ACE No. LUHI5 NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS WARTIME REPORT ORIGINALLY ISSUED August ishh as Advance Conf Identlsa Eeport IAHI5 THE NACA IMPACT BASIN AND WATER LANDING TESTS OF A FIjOAI MOIiEL AT VARIOUS VELOCITIES AND WEIGHTS By Sidney A. Batterson Langley Memorial Aeronautical Laboratory Langley Field, Va. NACA WASHINGTON NACA W/vRTIME REPORTS are reprints of papers originally issued to provide rapid distribution of • advance research results to an authorized group requiring them for the war effort. They were pre- viously held under a security status but are now unclassified. Some of these reports were not tech- nically edited. All have been reproduced without change in order to expedite general distribution. L - 163 Digitized by tlie Internet Arcliive in 2011 witli funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation http://www.archive.org/details/nacaimpactbasinwOOIang 17. ? ** KACA ACR No. li|.H15 NATIONAL ADVISORY COnrJITTEE FOR AERONAUTICS ADVANCE CONFIDENTIAL REPORT TliE KACA IMPACT B/iSIN AND VVATER LALjDING TESTS OF A FLO>.T i,:ODEL AT VARIOUS VELOCITIES AND .VSIGRT3 By Sidney A. Batter son SUMMARY The first data obtained in the United States under the controlled testing conditions necessary for estab- lishing relationships araong the nun':9rous parameters involved vi/hen a float ?iaving both horizontal and vertical velocity contacts a water surface are presented. The data v/ere obtairied at the NAC/\ ?j:ipact basin. The report is confinad to a presentation of the relationship between resultant velocity and i:')pact norri:al acceleration for various float weights when all other paraineters are constant. Analysis of the experimental results indicated that the raaxlmum impaot norr.ial acceleration was pro- portional tc the square of the resultant velocity, that increases in float v;eight resulted in decreases in tiie nsurimuni impact normal acceleration, and that an increase in the flight-path angle caused increased impact normal acceleration. INTRODUCTION Until the present tirie , almost all experimental v;ork relatijd to loeds on seaplanes landing on water Jias consisted of full-scale landing tests. Attempts to use results of these full-scale tests to establish rela- tionships aiiicng tre various impact parameters have not been very successful for tvvo reasons; (1)A prearranged test program involving the isolation of selected pararaeters could not be carried out since the values for a number of the variables were a function of piloting technique and the natural conditions of the wind and the sea were not controllable during the test; and (2) the available instrvjnents proved inadequate to suppl;/- sufl'iciently accurate results. CCN?IPENTL\L rACA ACR No. l1jJI15 In order to overco^ne the disadvantages of full-scale testing, an inpact bssln *n vy}i:'ch float models could be tested under controlled conditions v/as constructed at the Langley Memorial Aeronautical Laboratory at Langley Field, Va. The first data obtained in the- ?TAGA impact basin, which are contained in the present report, may be lased with the results of subsequent investigations to establish basic relationships among the impact parameters. Logical interpretation of results of flight tests investigating conditions beyond the scope of the NACA impact basin will then be possible. The present tests are confined to establishing a relationship between resultant velocity and Impact normal acceleration for seaplanes of various weights. The solution of the problem of determining landing loads must follow further investigations under controlled con- ditions in order to isolate the effects of a number of other paraineters such as fliglit-path angle, dead rise, hull shape, txnd trim. V resultant velocit;,' cl' float, feet per second % horizontal velocity component of float, feet per second V^ vertical velocity comporxent of float, feet per second g acceleration of gravity \32.2 ft/sec-y Y flight-path