSR AR SRESEaE ARES. : scgnserescaognsennnyacs . sogononoreazanssscat aPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE IX Mopet. Proreter A3/95/104 SRSESRE, nsgersagsess aggengagags]enS 2) BESERSSE : g|*| SESee8es | 2/7 a eergegess 7 «| Sase98. . - | SS982g0R98. - . |-| sseseneezz | 8888. HITE . | segasgesenas. "\a| 8882. "ls |, | agaggugeseegs. 5PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE X ‘Mopet, PROPELLER A3/95/124 = os Kr BESSESESATE. 2) ERSEZERSESREESEES- | eeseazesscadl i‘ PERRRESLLANSE | 28389829. | SeS8RResBEsesess. SESESECEESES. | BBBS8888 | | #| SESSSEEESSS28882, [2 | «| FRESHERS, i i: s| S8882893. REGSEREESECCRRSE. [1 | 2 | SERRGEREEEIE. ” enggeser? ASLLLSLRRSSSLAASS BRRRSSLLAARA: | 888383. She - | Seggdeessagesesagy. | S88822) | 3B282 | , 7) SSSScSSSedS2ee82 gasses | Seese 2 Beesecssossssesess Raggeg, SESREESIEE. a | S2228RSRRRP95SE95. SERLSLRRSSLLALA gaggerseasssecaasns 16PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE XI Mopet PRopELter A3/95/165 aa0w SRERSESESIEES: 09-0530 0-053 00838 8:0547 SESREESEIERE RRSSeraness yey SSESESESISTRT. eee ea SEEREADERERE. gevansagegnses seers | |. Regegassecanssseenseeres erauagssoranaensyansserss SESS 2SLES88S2. 2gea3 Eugeseeeese ZELZSGSESELERSIEI ERA | BORSSELNAARASTRASBELE BSSRESSSEELS ETE: SSSEERRARESSSESSS. egegesnaece: 29 SESEOSSE SE RSSSRY 535° angsenanaageanesece’ 47¥EKFURMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT TABLE XII MopEt. PROPELLER A3/95/206 SURSSIISE SEES SEE SERS SE. gpReeeenreeenygze | SE8825) SRR RanEeSeSEzS© enansneeausserszagssec’ SSQSESSEEEEEES CERLSARLGR LESS RSG BRS: ZVRSSSS SReLaeseanses SERSEREDSUERE. geangsarsagsed a0 SSSSESERARSALSSZENSE. SESEEEIESS 25 | gegugnageny | $PSRRSSSRC SS SCNAKZA ESS IS. SUSBESBSSSISERSESEESEEEE | SERSGESSRS SESSEREES. egngnssecs S2ESEESS SESS EER SERE SEE nengegusserssse: SSSRARRERRES SESsES. gserRa8 SSSn88.PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT DISCUSSION Mr. J. H. B. Chapman, C.B., R.CN.C. (Member of Council): As Mr. Newton has told us, the methodical series of propeller tests described by the two authors was sponsored by the Admiralty, but nevertheless T feel I should express the thanks of all of us to them for producing factual data whereby designers can select suitable propellers for high-speed craft. The informa- tion given in the paper, indeed, fulfils a long-felt need. Having said this, however, it is as well that I should call attention to the fact that all the information pertains to the performance of propellers in axial flow under the ideal con- ditions in a cavitation tunnel. Obviously the results obtained ‘with such experiments are not directly applicable to the full-scale ship for several reasons, not the least of which is the fact that in the full-scale craft the propellers operate in inclined flow. Consequently, it cannot be expected that the ship propellers could give the efficiency assessed from the model series. Nevertheless, itis interesting to record that the propellers for the Brave class fast patrol boats lie within the series described in the paper. They do, in fact, very closely resemble the parent propeller design which is “ringed” in column 2 of page 94, i.e. the propellers with a blade area ratio of 0-71 and a face itch ratio of 1°25. Tsay this is interesting because it is now an established fact that the Brave class fast patrol boats are attaining the highest speeds which, to my knowledge, have so far been achieved by planing forms. You may recall that a descrip- tion of the general design particulars of these craft was given in the recent RILN.A. paper by Revans and Gentry.* This paper ‘quotes a speed exceeding 50 knots. Its also perhaps well worth while recording that after several months of operation of these craft the propellers exhibit only slight signs of erosion, and this, is very gratifying, having regard to the very heavy cavitation conditions under which they operate. One final point I would like to make is that the authors have ‘gone to considerable trouble to include in the paper an example ‘of how to apply the information given in the series to design the propellers for a hypothetical craft. Tam sure that all students, and also perhaps any authority who in future may use the information in the new series, will greatly appreciate this. Tt is of great use in any paper and we are grateful to the authors for hhaving included it. Mr. A. Silverleaf, B.Sc. (Member): This paper has at least fout major virtues. First, as Mr. Chapman has pointed out, it presents a great deal of extremely useful performance data covering a very wide range of propeller operating conditions. Second, these data are presented in a form which enables us to make practical estimates with comparative ease; this will be particularly important in the near future as ships requiring such propellers become more common. Third, full results of the experiments are given; this, happily, is a growing habit, though itis perhaps a pity that here these Tesults have been tabulated from faired curves. Fourthly, the parent propeller has very 200d performance, showing that all the developments in fully cavitating propellers have not originated on the other side of the Atlantic, and that, indeed, successful results may well have been. achieved earlier in this country. However, the paper has some disappointing features. In dis cussing these, I do not intend any adverse criticism of the authors’ work, but hope only to stimulate and assist future work fon this extremely interesting and important topic. No clear information is given as to the design method originally used for the parent propeller; is it that given in Ref. 12, or, if not, what method was used? What were the precise operating conditions for which the parent model was designed? Presumably they + Sce TRANS. RILNA., 1960, p. 367. ‘were, as quoted in the table on p.99, J = 0-993, ¢ = 0-311, but which was the specified force coefficient? Was ky 0-121 the target, or was it kg 0-027? Clearly both cannot be inde- pendently specified. "Further, how were corrections made for the radially varying, finite ¢ values at which the blade elements operate? ‘The arbitrary, empirical snubbing of the Ieading edge of the parent model to achieve the desired performance is under- standable from practical considerations, but regrettable from a broader view. These modifications threw away the theory, so that it is no longet possible to make a serious comparison between design theory and experiment results. It also leads to minor difficulties for the reader. What is the datum line for definition of the angle of incidence of a blade element, including the equivalent section at 0:7 radius? Is it the nose-tail line of the original blade section before modification, or that of the section as finally modified? Indeed, how is the face pitch ratio of the screw defined” Is it that of a nose-tail line of the original, ‘unmodified section? In planning any methodical series of propellers, there are two alternative approaches. First, we can systematically vary the ‘geometrical parameters, including pitch ratio and blade area ratio. On the other hand, we can do something quite different; ‘we can systematically vary the specified operating conditions. By specifying a set of values for J, 0, and kr (or kg), a series of propellers can be systematically’ designed, for which all the ‘geometrical features are automatically determined. This is the method followed in Ref. 12, though there it was used only to design a series of screws and to compute their performance, no systematic experiments being made. The first procedure—that of simple geometrical variations—is acceptable for a series of moderately loaded, non-cavitating screws, since it has been shown (Ref. 13) that then variations in blade section shape do not have a major effect on efficiency. However, for propellers ‘designed to operate under severe cavitating conditions, the other ‘method—that of varying specified operating conditions—is the ‘only one which shouid be used. Otherwise we are inadequately informed; while we may assume that the parent propeller is the ‘best possible for its specified operating conditions, we do not know that those derived from it by simple geometrical variations (uch as a crude change in pitch ratio) are necessarily the most efficient possible at their own best operating conditions. suspect that the authors’ interpretation of their results in this paper may well be affected by just this doubt. Incidentally, the authors describe their series of propellers as “geometrically similat”; surely this is not correct for a series which are of the same size and which differ in important geometrical features. ‘The authors’ interpretation of their results and their con- clusions lean heavily on Lerbs’ equivalent blade section pro- cedure, The title of Lerbs’ paper (Ref. 10) emphasizes that this procedure is intended primarily for scale effect problems, in articular to correct model propeller data to full-scale con- ditions. To use it to obtain blade section characteristies from. propeller experiment data is possible, but stretches the method beyond its proper limits. Tt is tantamount to using a sledge hammer to crack a nut—and not a very accurately aimed sledge hammer at that, The proper way to obtain section characteristics is from experiments with foils, not propellers. ‘Admittedly the authors were, to a large extent, compelled to use the equivalent blade section analysis, but T suggest they have leaned on it more heavily than is justified. For example, they ‘conclude that lift-drag ratios for fully cavitating conditions vary with thickness-chord ratio and with pitch ratio, contrary to ‘theory; this conclusion may well be considerably influenced by ‘the way in which the methodical series has been derived from 119PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT the parent. Had the group of propellers been derived by designing the best series for a set of different operating conditions, the effects of thickness and of pitch might well have been found to bbe different. ‘Throughout the paper the authors emphasize that their design and analysis procedure is based on the assumption that the ideal efficiency of a cavitating propeller in a non-viscous fluid is, the same as that of a similar noncavitating propeller with the same distribution of circulation, and that differences in per- formance between cavitating and noncavitating propellers are due only to differences in the drag-ift ratios of the blade sections. Further, they maintain that this assumption is confirmed by the work in Refs, 11 and 12 and by their own work. I suggest that these views are basically incorrect. Circulation theory need have no place in the design of supercavitating propellers, as T pointed out in discussing Ref. 12, where it is applied quite illogically. This principle can be’ simply explained. For a fully wetted foil in a non-viscous fluid itis essential to introduce ‘the concept of circulation to explain the generation of lift; there is no need to do so for a foil with a long cavity, since potential theory without circulation will give lift and drag values which agree well with those measured in water tunnel experiments T believe that the tolerable agreement which has been obtained to date for supercavitating propellers between calculated and measured performance is due largely to a fortuitous cancellation of the many irrelevant and incorrect factors thrown into the circulation theory methods presently used. Indeed, I suspect, that potential flow, blade clement theory, with one or two neeessary corrections (as yet not fully determined) would give closer agreement between calculation and measurement for supercavitating propellers. In brief, I suggest that we still ack an adequate design theory for fully’cavitating propellers, and T hhave a strong feeling that it should be based on exact potential flow solutions for fully cavitating foils, as in Ref. 6, applied to a rotating propeller. Reference (13) SuvERLEAR, A., and O”BaueN, T. P.: “Some Effects of Blade ‘Section Shape on Model Serew Performance,” Trans, NECLESS., Vol. Th, 1955, Commander P. DuCane, O-BE,, RN. (Member): tis a ‘matter of some satisfaction to this writer that this paper has been submitted. Especially do we at Vosper appreciate the encouragement and active help received from the Director of ‘Naval Construction of the day (1955) and, of course, from Mr. Newton, It was his predecessor, Dr. Gawn, who encouraged us to go ahead with the purchase and installation of this cavitation tunnel at a time when orders for fast craft were already rather searce. It is an expensive item to purchase and man for a relatively small shipyard, but apart from economic factors it ‘can be considered well worth while, In the long term we feel convinced the outlay will even be justified in economic terms apart from the technological value of the results. As stated by Mr. Newton, when this piece of research was first discussed it was intended that the methodical series should be followed by actual running in a specially designed boat so ‘that correlation could be attempted with Full see. Despite the unquestioned value of methodical series in a cavitation tunnel, there must remain a number of imponderables in forecasting performance at full scale. It does not need to be ‘emphasized that propeller efficiency found from tunnel experi- ‘ments in themselves bear little relation to what can be achieved by the propeller at the end of an inclined shaft under a bull, ‘To mention some of the more obvious differences between the “tunnel” experiment and full scale: (1) The propeller shaft is inclined to the horizontal when running at high speed so that the flow into the propeller disc is far from axial as in the experiment. (2) Not only isthe propeller axis inclined to the flow direction in the vertical plane, but also in the transverse plane due to a cross-flow component arising under most planing Forms in way of propellers ) The drag of propeller shafting, bracketing, rudder, etc, will exercise a considerable influence on the overall propulsive efficiency as well as to some extent the effect of a propeller bracket on the flow into the propeller, ‘more especially in cavitating conditions. (4) For most ofthe above reasons itis, therefore, not possible, even with such knowledge as has been gained up to date, to be entirely precise as to such matters as optimum revolutions for each case. As always in attempting to solve such engineering problems, an element of com- ppromise has to result from considering the various ‘modifying factors in the case. It was most disappointing, therefore, when the full-scale boat experiment had to be abandoned by the Admiralty due to financial considerations, and we dare to hope that one day it will be restored in the programme. However, we decided the matter was of sufficient importance to justify a certain expenditure on an alternative experimental technique which could be adapted for use in our existing, tunnel.” "This consists in locating the propeller on a realistically inclined shaft including brackets and rudder. In this way at Teast a much closer approach to realism was obtained, and although the results have not yet been published we did explore the performance of the selected propeller A3/71/125 arising from the methodical series. The results are interesting and undoubtedly valuable. One somewhat surprising result is that at low values of Ky/J? the efficiency is substantially greater under cavitating conditions than for the axial flow case, In this, rig it was possible, as well as thrust and toraue, to measure the drag of brackets, shafting, and rudder. ‘Now there are some interesting and important points arising from the data presented in this valuable paper on which the authors’ comments would be of interest. Referring to Fig. 15, it is quite clear that optimum efficiency even for cavitating cases arises in conjunction with high piteh ratio, which seems to suggest a prior large diameter and low shaft revolutions. Quite obviously one of the most important operative factors here is the appendage resistance expressed as a percentage of total thrust, because in arriving atthe overall propulsive efficiency it is obvious that the large slow-running propeller at large shaft angle may involve a quite disproportionate reduction in overall efficiency from this cause, This is almost certainly the ease and the problem with which we are faced is to find the optimum. For this reason a methodical series of experiments to evaluate appendage resistance fora range of configurations and propellers is clearly desirable. However, this is not the whole story because if we consider Fig. 14 and refer to Table VIII, for instance, we can find an optimum efficiency at ¢ ~0-25 for P/D ~ 206 propeller A3IT1/206 to be 0-711 at J = 1-75, ‘This optimum point ‘occurs to the right of the “hump” in the Kg curve which we are advised in the text to avoid owing to the danger of bubble cavitation which has a damaging effect on the material on which the “bubbles” collapse. The blade angle of incidence (2) in this cease is O deg. 18 min. and local cavitation number 0°097, We are advised to move farther to the left on the curve (0 a point at J = 1-5, where the efficiency is somewhat lower, but Where sheet cavitation can be anticipated. If we now take the example of the propeller selected as parent in this series and 120PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT which was successful on the Brave class fast patrol boats at more than $0 knots and which, as far as we know, represents the ‘optimum, we find it in Fig. 14 on the curve for P/D = 1-25 and J = 0:98. Here the efficiency is 0-664 and angle of inci- dence 0 deg. $8 min. at local cavitation number (a) = 0-05, Given as the criterion for fully cavitating conditions earlier in the paper is a ~ 24 deg. associated with o ~ 0-08, which envisages a fully developed “blister” of sheet cavitation over the whole biade back, We must conclude, therefore, that the Brave propeller, even at or near full speed, is not really in the super- cavitating range. Referring to the photographs on Fig. 8, there isa good deal of difference between J = 0-9 and J = 1:0 as far as the develop- ‘ment of bubble and sheet cavitation are concerned. Certainly at cruising speeds of, say, 40 knots or even 30 knots, there will be a considerable amount of bubble cavitation. From this it would be expected that the rudder which is downstream would be badly eroded. “However, this is not at all the case on H.M.S. Brave Borderer after many hundreds of hours of high-speed running. This seems rather difficult to explain in view of the suspected liability for the bubble type of cavitation to cause damage. If this type of bubble cavitation does not in fact cause damage, then we could presumably design our propeller to the right of the hump in the Ko curve on Fig. 14 and gain ffciency. Where there is a strong liability for erosion of nacelles, this can be very important in hydrofoil propellers. However, be that as it may, it can fairly