THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING SOUND POWER SPECTRA FROM SUBSONIC JETS and SOUND POWER STUDIES OF MODEL SILENCERS Norman E. Barnett December, 1964 IP~689

( e 1T,

TABLE OF CONTENTS Page LIST OF SLIDES................................................. iii SOUND POWER SPECTRA FROM SUBSONIC JETS INTRODUCTION.......... o.. o c........................... 1 ACCEPTED DESCRIPTION OF JET NOISE.............................. 5 RESULTS...........-.e e.....o.o.Eooo............. 7 CONCLUSIONS................................................. 10 REFERENCES............e........ O.......o..... 15 SOUND POWER STUDIES OF MODEL SILENCERS INTRODUCTION........................................ 17 EXPERIMENTAL RESULTS WITH SILENCER............................ 18 REFERENCES. *oo o OO *O............. O *.......................... 27 ii

LIST OF SLIDES Slide Page SOUND POWER SPECTRA FROM SUBSONIC JETS 1 Air Flow System....................................... 11 2 Acoustical Instrumentation System...................... 11 3 Diffuse vs Free-Field Results.......................... 12 4 Description of Simple Jet Noise........................ 12 5 Jet Spectra; Different Nozzles......................... 13 6 Jet Spectra; Various Flow Rates........................ 13 7 Broadband Power vs Mass Flow Rate...................... 14 SOUND POWER STUDIES OF MODEL SILENCERS 1 Experimental Configuration........................... 24 2 Effect of Nozzle Extension............................. 24 3 Effect of Solid Shell................................... 25 4 Effect of Different Shells............................. 25 5 Effect of Water-Filled Shell........................... 26 6 Effect of Composite Muffler............................ 26 iii

SOUND POWER SPECTRA FROM SUBSONIC JETS

Thie experimental studies, described in thiIs paper, were motivated by a desire to understand- more fully the acoustical situation involved in silencing of jet engines during ground runup. Existing literature on this aspect of Jet no-se control is scanty and tends to neglect the physics of the silencing problem. In contrast, noise generation by ordinary jets has been studied extensively with the result that a comparatively complete physical description of noise-generation by a simple jet, unadorned with any auxiliary attachment, is available. In the present case, we desired to conduct experimental silencing research using cold model jets and to employ the radiated sound power as the measure of acoustical performance. The reverberation room method for evaluating sound power permitted convenient measurement of the radiated power with complete freedom to ignore the directionality of the noise sources, whatever they may be, so far as these measurements were concerned. The reverberation room met.od for accurate sound power measurement probably is not as familiar as tire free-field method employed either out-of-doors or in an anecoic room, However, the reverberation room method rests upon a firm theoretical and experimental foundation. A detailed justification of trve method lies outside of the scope of this paper except perhaps for the credence engendered by some data showing close agreement with free-field measurements. The experimental arrangement consisted of installing both the nozzle end of the air-flow system and the microphone end of an audio spectrometer system in a reverberation room~ Slide 1 shows schematically

the air-flow system used to operate the small nozzles. High pressure air at roughly 100 psig passed through a pressure regulating valve (Reg. 1) into a float-type flow meter. By suitably adjusting the air pressure existing at the flow meter with respect to the air temperature existing there (that is, effectively adjusting the density of the compressed air), the flow meter was riade direct reading with a full-scale value of 100 SCFM (standard cubic feet of air per minute). In this particular case, where the temperature ranged roughly from 60 to 130'F, the intermediate pressure was in th;e range of 40 to 50 psig. The directly-indicated mass rate of arY flow was used as the primary measure of the aerodynamic condition prevailing at the nozzle. Beyond the flow meter, the air passed through a second pressure regulating valve (Reg. 2) and a calming chamber, lined with acoustical absorption and flow-smoothing screens, before arriving at the interchangeable nozzle. The maximum airflow velocity in the calming chamber was less than two feet per second so that all interesting and noise-associated flow conditions occurred at or after the nozzle. The nozzle end of the settling chamber projected into t'ce reverberation room. The used air was exhausted from the reverberation room through a duct, one square foot in area and equipped with an acoustical filter. The filter served primarily to reduce the noise level in the room adjacent to the reverberation room for the comfort of the experimenters. The escape of the excess air from the reverberation room was essential in preventing short term rapid fluctuations in the barometric pressure there, which, in turn would have caused troublesome problems with respect to microphone calibration.