angle, degrees (See fig. 1.) T float trim, degrees n. impact acceleration normal to water su'^face, g DSSCHIPTIOK OF AFFARATTJ3 Pleat Model and MCA Impact Basin The m.odel consisted of a float designed to conform to exceptionally h:gh strength requirements. Care was COyFTD=?NTIAL IIACA ACK No. Ll4-ni5 CON?IDENTI/.L 5 exercised during the design and construction of this mo-^el to obtain a reasonably smootli bottom. The sheet- metal sVln and most of the structural members were made of dural in order to obtain the ralnimiim weight con- forming to the load specifications. The weight of the model was l\.0'J pounds; however, provisions were included whereby 2000 pounds of additional v/eight could be bolted onto the sides in increments of 25 pounds. The lines and pertinent dimensions of the float model are shown in figure 2. A feature of these lines is the absence of all chine flare . The NACA impact basin is essentialljr a concrete tank 560 feet long, 24 feet vi^lde, and 11 feet deep with a normal avater depth of 8 feet. Heavy built-up steel rails are suspended along the entire length of the tank with the exception of the last ]|.0 feet, v/hich is to be occupied by a v/ave maker. The rails and a portion of the tank are shovm in figure 3» Ths i.ipper, lower, and inner surfaces of each rail were ground straight and parallel within a tolerance of 0.002 inch and the saiae tolerance was hold during installation in locating the rails with resDect to each other and to the water surface. A main carriage em.bodying a drop linkage to which the model is fastened travels down the tank on the rails. Figure I|. shov;s the general arrangement of the carriage Y/ith m.any of the secondary members omitted for clearness. The basic carriage structure consists of chrome -molybdenijim steel tubing (fig. ^); the total weight, without model and instruments but including the dron linkage, is approximately 5^00 pounds. It may be noted that both lower and upper wheels are provided. The upper wheels are arranged in sets of two and located in trucks v/hich sv^ivel in a vertical plane parallel to the longitudinal carriage center line so that the load is eqiialized betv'een the tv;o wheels. Solid-rubber instead of pneumatic tires are used to reduce to a minimum deflections mader load. Before installation, the outside diarieter of each v/he&l v/as ground concentric to the axle bearing and thun balanced. The lov/or wheels may be jacked up against the lower surface of the rails until both upper and lower wheels exert a predetermined pressure on the rails. Oscilla- tions transmitted to the carriage are limited by this arrangem-ont to very small amplitudes and therefore have little disturbing effect upon tho actual drop process and COKPIDENTIAL C0MPID3NTTAL NACA ACR Mo. ri4Jil5 the instruments. Lateral restraint is provided by four side wheels bearing upon the inner re.il surfaces. The drop linkage consists of the bocrn and the upper and lower linkbars, which are pivoted 3.t both ends and with the boom forrn a parallelogram t;rpe of linkage (fig. Ll) . The model was fastened rigidly to the lov/er boom fitting by means of bolts through three lugs built into the float (figs. 6 and 7)» By this attachment, the float was rastrained in all directions with respect to the boom but had freedom in the vertical direction since it was attached to the parallelogram linkage. The lower boom fitting provides a means for setting the float at various trims and angles of yaw, "rhe float may be dropped from any height up to I;, feet, depending upon the vertical velocity component desired, by engaging the corresponding rack tooth v;ith a latch on the carriage. This latch is released by means of a trip ca;;i located at the proper point along one rail. Releasing the latch allows the boom and the float to drop freely except for the restraint Imposed by the upper and lower linkbars, which keep the boom vertical as the float drops. The motion imparted to