be contended that the authors have produced some highly successful propellers both from the point of view of performance and avoidance of erosion, and this is a matter of great significance in posing a number of hypothetical questions as to how matters might be improved ‘The writer would like to put forward a suggestion that a propeller should be tested under a boat at some stage without appendages of any sort, except possibly a rudder. ‘Theoretically it is believed the arrangement is selfstabilizing and there are many stories from operational flatillas during the war of boats returning with propeller brackets broken off from the hull and where no one on board had observed any difference in per- formance or vibration. The appendage drag eliminated by such fan arrangement would be considerable, leading to favourable overall propulsive efficiency. ‘The authors are to be congratulated for producing what will bbe considered a classic on this subject, and the writer of this discussion wishes to acknowledge his appreciation of the extraordinary amount of skill and meticulous work which was involved in the execution and recording of data accomplished by Mr. Rader and his staff Mr. J. H, Bedbrook: I would like to add to Mr. Chapman's comments concerning the limitations of the methodical series data presented in the paper. Some of those present this evening will be aware that separate contributions were presented by the same two authors to the 9th LT.T.C. in Paris, which bear directly upon the effects of inclination of fow on Ky, Ka, and 7, to which Mr. Chapman has referred. ‘Very briefly, experiments were conducted in the Haslar Cavitation Tunnel with a model destroyer propeller to deter- mine the effect of 10 deg. inclination of flow, About the same time experiments were also being conducted in the Vosper tunnel ‘with a model of a propeller for a fast patrol boat, with the shaft inclined 124 deg. Although these two investigations were carried ‘out quite independently they indicated similar trends as regards the effect of inclination of flow on thrust, torque, and efficiency. ‘The Haslar experiments showed that'at cavitation numbers applicable to the full-scale destroyer, about o = 1°5, there is a drop in thrust and efficiency over almost the entire range of ‘advance coefficient. Observations indicated, as expected, that the cavitation pattern on each blade varied considerably with its orientation on the shaft due to the wide variation in angle of attack of the blade sections caused by the shaft inclination, Consequently, at all advance coefficients one or more blades ‘were losing thrust due to either back or face cavitation. The loss in efficiency was by no means small. At advance coefficient 1-0 and ¢ 1-5 it amounted to about 3 per cent, and is therefore ‘worth while accounting for in a new design ‘The Vosper experiments were carried out over a wider range and in particular down to a lower value of cavitation index. ‘These experiments indicated a very similar trend as regards cffciency at the higher cavitation index values, but as the cavita- tion number was reduced so the loss in efficiency reduced. Indeed at values of cavitation number below 0-7 a reverse effect took place, the efficiency increasing and the magnitude of this increase grew as the cavitation index was reduced below 0:7. In this case also the increase in efficiency was not small and well worth accounting for in a new design. Mr. L. A, Persson (Associate-Member): This paper is the first in which a methodical series of model experiments with propellers specially adapted for high-speed craft has been published, and therefore it is of the greatest interest, The ‘varying cavitation index is an extra parameter which the designer of propellers for high-speed craft has to consider, which makes the choice of the main data still more important than for pro- pellers not working in the cavitation range. This paper, there- fore, will help the designer from the very beginning to find the ‘optimum propeller for a particular case. T would also comment on the teference to the very good agreement between the results of the model and full-scale test, about which we should be very satisfied, ‘The other type of fully cavitating propeller is that with wedge- shaped sections. In the K.M.W. cavitation tunnel at Kristine- hhamm some twelve propellers of this type have been tested, and fone of these KaMeWa models has been tested in the Vosper tunnel for comparison. =030 Rr Fo. 16-—T20\/K MW IIL, ‘Tonnes tests, Monet. KMW355-B In Fig, 16 the efficiency of the model propellers A3/48-95, With piteh ratio about 1-04-1-66, has been plotted on a basis, of Ky/J2 There is also a plot of 384-B with wedge-shaped sections, riPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT ‘The B.A.R. of 354-B is only 35 per cent, which, taking the large boss, 29 per cent, into consideration, corresponds to a B.A.R. of about 41 per cent for a propeller having the same boss as the A3 series. The comparison indicates that possibly some gain in efficiency could be obtained at a higher B.A.R. ‘A propeller of the 354-B type has been used for a. small experimental craft of the Royal Swedish Navy, 1,550 shp at 2,000 rpm. The ship speed was around 42 knots. Photographs In Fig. 17, taken at 2,000 rpm, Were taken on board the boat. In my opinion the authors have wisely chosen a fairly con- ventional thickness distribution for their blade sections, obviously aimed at obtaining a reasonable performance through= out the whole speed range. In the non-cavitating condition the blade sections have a higher effective camber ratio than in the fully cavitating condition. This seems to be the reason for the fact that with increasing slip, efficiencies are comparable for the non-cavitating and fully cavitating conditions. It is noted that the authors originally chose the N.A.C.A. Fig, 17. No. 4201 Fig. 18, No. 4213 PID = 1-189) 0, = 0-382 1,500 bhp mee J2,014 rpm—A-2 knots the clear cavity can be seen. Corresponding photograph ina = 1 mean line for the face shape of the blade sections. Since the cavitation tunnel is shown in Fig. 18. The propeller has now had 130 hours of service at shaft speeds of 1,800 rpm or more without showing any sign of erosion, which is a very promising feature of this type of propeller. Therefore, I believe it will be appreciated that there are two main advantages, the possibilities of very high speed and more erosion-free propellers. Finally, I would like to say how grateful we all must be to the British Navy that they continue their tradition to publish methodical series of propellers, this time in association with Messrs. Vosper. Mr. G. C. Cox: One of the most important problems associated with the design of fully cavitating propellers is the avoidance of face cavitation, since otherwise a drastic loss of thrust and efficiency can result. ‘The straightforward solution, which is to increase pitch, can incur a high drag penalty. "At present, only methodical series such as presented this evening can provide the necessary guidance. ‘Since this type of propeller will be fitted to an inclined shaft, it Hs suggested that the next logical step is to obtain thrust, torque, and onset of face cavitation data for the present series ‘of propellers in inclined flow representative of full-scale conditions. Due to the possibility of face cavitation occurring under certain conditions when using propellers in inclined flow which have been designed from the presented data, it is possible that there will be eyelie variations in thrust and torque. Do the authors consider this could lead to serious vibration problems? subsequent modifications had to be made, could they give the reasons for their choice and state the order of improvement gained by making the two modifications. Reference to the photographs shows that atthe design advance coefficient and cavitation number of the parent propeller, the cavitation is of a burbling type as opposed to sheet cavitation, beginning at the leading edge and extending well beyond the trailing edge. Could the authors say whether, in their opinion, this has any detrimental effects as regards performance and cavitation erosion. Mr. W. A. Crago, BSc. (Associate-Member): 1 find it very ‘gratifying to have listened to a paper of such importance as the ‘one read today, which contains data obtained by a commercial concern who were allowed by their customer to publish so fully the details of their experiments. We must be very grateful to the Admiralty and to Vosper Limited for a paper which will be often referred to by those of us who work in the high-speed regime. would like to raise one or two queries. First, is it not time ‘that we stopped talking about manufacturing tolerances in terms of + so many thousandths of an inch? Theoretically the propellers discussed in the present paper could have had a 0-020 in. discontinuity but would still have been within the tolerance quoted. Ido not, of course, suggest that the propellers were as badly made as this, but it does suggest that some other criterion is required and should be quoted, T notice that in the experiments, the shaft and dynamometer 12PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT were upstream of the propeller, whereas more often they are located’ downstream. Does this imply that in attempting t0 ‘compare the results of the present paper with data produced, say, by Gawn and Burrill, that some form of correlation factor has to be employed? If no such factor is required, why should the position of the shafting