The acoustical instrumentation is shown schematically on Slide 2 and is straightforward in concept. A one-half inch diameter condenser microphone was selected in order to enjoy the convenience of a flat diffuse-field response from low frequencies up to the vicinity of 20 kc. The microphone calibration was checked using an absolute diffuse-field reciprocity t echnique (Reference 1) and was found to agree with the manufacturer's calibration to within a fraction of a decibel. Sound pressure measurements were taken of the jet noise in the 24 one-third octave bands whose geometric-mean frequencies ranged from 100 cps through 20 kc, The levels of the individual bands were observed on a damped rms meter. Sound power determination by the reverberation room method requires a knowledge of the decay rate of the room, band for band, and Slide 2 indicates the decay rate instrumentation also. At the bottom of Slide 2 is the formula relating sound pressure and decay rate to sound power (References 1 and 2), where:W -a acoustic power per band, watts; p2 space and time averaged sound pressure squared per band, measured with. calibrated microphone; D space and time averaged decay rate per band, measured quantity; V - volume of th;.e reverberation room; p,c = density and velocity of sound respectively for the atmosphere in th.e reverberation room; a = Sabine absorption coefficient for the reverberation room, computed from decay rate and only of second order imponrtance - Thke broadband power, of course, is computed by summing the contributions from the individual one-third octave bands.

The experimentally determined decay rates were always utilized for computing sound power. O;nly negligibly small changes in these decay rates occurred during a group of test runs as a result of the compressed air being released into the reverberation room. However, significant changes in decay rates, especially at th-e higher frequencies, were encountered over longer time intervaIctl: and were due to the normally fluctuating moisture content of thte atmosphere. The stated upper-frequency limit of 20 kc may have aroused some concern about the validity of results. Measurements in reverberation rooms for architectural acous-1-.lcai purposes are seldom extended upward as far as 10 kc. Th.e University's 5400 cubic foot reverberation room seems to provide satisfactory, although less perfect, diffusion through 20 kc. Extension of tkire measuring range to still higher frequencies could not be achieved satisfactorily Ln a room of this volume, Slide 3 shows a comparison of jet noise data obtained in the reverberation room with that obtained in an anechoic room. One of the nozzles used by W. C, Sperry Re-e.reene ) was duplcated and tested, The solid line is Sperryts partly tS-eoreti-cal, partly emperical reference curve, Thle circles are:i"-s experimental data and. the crosses are the Michigan experimental data;. 7 ke results agree almost perfectly - even the departures from.t..-e refrene e curve are duplicated.

ACCEPTED UE;SCRT=- TEON OF JET.NOISE As mentioned earlier, most of t. he definitive literature on jet noise refers to a simple subsonic jet exhausting into the surrounding atmosphere, It seemed wise, as a portion of this research program, to conduct some experiments whose results could be compared with the literature..Thi.ls literature, taken as a whkole, leads to a widely accepted description of simple subsonic jet noise. Many people have contributed to this description.; M. j, L-gkthr-illl, Alan Powell, and Ho So Ribner, to mention a few. ThIeis description, as I understand it, is summarized on Slide 4. Th.e t.ot al radiated acoustic power generated by a stationary cold ambient-air jet iso 8 d-2 -c id c< watts Wacoustie -- O f; 8dA20"5 wat' t where o poCo' respectively thKe density and velocity of sound in the ambienti air; K -- acoustical power coefflc.aient, emperically evaluated. and, generally of order; x.O'; U T: mean s-. eam v-eloci.ty +f" Y-e 3et flow; d -- diameter of +,te nozzle exi.t, Th. is equation;'s one form of t:.e wellrknown eighkthk power law and states that thke total radiated sound. power -%ar es as thte ctgtir power of tne efflux velocity and directly as th:e cross-sectional area of thLe jet. A generalized, spectrum sLhape i~s described as shown in Slide 4. It ex:h ibits a broad maximum and w'Len t;-.e frequency dependence is plotted in proportional band widt;.:s, it..as a slope of roughly +9 db per octave below and. -2 db per octave aJ:ove theF max imum. Also, the location of this

maximum with respect to frequency has been found to obey a Strouhal number relationship~ fm = SmU/d where the corresponding Strouhal number, Sm. is emperically about 0.2, These features cons- itute the widely accepted description of simple jet noise: a spectral distr-.bution whose shape is invarient -to changes in size and velocity; the location of the frequency maximum, with respect to frequency, dependent upon the ratio of velocity to diameter and, with respect to level, dependent upon U+8 and area. (References /4 and 5)