the model is not perpendicular to the water surface during a large part of the drop. Since the immersion occurs v/hen the linkbars are practically level, hov/ever, any horizontal component contributed by the linkage arrangement is at a m.inlmiTOi during impact and is negligible . The dropping weight may be varied b;/' the addition of lead bars fitted around the boom, and bolted together as shewn in figure 6. The total weight of the boom and llnkoars alone, and hence the minimum, dropping weight, is 700 pounds. In order to simulate v^ing lift, an air-cylinder and piston iirrangem.ent that can appl>' any desired lift on the model up to 2l).00 pounds is incorporated in the carriage. This mechanism is referred to as the "buoyancy engine.'' The lift is applied to the model by so connecting the boom, and the piston of tlie buoyancy engine with a cable and sheave system that the piston is forced to travel against .the air nressure in the cylinder as the float drops. The amount of lift exirted on the model depends upon the Initial air pressurr supplied to the cylinder before each run. The rod nannlng upward at an angle from. th.i bottom, roar noint of the boom lifig. 6) is the low3r--snd connec- tion of thi cable system. ;,7ith this arrangsmient the' application of the lift may bo witliheld throughout the downward travel of the boom until just be fore "the COKPIDE.NTIAL NACA ACR No. li|H15 GONFIDEivTIAL float contacts the v.-atsr. The float is thus allov/ed to attain the desired vertical velocity component . The carriage has no self-contained drive but any desired liorizontal velocity up to 110 feet per second may he attained hy means of a catapulting system. The catapult is of the type used hy the Navy on shipboard for launching service planes and accelerates the carriage to the desired speed in a distance of 60 feet. The drop linkage is released at such a point that the impact occurs approximately 100 feet from the point at which the catapult stroke ends. This procedure allows a period during Vi'hich most of the irregularities and oscillations inherent in the catapult run are damped out. Pol lowing the imoact, the carriage ru:i is terminated by a Navy arresting gear capable of dis- sipating the total kinetic energy of the carriage in less than 100 feet. instrumentation The operation of the horizontal-velocity recorder is dependent upon 1-inch strips of thin metal, referred to as ''Interrurters, " that protrude about I4. inches below the lov/er inside corner of one rail at 1-foot intervals throughout the length of the tank. These interrupters may be seen on the left rail in figure 8. A photoelectric cell is located on the carriage in such a manner that each interrupter causes a break in the photoelectric-cell circuit as the carriage travels dov:n the tank. The current is then fed to a high- frequency galvanometer element of a recording oscillo- graph in which a shift occurs in the record line each time the photoelectric-cell circuit is opened by an interrupter. In addition, the oscillograph record contains —^ — second timing lines, Inasm.uch as the 100 carriage is traveling at oractically constant velocity between the end of the catapult stroke end the impact, this velocitj'" can be determined by dividing the number of interrupters passed dviring this Interval by the time . The displacemient of the boom, and its velocity in the vertical direction are also recorded by the oscil- lograph. The displacement is recorded by a galvanometer CCNPIDENTIAL 6 COKFIDESfTlAL NACA ACR No. LLi.H15 element, v;hich deflects in proportion to the amount of current that flows through a piece of resistance v/ire. The effective length of this wire is varied 'jvith the position of tho hoom by completixip; the circuit through a sliding contact. The displaoerient of this contact along the v/ire follows the toora trave]. , The sane apparatus is used in the determination of the vortical velocity con.ponent. In order to determine