and dynamometer relative 10 the propeller be changed? would like also to ask the authors to comment on the extra- polation tothe full-scale regime of the results they have obtained ‘when they have no idea what the effective Reynolds numbers of the blade sections are. ‘The data ofthe present paper have been presented in terms of Kz and Kg. Many practitioners in the field of propeller design, employ Bp and 5. Is it not possible to present cavitation ‘phenomena in these terms? ‘Turing now to the example given at the end of the paper, I would like to know how one proceeds from the cavitation number of the fullscale propeller to the equivalent tunnel ‘number. Also it seems to me that one cannot make use of the data in the paper even withthe aid of the worked example unless fone has an idea of the quasi-propulsve coefficient factor. This factor can, according to one authority, become as low 2s 0-7 and is supposed to account for “extraneous” effects not normally taken into account. Are we to understand that if we use the data inthe present paper that this factor ether becomes unneces- sary or reduced? Ifthe factor still exists ands, in fact, as large as [ have suggested, then it seems that experiments aimed at providing design values for it would be of far more value than Any Further detailed tunnel tests on systematic series of propellers Mr. J. F, Leathard, B.Se., Ph.D. (Member): The concept of applying’ circulation ‘theory’ to the calculation of a super- cavitating propeller is intriguing. Mr. Silverleat suggests that such a theory has no application to the problem, but the fact, remains that the paper suggests a method whereby cavitation results may be assembled from model data into a reasonably ‘coherent whole, ‘A simple example of this concerns the “ideal” or “hydro- dynamic” performance of propeller. Under a given set of conditions existing in the wake, the circulation at any radius rust be related to the lift coefficient at the propeller and conse- quently the pitch angle at the corresponding blade section must bbe chosen so that the angle of attack is sufficient to enable this lift coefficient to be developed. Under cavitating conditions, however, its reasonable to assuime that the effective shape of the section is altered, due to the suction cavity on the back and an effect, not unlike that produced by flow curvature, is apparent, whereby the effective angle of attack, and consequently the pitch angle, has to be increased, ‘Thus for a given power to be developed under cavitating con- ditions a relative increase in pitch ratio is required as compared ‘with a non-cavitating screw. Although conclusion (1) suggests that the efficiency of a cavitating propeller of the type tested ‘can be made to approach that of the non-cavitating propeller it must be emphasized that there is always an associated loss of thrust and torque which, however, tend to balance each other and maintain nearly constant efficiency. ‘An analysis of the results of the wide-bladed propeller series presented in a paper to The Institution by the late Dr. Gawn® shows that the optimum thrust coefficient varies little on a given blade area ratio with change in pitch ratio, This feature can produce considerable simplification in design and it would be interesting to analyse further the authors’ results to see whether a similar type of effect exists at a given cavitation number. Itis believed that, if this were found to be the case, the work involved in the authors’ example could possibly be reduced. + Gawn, R. W. Li: “Effect of Pitch and Blade Width on Propeller ‘Trans. LNA. 1953, p. 137. One great merit of the paper lies in the presentation of its data. tis far simpler to read values from a table than to lift them from a comparatively small chart. For the naval architect who ‘wishes to use the data to design cavitating propellers, the paper is full of useful information excellently presented. ‘The Chairman, Mr. L. Woollard, M.A. (Honorary Vice- President): We have had a good paper and a good discussion nit. I particularly commend the paper, and I ask you to give a hearty vote of thanks to the authors ‘Written Discussion Mr. K. C. Barnaby, O.B.E., B.Se. (Honorary Vice-President): ‘The authors of this very informative paper have reserved what appears to be a major difficulty in the design of propellers, ‘working at low cavitation numbers, to the penultimate paragraph of the appendix. This postulates a requirement that, in order to obtain a stable cavitation pattern, itis necessary for the torque coefficient to “plot on or just to the left ofa line connecting the minima (fo the left of the humps) of the torque coefficient curves, as shown in Fig. 14.” It is presumably this restriction that has, Jed to an anomaly in the table for B.A.R. 0-71 on p. 103. This shows an improvement in efficiency by either increasing or decreasing the blade area of the two larger propellers. This sounds absurd, but itis evident that the efficiency of these two propellers could be increased by adopting a slightly lower pitch ratio and a correspondingly larger diameter. Although the appropriate torque coefficients would still agree with the experi mental results, they would plot somewhat to the right of the ‘connecting line on Fig. 14. Perhaps the authors would state hhow sacrosanct they consider this line and what are the penalties incurred by straying into forbidden territory. ‘Some thirty years ago it was necessary to design propellers for an estimated speed of 100 knots and a cavitation number of only 0-073, or about a quarter of the lowest number con- sidered by the authors. In the absence of all information such as that given in this very useful paper, the propellers were {designed on the assumption that the relative efficiencies at various pitch ratios varied in much the same way as in the atmospheric condition, but the rpm would be increased. This apparently rather erude assumption worked! The propellers gave over 103 knots when absorbing the designed power at about the required rpm. If we apply this device of a revolutions factor to the pro- pellers of p. 103, we can obtain a close idea of the most suitable rom from normal data. For the Taylor-BP-ND|V_method Of presentation, the required BP values are, respectively, 2-65, 3-82, 5-0, and 6-16. The very small improvement in efficiency below BP = $-0 would rule out rpm of 900 and 1,300 as being insufficient to offset the increased weight and appendage resistance. For the lower two pitch ratios of the table, normal data would suggest NDJV values of about 88 and 101 respec tively. Applying a revolutions factor of 6 per cent, these ND/V figures become 93°5 and 107 and give J values of 1-084 and (0-947, which are almost identical with those given by the authors. It i submitted, therefore, that a factorial method gives in a few minutes a close approximation to the most suitable rpm and dimensions for any given conditions, provided they do not involve only partial cavitation, Detailed calculations can then bbe made from the information of the paper. It is, of course, true that a factorial method gives no clue as to suitable area, Dut for highly loaded propellers, this will largely depend on strength considerations and not on relatively minute changes in efficiency. Tn conclusion, it is most satisfactory to find that low cavita- tion numbers can be associated with the high propulsive efficiencies reported in this very valuable paper. 123PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT ‘Mr. G. S. Selman (Member): A number of us who have been ‘associated with high-speed craft have for a long time used propellers which had crescent-shaped sections. Where severe cavitation exists such propellers have proved more efficient than any other. Unless one had the means of experimenting, the design of such propellers was a hazardous undertaking. ‘The present paper is so very welcome since it provides data which will tenable the designer to modify and improve and to venture, ‘with assurance. ‘Such sections are probably best associated with relatively small disc area ratios and large slips. In such conditions crescent-shaped sections give better results than others providing that one can locate the stagnation point at the centre of the Teading edge and in addition the centre camber line at zer0 angle of incidence. If a negative angle exists at any point during its rotation such propellers will suffer face cavitation and erosion. They are, therefore, useless unless a constant wake and even flow into the propeller disc can be guaranteed. ‘To an extent the presence ofa forward strut has a bad influence bby inducing a local wake, and the authors are to be commended for its use in their experiments. However, one could choose to place the strut aft of the propeller, and it would add to the value of this paper if experimental results with this alternative could be provided. If such were forthcoming a direct com parison with Gawn and Burril’s 1957 paper could then be made. ‘Where there exists a complete breakdown on the backs of the blades I hold the opinion that a sharp leading edge is no longer necessary or indeed desirable. I would suggest that a thick, round-nosed aerofoil section may in such conditions fill the cavity and give a better result. Such sections would also lower bending moment stresses. Many years ago I designed propellers of this type for Miss Britain IM and they proved very successful Since the trailing under edge, in effect, makes up for pressure lost at the leading edge, I would like to ask the authors if they have experimented with sections having the maximum under ‘camber closer to the trailing edge. ‘The propeller described in this paper has under camber at all radii. May L ask if face cavitation, when experienced, was ‘more severe at sections close 0 the boss? Fig. 2 gives a side elevation of the propeller and appears to indicate that under camber dies away at the extreme tip, resulting in the peculiar hooked tip shown in this diagram. Js this correctly drawn? Fig. 6 indicates that at high slips the efficiency declines, but little, as cavitation conditions increase, and that both thrust and torque reach a minimum at quite modest cavitation numbers. This might be reasonably inferred after complete back breakdown had been achieved. However, at low slips there is a continuous reduction of thrust, torque, and efficiency as cavitation numbers decrease. ‘This latter is, I think, characteristic of crescent-shaped sections and at J values above 1-2 I would prefer an ordinary segmental section, ‘Would not Fig. 9 be more useful if recast on a slip base? If this is done it will be seen that face cavitation occurs at slips Jess than 20 per cent, even at very low cavitation numbers indeed. Surely nobody wants to use such low slips anyway. In this instance the authors’ four curves may be reasonably replaced by ‘one, which shows how misleading the use of J can be in cavitating ‘conditions where slip or V/P N is of paramount importance and fan altogether better criterion. "Am I right in an impression gained by reading this paper that the authors regard supercavitating conditions as desirable? T would like the authors to explain what they mean by con clusion No. 6. Mr, J. W. English, B.Se, (Associate-Member): The data con- ied in this paper provide a very welcome addition to the rather limited amount of propeller performance data under cavitating conditions. The use of an upstream shaft in work of | this nature must be particularly important when considering the free formation of the core vortex and the low pressures asso- ciated with it acting over the surface of the boss. From a cursory glance at the flow conditions associated with upstream, and downstream cavitation tunnel shafts, it looks as though the flow patterns may be amenable to simple calculation with not 100 restrictive assumptions, in which case first order correction factors could be derived for correcting from downstream to upstream shaft conditions. “Having concurred on the importance of using an upstream shaft the authors will agree that such corrections, if of sufficient magnitude, should be included in work from tunnels possessing downstream shafts, the results from Ref. 7, for example. Have the authors any knowledge of such corrections being in existence? ‘The practice of presenting faired experimental values is admirable, for then one can include tunnel wall corrections of ‘one’s own choice, however, for the propeller designer prepared {0 use the Wood ind Harris corrections diagrammatic presenta tion of the results, as in Ref. 7, would have been greatly appreciated and need not have taken much space. AS this type of propeller is mainly used in craft in which there is normally a high shaft angle, the effect of shaft inclina~ tion on propeller performance becomes a factor of high impor- tance, and the separate contributions of the authors to the recent LT-T-C, meeting have been noted with interest, In this con- nection, therefore, it is necessary to have rather more guanti- tative data on the effect of shaft inclination, and itis to be hoped that the authors may soon publish more results from their tunnels filled for inclined shaft work, Turning to the design of the parent propeller, would the authors confirm that a propeller design method, as indicated in the list of references, has been used to design this propeller and, if so, which one.’ Whichever method or methods have ‘been employed, the fact that the design requirements were met only after an arbitrary snubbing process had taken place seem to indicate deficiencies in the design method. The subbing at x = 0°7 on the parent represents an increase in incidence of the Chord line of the section by about 0:92.deg., a substantial amount. Once the deficiency in pitch had been discovered in the parent, a more satisfactory procedure to adopt would have bbeen to alter the design conditions of the parent rather than alter the blade shape in such an arbitrary manner, The authors’ conclusions that the propeller vortex theory as developed for non-cavitating propellers is applicable to cavitating propellers, as propounded previously by other writers, is only a rough ‘engineering assumption, and must remain in use until a method mote descriptive of cavity flow processes is evolved. At the moment the vortex theory technique appears to be a poor substitute. Dr.-Ing. F. Gutsche: The authors are to be congratulated on having performed the propeller cavitation tests with success and in such a thorough manner, and presented all the data in such a convenient manner as in Tables I 10 XII. The results show that the type of sections has some influence on the performance of the propeller, especially at working conditions with higher propeller loading, and in this respect itis to be welcomed that the tests are extended to the regions of propeller work in which the sheet cavitation on the suction side emanating from the leading edge is fully developed over the blades in the so-called state of “supercavitation.”” It is interesting to learn that the authors, in the first stage of their work, had been induced to alter the nose part of the sections to avoid face cavitation. By this alteration they got a 14PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT section with an increasing effective curvature from the leading to the trailing edge. The fundamental behaviour of these sections in the fully developed suction side cavitation seems not to be very different from that of the usual ogival sections flat face. For instance, comparing the influence of the cavita- tion number ¢ on the lift coefficient Cy the simple Betz formula Cy = Ge > @ gives for the lower values of 7 <0-1 a very good approximation to the lift coeficients shown in Fig, 12. For the dragelift ratio « one gets with full suction side cavita- fon the formia ema + SESE, wih C= oie of Sate surface friction sestance and C, ~ cosficient of viscous ‘pressure drag induced by the finite thickness of the leading edge. By comparing this formula withthe results collected in Fig. 12 and usinga vale Cy ~ 0-003, the dag factor C, becomes about 0-008 for the propellers with a blade area ratio B.A.R. = 0-48 tnd 0-005 for BAAR. 0°95, respectively. Theae values are in Close areement with the experimental resus ofthe Walchner testo on three ogival sections. "The authors statment that the Cavitation pattern with sheet cavitation emanating fom the leading edges of the blades is stable with angles of incidence ony > 2 deg i alo endorsed by Walser’ observations, thelintrpotation of which leads to the formula, ) bo vie a =015(1~ ‘with «= angle of incidence in radians; 1 = thickness of the ogival section; ¢ = chord length; ‘$o = centre angle of the circular arc in agreement with the profile suction side. ‘The results of the quoted formula with cavitation number are for BAR.=0-48 4 =3-3 deg BAR. =0-71 2 =2-Tdeg. BAR. =0:95 «= 2-4deg. It would be outside the scope of the discussion to give special further information on. the relationship of cavitation number ‘and angle of incidence for the stream flow pattern with cavities vanishing behind the trailing edges of the blades. Regarding the dependence of cavitation number on blade section radius, the agreement of the authors’ result op,7< 0-08, with the ‘experience built up by Walchner’s observation, is also a good one. Regarding the definition of the angle of incidence of equivalent bb :de section at 0-7, I would be very much obliged to the authors to learn what should be the exact meaning of “corrected for loaded surface effec.” Should it mean the correction induced by curved streamline flow worked out by Ludwieg-Ginzel? ‘And, furthermore, which is the profile reference line in Fig. 3 for the angle of incidence after the alteration of the blade sections? ‘Authors’ Reply Mr. Chapman very correctly calls attention to the limitations of the series data presented in the paper to axial flow conditions, i.e, the results take no account of the effects of such factors as inclination of flow, variation of velocity across the propeller dise, air content, and scale, The effects of any of these can be appreciable; for instance, as Mr. Bedbrook has pointed out, the inclination of the shaft alone can affect the efficiency by as much as 3 per cent if the flow is inclined by about 10 degrees, and, furthermore, this effect can be to advantage or disadvantage, = Warcusen, 0.: Kavitation,” | Hvdro- ‘mechanische Probleme d. Schiffantrebes, Tet 1, Hamburg, 1932. according to whether the cavitation index is low or high. More details of the results of the A.E.W. and Vosper investigations in this particular direction, as presented to the 9th LETC. in Paris, September 1960, by the authors, are given Tater, ‘The discussions at that Conference brought out an important requirement for future propeller investigations. This is the need to simulate the full-scale ship conditions more closely when conducting cavitation tunnel experiments. Unless this is done the results for torque, thrust and efficiency can only be regarded a8 qualitative, and it has now been clearly