the vertical velocltj'- component, hov»ever, the current derived from the slide- v'lVQ is directed into several high-capacity condensers and thus eloctrioallv differentiated. The galvanometer element records this change in current, v/hich is a function of the boom vertical velocity. The velocity is then derived from the recorded change in current by reference to a suitable calibration curve. The impact normal accelerations were initially determined with an aocelerometer' that recorded the flexure of a cantilever vane as measured oy a strain gage. The frequency of the accelerometer v;as 12,5 cycles per second. An amplifying system was required, hov/ever, to adaot the accelerometer to the oscillograph that recorded the other values. Since amplifying equipment vvps not available in time for the tests, a special recorder was necessary for this particular InstrTUaenc , The records obtained during the first ncrt of the test shov;ed that extraneous vibr^tion^ \.ere disturbing the galvanometer element which recorded the i?npact accelerations. An accel- erometer that recordoQ the angular displacement of an unbalanced galvanom.eter v/as therefore substituted for the rebt of the test. This InstrLU^-ient had a self-contained optical recording system., had a frequency of 10. 5 cycles oor second, and was apparently undisturbed by extraneous vibrations. The damping was between five-tenths and six-tenchs of the critical damping. The instrument was enclosed in a box and mounted rigjdl:/ on the boom between the front and rear float fittings. The mounting may be seen in figures 6 and ( TEST PROCEDUHE The data presented herein were obtained during the initial calibration runs at the NACA Impact basin, oonfii)]:ftial NAG A AC?i No. rii.H15 COrJFIDErJTIAL 7 The test program thus depended upon the calibration rsquiroments, v/hich necessitated rims at varying weights and landing velocities. A test progran was therefore foiTCulated that consisted of a systematic series of runfj frori which the variation cf ma^iimijin ncrrn.al acceler- ation was obtained as a function of resTiltant velocity and float weight with all other parameters constant. The model was tested at dropping weights of IICO, 1500, 1950, and 2I4.OO pounds. A complete series of runs •vas made for each weight, vv'lth hori:^ontal velocities covering an arrroxiriate range frow 55 ^^ 95 f-^et r^er second. An atteinpt wa.=! made to maintain throughout the tost a constar.t ratio of vertical velocity con- ponent V^ to. horizontal velocity coruponent Vj^. The value of V\,/Vi-. is designated tan y* where v is the flight-path ar.gle . The \'aluo of tan y was selected as 0.12s for the pres-snt tests. In order to check the effect of an increased flight-path angle, four addi- tional runs in v;hich tan y •'■'ss approximately 0.200 were inade for the dropping weight cf 1100 poiinds . The trill] v/as 5^ and the angle of ya'o; v/as 0^ throughout each series. During the i.ioact process, a lift equal to the dropping weight in each case was exerted on the float Ly means of the buoyancy engine. The normal accelerations were recorded and the m.axiinu.n value was noted for each run. PRiiCISIOIT The apparatus used in the present tests gives measurements that are believed correct v/ithin tjie foil o w i ng 1 ir:' its; Horizontal velocity, feet per second ,,,,, u ....... . •^0,5 Vertical velocity, feet per second .,,, 1 .