established that the cavitation patterns on a propeller in axial flow are very different from those pertaining to the full-scale conditions. Consequently the effect of cavitation as regards erosion cannot be judged upon axial flow model experiments. ‘Tn the meantime, however, as Mr. Chapman has implied by his reference to the Brave class F.P.B. propellers the experi sent results can be relied on toa large degree for comparing the Performance of different designs of propeller to arrive at the ‘optimum. It is then necessary to apply correlation factors based upon previous experience in the field to arrive at the full-scale performance. Having acknowledged the practical value of the paper, Mr. Silverleaf then expresses certain disappointments which he declares not to be adverse criticism, but which to the average reader must appear to warrant this ater description. He seems to be torn between the alternative of believing data obtained by experiments on practical propellers and that which can be derived from pure theory which few would claim, as he seems to do, to represent the truth. For instance, whilst the paper clearly states that the mathematically shaped blades of the propellers had to be slightly modified atthe leading edges to eliminate face cavta- tion and so get good performance, Mr. Silverleaf favours taking a “broader” view. Precisely what this broader view can be, ‘ther than a purely theoretical approach and placing complete reliance upon theory employing assumptions which do. not entirely cover the known physical aspects, is not clear. “The teuth is that real progress in any’ field is made possible only by a healthy appreviation of the inability of theory to represent fully the physical aspects and the fact that no theory can be proven to any degree without practical experiment. Only by the marriage of theory and experiment can the accuracy of the former and the perfection of the later be brought near to the correct physical conception. Constructive criticism is the catalyst in this process and on the premise that the remarks of Mr, Silvera lie in this category, the authors will attempt to answer some of his numerous questions. The design method is not that given in Ref. 12 but similar in ‘many respects. Ref. 12 had not been published when the parent propeller was designed in July 1955. “The method used for the design of the parent propeller is based on the well-known, propeller vortex theory using L. Rosenhead’s (Ref. 13) approx ‘mate formula for calculating the section lift coeficients for 2er0 cexvitation number and the elation C=C.) [1 +9] to allow for the radially varying foal cavitation numbers. Only one force coefficient was specified but the other was, of course, calculated and both, as well as efficiency, are given 4s desig values in the Table on page 72. The datum line for the definition of the angle of incidence and of the face pitch ratio quoted in the second Table on page 67, is the nose-tail line of the section as finally modified. ‘Whether itis more useful to plan a methodical series of pro- pellers by systematically varying pitch ratio and blade area ratio or by systematically varying the’ specified operating conditions is very much open to discussion. “The aim of the present series wwas to cover the range of blade area ratios and pitch ratios which appear practical for engineering applications under severe 3 1sPERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT cavitating conditions. ‘The numerical example in the Appendix shows that this has been achieved. ‘The authors fail to see any Justification for Mr. Silverleat"s categorical statement that the ‘method of varying specific operating conditions is the only one which should be used for propellers designed to operate under severe cavitating conditions. Mr. Silverleaf says that to obtain blade section characteristics from propeller experiment data by means of the equivalent blade section procedure (Ref. 10) is tantamount to inaccurate aiming of a sledgehammer to crack a nut. Whilst the authors do not claim that the process is perfect, nor regard the alternative of ‘obtaining section characteristics from foil experiments as in the same category, they wish to point out that very little data ‘on foils was available at the inception of the series reported. Detailed study of the behaviour of propellers under severe cavitating conditions were commenced at A.E.W. by Mr. Rader in 1953. ‘The equivalent blade section method was then used to analyse the results under cavitating conditions for a number Of widely differing propellers. The lift coefficients and drag-lift coefficients obtained in this way varied similar to those obtained fon isolated two-dimensional foils which were known at the time, Moreover, the analysis showed that the ideal efficiency of cavitating propellers can be assumed to be the same as for non-cavitating propellers provided the thrust loading coefficient, the distribution of thrust, and the rate of advance are the same. ‘The authors also fail t0 see why the conclusion that, contrary to theory, the drag-ift ratios for fully cavitating conditions vary with thickness chord ratio and with pitch ratio should be influenced by the way in which the methodical series has been derived from the parent, In the authors’ opinion a theory which stipulates infinitely thin leading edges and non-viscous fluid ‘cannot be expected to give results which agree with experimental results obtained for sections with finite leading edge thickness in viscous flow. ‘As far as design and analysis procedure are concerned the authors used what, in their opinion and based on previous ‘experience, were the best methods available at the time. No doubt a great deal of work is still required to develop a reliable design method for fully cavitating propellers. However, the authors cannot follow Mr. Silverieaf’s reasoning for suspecting that “potential flow, blade element theory, with one or two necessary corrections would give closer agreement between calculation and measurement for supercavitating propellers.” If this argument were correct one would expect it to apply also to non-cavitating propellers. It is well known, however, that the propeller vortex theory is more accurate than the blade element theory, as itis known today. ‘Commander Du Cane adds still further to the list of imponder- ables which make the approach to fullscale conditions in a cavitation tunnel so difficult, and his disappointment at the abandonment of the intention actually to conduct tests in a fullscale boat is shared by the authors. Maybe this paper will inspire other investigators with a more immediate interest in the development of planing forms to carry out such tests ‘Commander Du Cane is quite correct in concluding that the Brave propeller is not really in the supercavitating range. However, as the full-scale propellers work on inclined shafts in a divergent flow, the cavitation pattern cannot be judged by the photographs in Fig. 8, In that part of the propeller disc where the inclined flow causes an increase in angle of attack and relative velocity the cavitation pattern is more like that shown for J = 0-9 and ¢ = 0-25, whereas in the other half of the disc, where the angle of incidence and the relative velocity are reduced, there will be less bubble cavitation on the backs of the blades than can be seen in the photograph for J = 1-0 and ¢ = 0-3, but there is a certain amount of face cavitation. It is difficult to explain why the rudders on the Brave class Fast Patrol Boats do not suffer severe damage from cavitation ‘erosion. Owing to the inclined shaft effect on the cavitation pattern there may be fewer individual bubbles in the propeller slipstream than might be concluded from the axial flow photographs. ‘may also be a question of bubble size, Professor Knapp (Ref. 14) has shown that cavities of a given diameter collapse at a fixed distance from a guiding surface and a slight shift in the range of cavity size may have a large effect on the damage ppotential, It is also conceivable that the air which is liberated from the water by the action of the propeller will give some cushioning effect for cavitation bubbles impinging on the rudder. The authors doubt whether itis safe to conclude that the type of bubble cavitation on the Brave propeller does not ‘cause damage under all conditions. Therefore they do not think it advisable to design propellers which work to the right of the hump in the Kg curves of Fig. 14. ‘The experience quoted by Commander Du Cane of M.T.B’s returning without propeller brackets and yet without any apparent difference as regards performance or vibration is most interesting, but perhaps quite explicable in the light of knowledge gained since the occurrence. In the first place, as he points out, adequate investigations have now been carried out to prove that the added resistance of brackets and other appendages reaches high proportions as speed increases, and equally that obstruction 10 the flow to the propellers’ adversely affects efficiency and cavitation. Furthermore, it is now well known that any discontinuity in the wake arising from appendages ahead of the propeller is one cause of hull vibration induced by the propellers. "The authors are indebted to Mr. Pehrsson for the information he has presented comparing the tunnel efficiency of one wedge- shaped fully cavitating propeller with that of the propellers of this series. The information that wedge-shaped propellers have actually been tried in a craft at sea is particularly