-,.,, ^ ,,.. , --0.2 The strain-gage accelerometer v;ar^ us3d throughout the series in which the dropping weight was I5OO pounds but was replaced b^" the galvanometer accelcrometei' for the other bhree seiies of tests. The accuro.cy of 2^5 percent for acceleration measurerients refers to the galvanometer accoleroraeter . This accuracy represents the extreme limiiis of error possible T-.hroughcv.t the range of applied load frequcnciec from the stable to COiJPIDi:RTIAL 8. COiTPIDSI-ITIAL 17 AC 4 ACR No. IJ4.HI5 the nati.ir.?.l freouency of the accelerorrieter and is based \apon observations of a freqi^enoY-responss curve aerived experimentally for the Instr-uunont . Inasvnuch as the test results shovved no marked decrease in aocelerafcion vr.lues at high-frequency loads, it was concluded that the natural froquenoy of the accelarometer was not exceeded. In the lovifor velocity ranges, the atternpt to maintain constant flight-path angle v/as not very successful, beoa"ase no calibration data v/ei'-e available and the horizontal v=^locity components had to be esti- mated. In addition, dsviationv^ ^occurred betv;een the vertical velocity oornr-onents expected I'rora drop calibra- tions made \fin.th the carriace at; rest and the vortical velocity ccrrponents obtained during the tost runs. The magnitude of these deviations docroased loss rapidly than the corresponding^, vertical velocity components and therefore decreased^ the accuracy of the flight-path angle less at high tlian at lov; velocities. RESULTS AND DISCUSS lOIT The results of each series of tests for dropping weights of 1100, 1'300, 19yf^'j ^'^'^'^■- 2'|00 pounds are shown as logarithmic plots in :''i£_:uros 9 '^^ ^-2, respectively. The impact normal accelei'ations in g units are derived from the accel'^-rcme uer record. Inasmuch as the buoyancy engine contributed a force equal to the dropping woight, Ig v/as subtracted from the vaiu.es obtained frora the accelerometer recordi to isolate tho force resulting from the impact. The re:.mlts of the foui' runs for v;hich tan y was approximately 0.200 are ulotted in figure 9, which clearly indicates tiiat an increase in the flight -path angle increased the normal acceleration. The short-dasxi lines in figures 9 to 12 have slopes that represent the proportion n^ cc V2 and pass through the experimental points that correspond to tan Y ** 0.125 -^s deter-ained from figure 15, which shows th.e variatiDn in flight-path angle with i-esultant velocity for tij.e four dropping weights. By referring the faired experimental curve of figure 9 for CONPIDSNTIAL • ' . -TT C NACA AGR Ho. I4HI5 COI'IFIDEI'TIAL tan Y~ C'«12[; to figure 15(a), the ma^ximum normal acceler- ation Gun be observed 'co be direotlj'^ proportional to V"^ when the flight-path angle remains constant. Below 72 feet per second, however, this proportion no longer holds and the maxl:n-un norrr^al accelerations; show a larger increase with resultant velocity than is indicated by the line for nj_ cc v2. This trend is expected in order to be consistent with the four points of figure 9 that were obtained at tan v ** 0,200. This analysis can be applied to figures 10 and 12 and figures 13(b) and l$(d), respectively, althoiigh the range in which the flight- path angle remains const&nt and the amonnt by which tan Y varies differ somewhat in figures 15(b) and 13(d), 3y applying this analysis to figure 15(c), it would be expected that the values of maxlvjum normal acceleration and resultant velocity that correspond to values of V from "^6 feet per second (the point at which tan x ~ 0,125) to the maximixm velocity would show some proportion other than V- since the line ro pre stinting tan y "^'^^ a definite slope within this range of V; that is, at V > 76 feet per second, the values of raaximuin normal acceleration should lie below the curve for n.cc V^ whereas, at V < 7b faet per second, the values of maximum normal acceleration should lis above the curve for n^ or. v^. Figure 11 shows that such is the case. The dashed line for n-j_ cc V?