interesting, and is believed to be the first successful attempt to do so. From, the model results it would appear that under cavitating con- ditions propellers with this wedge-shaped section can have as high if not higher efficiency than propellers of more orthodox design. One cannot be too categorical on this point, however, until carefully controlled measured-mile trials have been con- ducted using propellers of both types, and it might be as well to reserve judgment until then, and in the meantime take the lead given by Mr. Pehrsson that the wedge-shaped section offers distinct possiblities as regards performance and the avoidance of erosion. One disadvantage of the wedge-shaped section is its high drag ‘under non-cavitating conditions caused by vortex formation behind the blunt trailing edge. For this reason the efficiency of propellers with wedge-shaped sections under non-cavitating. conditions is rather poor. This is a point which has to be considered very carefully when this type of propeller is used for the propulsion of craft with a pronounced hump in their resistance curve such as hard-chine craft and hydrofoil boats, especially if they are driven by internal combustion engines. ‘Appropriately cambered high-speed sections with sharp trailing edges as used for the propellers of this series have good charac- {eristics in the whole speed range and give good acceleration capacity. It might be mentioned here that H.M.S. Brave Borderer reaches het maximum speed of over 50 knots from rest (engines idling, gears in neutral) in 30 seconds. ‘As stated in our reply to Commander Du Cane, the authors, agree with Mr, Cox that any variation in the velocity distribution 10 the propeller, whether this be brought about by inclined shaft, appendages, or hull form, must introduce what are now com= monly termed “mixed” wake forces which are transmitted 10 the hull through the brackets and induce vibration. Precisely how serious this effect is will depend on the hull form charac teristics, the actual degree of inclination of shaft, the form of 126PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT the appendages ahead of the propeller, number of blades in any one propeller, and its position relative to middle line. Investigations have shown that mixed wake forces arising from the hull wake are much more serious for centreline propellers than for wing propellers. ‘However, experience with several full-scale installations shows that the cyclic variations in thrust and torque associated with inclined flow do not lead to serious vibration problems, but as Mr. Cox would probably agree much depends here ‘on the design of hull and shaft supports. ‘The advantages of a fairly conventional basic thickness form with sharp trailing edge have already been pointed out in the reply to Mr. Pehrrson’s contribution, ‘Mr. Cox asks why the N.A.C.A. a = 1-0 mean line was used for the face shape. He probably meant to continue “in pre- ference to a supercavitating face shape.” Tulin’s super- cavitating sections were, of course, known when the parent propeller was designed but only as’a mathematical shape for two-dimensional flow. No experimental data were available at the time. However, a limited number of individual high-speed propellers with a fair amount of camber in the blade sections hhad been designed and tested at A.E.W. and analysed by the equivalent blade section method as mentioned in the reply to Mr. Silverleal’s contribution. In these propellers the N.A.C.A. mean line a = 1-0 had been used as mean line proper with the thickness form arranged around it in the usual way. Under cavitating conditions these propellers proved superior to pro- pellers with flat-faced sections. At the same angle of attack and cavitation number the lift coefficients were higher and the dragclift ratios lower. The results of this analysis were used as ‘uidance in the design of the parent propeller for the present The order of improvement gained by modifying the leading edges twice is shown in the following Table for the cavitation number 0-3, ‘The turbulence level of a similar tunnel with the same con- traction ratio and dimensions of working section but reduced height, was found to be 0°75 per cent, This value could be used as a guidance for Reynolds number correction ‘No attempt has been made so far to present the data of this series in the B, and 8 form. It should be possible to do but it is doubtful whether the number of propellers tested is sufficient. ‘The artifice of using the equivalent tunnel cavitation number ‘was introduced to overcome the many imponderables mentioned before. When model propellers are tested under physically more realistic conditions (inclined flow) this should no longer be necessary, as shown by the comparison of the inclined flow tunnel and full-scale data shown in Fig. 19. Original section 0-602 First modification 0-653 Sections as used for series 0-670 ‘The question relating to the effects of burbling cavitation on performance and cavitation erosion has been dealt with already in reply to Commander Du Cane’s contribution In reply to Mr. Crago, the placing of the model propeller in the tunnel to accord with that in the ship, i, downstream from the shaft, has, by and large, two advantages which Mr. English also calls attention to. First, the flow of water to the propellers is more representative of that atthe ship and, second, the forma- tion of the cone vortex can be observed and, if necessary, reduced by attention to cone shape. No correction factor is considered to be necessary, however, between results obtained in experi- ‘ments with the propeller upstream and those obtained with it downstream, because the measured results have been corrected for “idle” thrust and torque generated by the boss and shaft alone in accordance with normal practice, ie. the Ky and Kq values quoted are for the blades only. This also answers one Of the questions by Mr. English. ‘The upstream shaft will of course produce a small effective wake. AS it is confined to the sections close to the root, the cffect on the performance characteristics is negligible. “The Uncertainty of the effective Reynolds numbers of the blade sections should have no great effect on the accuracy with which the blade element efficiency can be scaled. ‘The uncertainty of ‘the roughness correction is probably @ greater imponderable. alae Foo i = i ; . a ‘MODEL PROPELLER ie etter Z : Fas a ‘weuneD row. foo A. fe 1 aa “MODEL PROPELLER : eee fo — . ar CAVITATION NUMBER Oj, ‘om FULL scate PRopeLien Fig. 19, For rather costly high-speed craft it is always advisable to test ‘a model of the full-scale propeller with inflow conditions simu- lated as accurately as possible. Ifthe experiment is so designed that the drag of the appendages can be measured at the same time, the use of a quasi:propulsive coefficient factor becomes superfluous. A systematic series of tests on appendages would cettainly be very useful, but it would not of necessity obviate further detailed tunnel tests of systematic series of propellers. Mr. Crago also raises the question of how tolerances ‘machining should be specified and appears to prefer a definite v7PERFORMANCE DATA OF PROPI figure rather than a “plus and minus” which is usually quoted. He infers that the latter can involve a larger error in the finished ‘product in any ofits linear dimensions, and a more non-uniform finished surface. In the latter he may be correct, but in the former not necessarily so because so much depends upon the figure actually specified. For instance, if the surface machined to = tolerances is found to be unacceptably undulating then surely this can be cured by reducing the figures specified. The trouble is that this practice of specifying + tolerances is not only accepted by propeller manufacturers but is also preferred by them. One way of overcoming the difficulty is to specify rot only the + tolerances on, say, thickness, but at the same 1 to place another restriction on the finished product, making it clear that the actual finished thickness should nowhere be reduced or increased to unacceptable proportions. For instance, the Admiralty tolerance on thickness for full-seale propellers is specified as +440 thousandths of an inch or +1 per cent of the local thickness, whichever is the greater tolerance, ‘As regards the use of a correlation factor, it is clear from previous remarks on the wide differences between tunnel and ship conditions that this is essential to arrive at a propulsive coefficient for the ship. The value of the quasi-propulsive Coefficient factor or any other correlation factor used in the process of design, can only be selected from consideration of the design of hull, number of propellers, speed, and other characteristics in comparison with previous designs. Suffice it to say that, generally speaking, selection of such factors is not so precise as for normal displacement forms for which more data from previous experiments is available. ‘The authors are grateful for the remarks of Mr. Leathard endorsing the validity of the method used for the analysis of results and they agree with his views on the relation of circula- tion to lift coefficient and pitch under cavitation conditions. ‘No doubt it would be very gratifying if the work involved in the design procedure exemplified by the authors could be reduced. Inspection of the tabulated data shows, however, that for cavitating conditions there is no obvious feature such as constancy of optimum thrust coefficient with pitch ratio for given blade area ratio which Mr. Leathard found for the propellers of the Gawn 20-in, series propellers in non-cavitating flow. ‘Mr. Barnaby asks how sacrosanct ig the line connecting the minima of the Kg curves in Fig. 14. This line is only a criterion for supercavitating conditions. As pointed out by Commander Du Cane, the Brave propellers, even at or near full speed, are ‘not really in the supercavitating range and yet there is relatively litle evidence of cavitation erosion. It appears, therefore, safe to conclude that the region between the humps and hollows in the Kg curves can be used without incurring serious trouble from cavitation erosion. ‘The authors appreciate Mr. Barnaby's interesting remarks on the application of a revolutions factor. The agreement of the results obtained by the factorial method with those given in the paper is remarkable. ‘Mr. Selman points out quite rightly that the superiority of crescent-shaped blade sections for propellers which work under severe cavitating conditions has been known for a long time. Such sections were considered in the early stages of the design study, but it appeared from the results of the preliminary analysis mentioned already that the type of section eventually chosen for the series propellers would be more efficient. ‘There appears to be some misunderstanding as regards the presence of a forward strut. In the Vosper Cavitation Tunnel the shaft support struts are in the contraction upstream of the working section at a distance of more than $ ft. from the pro- peller under test. The influence of these struts on the flow into the propeller should be negligible. ‘The question of comparison with the data from the Gavn— 128 ELLERS FOR HIGH-SPEED CRAFT Burrill 1957 paper has been answered already in reply to Mr. Crago’s contribution. Mr, Selman’s comments on the shape of the blade section leading edges are very interesting. The basic thickness form of the sections used for the propellers of the series is elliptic in the forward half. Although Fig. 3 may give the impression that the Teading edges are sharp, there is a certain amount of roundness. It appears from theoretical and experimental information available that the elliptic thickness form is very close to the optimum, ‘AS far as face camber is concerned, the authors have not so far experimented with sections which have the maximum face ‘camber closer to the trailing edge. It is known, however, that the Johnson three- and five-term sections, which fall in this category, have been used by other experimenters and it is believed that the three-term section has given very good results. Mr. Selman’s question whether face cavitation, when ex- perienced, was more severe at sections close to the boss requires two answers. As far as the model propellers in axial flow are concerned, the answer is no. On some full-scale propellers, however, there is evidence of slight erosion on the face of the root sections. It is believed that this is caused mainly by inclined flow effects. Fig. 2 is correctly drawn and the hook at the blade tip does exist. ‘AS regards Mr, Selman's remarks on Fig. 9 the authors fail to see why slip should be a better criterion for cavitating con- ditions than rate of advance. ‘Whether supercavitating conditions are desirable depends on ‘many factors. In the authors’ opinion they are desirable if the speed of the craft which has to be propelled is so high that itis impossible to design a non-cavitating propeller. ‘The modifications mentioned in conclusion No. 6 would ‘consist in further lifting of the leading edge. When applied to the existing model propellers this would, of course, also involve further shortening of the blade sections. Both Mr. Bedbrook and Mr. English have referred to the effect of inclination of shaft and Mr. English has requested further information, There are, of course, several components 10 the inclination of flow to the propeller of which shaft inclination is only one. The others are due to the hull form, induction of the propeller itself, and the orbital velocity of the wave at the ster, ‘The resultant inclination of flow can be appreciable, To ascertain the order of the effect on performance of the propeller, inde- pendent experiments were carried out at Messrs. Vospers’ and at A.E.W., Haslar, the former on a propeller for a high-speed ‘craft for 124 deg” inclination, and the latter on a destroyer propeller for 10 deg. inclination, both relative to axial flow. ‘The results of the Vosper experiments are given in Fig. 19. ‘The thrust and torque values of the full-scale propeller measured during trials are corrected for the resistance and torque of the shaft and the corresponding components of the friction in the bearings. No other corrections were applied to the results Pechaps the most interesting result is that the efficiency in inclined flow increases with decreasing cavitation number. This, is even more pronounced at higher rates of advance and in ‘complete contrast to the results in axial flow. Itis believed that this improvement in efficiency is due to the changes in cavitation pattern associated with the cyclic variations of the blade element angles of incidence, relative velocities, and local cavitation ‘numbers in inclined flow, but a full explanation requites further experimental and theoretical investigation. In the case of the destroyer model propeller tested at A.E.W. the results confirmed those for the Vosper tunnel in that at nominal ship cavitation numbers, ie. above the cross-over point fof «about 0-7, the efficiency was reduced. For o = 1°5 the reduction was 2°5 per cent on 67 per cent. Mr. English’s question relating to the method for correcting.PERFORMANCE DATA OF PROPELLERS FOR HIGH-SPEED CRAFT from downstream to upstream shaft conditions has been answered already to some extent in the reply t0 Mr. Crago’s contribution. A definite answer as to the magnitude of such corrections, however, requires accurate comparative tests on an upstream and downstream shaft in the same tunnel. AS far as ‘the authors are aware, this has not been done so far. ‘The question concerning the design method has been answered already in reply to Mr. Silverleaf's remarks. “The fact that the leading edges of the blade sections had to be ‘modified does not of necessity indicate deficiencies in the basic design method. If a face shape with less or zer0 slope at the Teading edge had been used instead of the N.A.C.A. mean line a = 1-0 face cavitation at the design rate of advance would have been less severe and the original design would most likely have fallen less short of the requirements. As the parent propeller ‘was designed (o propel a certain craft at a certain speed it was not possible to alter the design conditions of the parent rather than alter the blade shape. The reasons for adhering to the original programme, which involved applying the modifications to all the propellers of the series, are fully explained in the paper. In reply to Mr. English’s remarks concerning the applicability of the propeller vortex theory to cavitating propellers, the authors refer to their comments in reply to Mr. Silverieal’s ‘contribution. ‘Dr. Gutsche’s comparison of the lift coefficients and drag-tift, ratios of the equivalent blade sections with Betz’s simple theory and the results of Walchner’s experiments is very interesting. ‘Although the agreement between the results is very good it ‘would, in the authors’ opinion, be incorrect to conclude that, in the supercavitating state, the behaviour of the blade sections used for the methodical series propellers is not very different from that of the usual ogival sections with flat face. ‘A comparison of the results from this paper with those from. the Gawn-Burrill paper mentioned already will show that, under severe eavitating conditions, propellers with hollow-faced blade sections are more efficent than propellers with flat-faced blade sections. For a direct comparison of the blade section lift coefficients, ‘and drggsift ratios with two-dimensional foil data the blade section lift coefficients would have to be corrected for the curvature of the streamlines. This correction was not applied because the actual lift coefficients obtained are more important from the propeller designer’s point of view. If the correction for the curvature of the streamlines had been applied, the blade section lift coefficients would have been higher and the drag-lift ratios lower. ‘The correction of the angle of incidence for loaded surface effect is given by the approximation Pwo y Aan ond K) where fyeom = the geometric camber of the blade section; "C= the chord length of the blade section; and K = the Ludwieg-Ginzel camber correction. This approximation is based on the method outlined by von Karman and Burgers in Ref. 1S. In the simple form given here it involves the assumption that the centre of pressure is at or close to the mid-chord position. In conclusion, the authors wish to express their gratitude to those who have contributed to this interesting discussion, and their hope that the data presented may prove valuable in the design of propellers for high-speed cra References (13) Rosenneap, L.: “Resistance to a Barrier in the Shape of an Are of Circle,” Proc. Roy. Soc. (A), Vol. 117, p. 417, 1928, (14) Kwapr, R. T.: “Further Studies of the Mechanics and Damage Potential of Fixed Type Cavities,” Cavitation in Hydrodynamics. Proceedings of Symposium held at the National Physical Laboratory, Teddington, Sep- tember 1955, (45) Duranp, W. 1 Aerodynamic Theory, Vol. 1, p. 232. 129