- is determined by the tb.ree points in figuro 13(c) at which tan y ~ 0.125 , If thus appears that, pi-ovlded the flight-path angle is constant, the maximum non:ial accelerationc resulting from the Vi^ater landing impacts vary directly as the square of the resultant velocities. This conclusion agrees with the Nev.'tcnlan or V^ lav; of fluid resistance since the Reynolds number w?.s very large. The slope of the lino through the points for tan y = 0.200 in figure 9 cannot bo relied upon as indicative of the true trend since only four points were obtained and the r5.ng& of y '"^^ approximately 10 percent. The dashed lines of figures 9 to 12 are replotted in figure lIi, vmich therefc^e presents the experim.ental variation of maximum impact normal acceleration with resultant velocity for various dropping weights with all other parameters constant. It may be noted that the maxim.um normal acceleration decreases as the weight increases. The curve representing a ".veight of I5CO pounds shifted, as was mentioned previously, b?^- cause the galvanometer elem.ent that recorded the acceleration for this series of tests w as out of b a 1 an c e . GCNFIDhivTIAL 10 GOrTPIDENTIAL NACA ACR No. li|H15 superimposed upon the aocelerometer record and accelerations gruater than the actual Imnact accolerations were ccnso- quently recorded. Ko atte^npt was made to evaluate the decrease in raaxi'.num nonnal acceleration that resulted fr.'oiri the Increase in v/elght since the data appeared inadequate for this purpose . CONCLUSIONS Re su.lt 3 cf tests ui the N/vCA ir.ipact basin of the variatiC'n with resultant velocity and weight of the noi'-mal acceleration resulting from landings of sea- planes on water indicated the follov/ing conclusions: 1. The riaxinuin impact nor:!ial cxceleration v;as proportional to l:he square of the resultant velocity in accordance with the 7'^ lav/ of fl\iid resistance. 2. The riaximuiti impact normal acceleration decreased as the weight increased provioed all other condibicns remained constant. 5. An increase in maxim^an impact normal acceleration accompanied nn increase in r].ight-psth angle provided all other conditions remained constant. Langley Memorial Aeronautical Lahorato:'^y National Advisory Committee for Aeronautics Langley Field, Va. CONFIDENTIAL NACA ACR No. L4H15 Fig. Q O O k ii _l o k ^i z: S * s X CJ k R < ^J tH 3 2 Ed Q 1 — 1 ^ 2 o o k ^ .^ ^ a ^ ,^ ^ ^ [ij S :3 ^ NACA ACR No. L4H15 Fig. _3 a, o o k I ;:: ^ K a; o -J I s <--) NACA ACR No. L4H15 CONFIDENTIAL Fig. 3 Figure 3.- Tank and rails of NACA impact basin, CONFIDENTIAL NACA ACR No. L4H15 Fig. ►J < I— I z Ixl Q »— < Z o o ^ ^ ^D k ^) ff ►J 'S < <: 1— 1 ^ Eh 5^ to Ed ^ Q ^ 1— 1 2 O L.I O VI 'J: tv IV (5 ^J^ U| IV -■^ ,'■'1 t^ NACA ACR No. L4H15 CUNFIDENTIAL Figs. 5,6 Figure 5,- Front view of carriage in NACA impact basin, HflTft 36Z03 Figure 6.- Side view of model fastened to boom in NACA impact basin. CONFIDENTIAL NACA ACR No. L4H15 CONFIDENTIAL Figs. 7,8 3fcZ04- Figure 7.- Front view of model fastened to boom in NACA impact basin. Figure 8.- Photoelectric-cell interrupters in NACA impact basin . CONFIDENTIAL NACA ACR No. L4H15 Figs. 9,10 -^ ^ ^ ^^ s 3 ^ k y 3 S •\ X^. '« I v; «i §- lO ^ « 1 1 1 • i^ i > 'uoti OJS/3J JO (DLUJ ou 4.ooduj/ UUniM/KOl/^ I -C p s; 53 ij O^ ~S 1 8^ t^P § iT ^ 5^1 2 J^ o ^vg §. ^ s^§ ■k 5 a^ s tv i 9 c -c <*; ■t- G *. St 5 O 5 ^ NT "O 1 >^ ^ O ::; 5§: 0, i- Q- h^~^ ^ Kl "-5 '3 N V. ^ ^ Z °^\. K3 - '^ N \-^ X^, |g P i o ■ 1 V \ ^ -K^^, ' X X \ X \ 6' 1 '0//0. ia/303 C jOIMJ ou 4:indiMi ijun'jj/xo^ «r o -s. fl, ■c (b ^ Q ^^ — ^ ^ ? ^ O ? V 4~ 1.^ -. ij 6 c; V - ? cx. § \ ^ ^ ^^ c 6 \ 't? u •o -<< 0) c C Ch ^ <:' rj -K, ^-1 i 1 Si, "> -r Q "il ^ it: ^, 'C NACA ACR No. L4H15 Figs. 11,12 "^"^ S ^ ^^ ^ "> is 1 3^ '^v. C 1 ^^ » li t5 > 1 1 £'.'(/ 'L/o//oja /a:>De /oujjc >i/ 4.Doduui wnujiyo W ^ 9 ^ iC In I a 8^ 4. 5^1 iji c eg ^ Q I I I 1 1 -K O Q. 3 ^ NACA ACR No. L4H15 Figs. 13,14 < I— ( Q o a 8 Si ■^~X> ■^- 1 ^ \ c-^ ■~-^-> ■ ^^\,,^ ^1 1 ^ HJ 1 a >\ ii r-1 .^1 s ^ r 1; i 1 : 1 1 S \J ■6 p 6 !u ', JO 140. js/sooi lOUJJC 7U ^ouduji wna/ixof^j 5 01 «5 G ■ ) G b 1 -0 ~i / / / M / L I I 1 1 / / t / J u/ I I I i / 7 i> / OJ / I § r I O § .^ I 1^ -K ^ ?^ ft) W 15 i ?^ < .5^ -K to to ^ ?! f " X. z/£uD i/.^j>o^y.L/£/// J.O ^ua£uDj_ I UNIVERSITY OF FLORIDA 3 1262 08106 485 8 :^N-,'-no!TY OF FLORIDA NTS DEPARTMENT 3T0N SCIENCE LIBRARY 117011 ■-.. '.i'^tciVILLE, FL 32811-7011 USA r-