RESULTS A series of experiments were undert aken which attempted to operate within the above framework. Slide 5 presents the expected spectral curves for tree small nozzles whose areas are in the ratios of 1 o 2 o 42 all operating at the same total mass flow rate of 100 SCFMo The dots represent the experimental dCata for three corresponding smoothapproach nozzles indicated at the right of the slide. Obviously, a general agreement exists between r;e expected and the observed behavior but there are also clearly discerni'.^ e systematic differences. First, consider the uppermost curve for a one-half inch jet at 100 SCFM, corresponding roughly to Mach one (actually 1.05 based upon observed pressure ratio). Using a value of K 6' 0 x 105, a total sound power levrel of 115.5 db reference 1 x 10l12 watt would be expected.?;ht.e broadband. sound power level calculated from band measurements through 20 ke is 114.4 db; a very close agreement considering the fact that even -.gker frequency bands would be expected to contribute something to the to-,al power. Assml.ng a continu7ed -2 db. per oct-ave behavior, another 1 j db mig.W, be added to tr'.e experimentally observed, power for a grand tcot;al of 115,7 db compared wi;t-. 115..5 dbo Certainly the total power obervred for thv e one-half inch'r nozzle at, sonic velocity is in very satisfactory agreement with thie predicted value. If th'e_ peak of the expected curve is located at 6'00 cps, as suggested by thLke best visual fit to the experimental data, then the r4o..al rumber is 0.2?; also in sat sfactory agreement with an expected 7alue of aboeiut 0.20

Considering thre two larger nozzles, we would anticipate successive power reductions of 2-7 or -21 db and downward frequency shifts of 2-3/2 or 4-1/2 one-third octave bands. Experimentally, power level reductions of first 18.9 db and then 15.4 db are observed; considerably smaller than expected, The spectrum shape flattens and it becomes difficult to state explicitly what happened to the spectral peak except that it did not behave as expectedo4Three different nozzles were used to obtain the data shown in Slide 5 and consequently there is a doubt about whether these nozzles really provided comparable situations. Perhaps some of the flow parameters, something related to boundary layer (Reference 6), for example, were altered inappropriately. Therefore, the one-half inch nozzle was tested again, only this time the mass flow rate was varied. Slide 6 illustrates the results. Thke soled curves again represent the predicted behavior and the dots the corresponding experimental data. Alternately open and closed dots have been used to aid in identification of the parameter. As t.t e mass flow rate (velocity) decreases, tghe deviations between the expected and the experimental results grow until at 25 SCPM the curves would hardly be recognized as related. The deviations follow the same trends as illustrated on the previous slide (Slide 5)~ The spectrum flattens as the flow velocity decreases and proportionally more'XDro Alan Powell of U.C.L.A. suggests that the observed deviations are due to a flow separation occurring upstream of the nozzle in the calming chambero If this s s so it would change my interpretation of the observed results. This point has not been fully resolved at present.

of tithe energy appears at higher frequencies. On the basis of broadband power, agreement would appear quite good as low as 40 SCFM but the spectral data clearly evidence consistent though compensating departures at 60 SCMo A more critical examination of these broadband data on Slide 7 reveals tlhata a consistent trend existed here also. At the lower flow rates, more sound power is observed than expected on the basis of the eigr:ithb power law and the slope of the curve suggests a dependence on velocity to some power smaller than the eighth. We might expect that at low enough velocities, the sound power characteristics of jets ought to tie up with those of ventilation system grills. I have not been able to find information on grills in appropriate form for a detailed comparison however there is some confirmation of te rng spetral shape and a velocity dependence to some lower power th-an th.e eigh.th. (Reference 7 and 8)

CCN;NCLUSONS LTe experimental results reported here, taken in conjunction with existing literature on jet noise, suggest tie.at t.e usual descrip' tion of jet noise be reserved for th.le trigt* subsonic velocity range and be incorporated as a sr eoial case wit',cin a broader description of jet ncise. Suc-l a description is only vaguely perceived at present but th.ese results do provide some hl-.int of wk.at it must account for, One can even begin to visualize wthat t-e theory might contain. For example, regardless of t-h.e jet velocity _aL;r1 l-:erefore the rate of energy transport), tte shear ust outside +ure nozzle must be very large. As a consequence, an initial strong small-scale turbulence, radiating high frequencies, should be expected thereo But whrether or not powerful largescale turbulence will develop mig t depend on the amount of energy available initi ally and on th-,e competition among several dissipative mechanisms. On the experimental side, much work remains to be accomplished particularly at low and medium velocities to delineate tP: e physical bekv.:or l.early. c''.e present researc< program could not proceed furtt-er s.nce isa. pr:.p p urpose lay in a dofferent direction. It i:as h owo ever, provd-ed very sttrong evidence of some unf inis ed acoustical bus L n.ess w.ic: -,l -,opeful.ly some among us may be able to complete. -10

KiN PRESS. TEMR. l < LINING B SCREENS 0-20 PP1G. = k~ S._ ~ H ~MCOMPRESSOR AIR FLOW SYSTE M Slide 1. Air Flow System. nCOND. MIKE SPEAKER nCOND. CALNUMIKE REV. ROOM REV. ROOM / 1/3 OCT. AMP. 1/3 OCT. AUDIO AUDIO SPECTRO- I SPECTROMETER METER WARBLE OSC. /;<\~~ ~TIME INTERVAL METER RMS SLOW DB METER BAND PRESSURE LEVELS DECAY RATES ACOUSTICL IST RUMENTATION SYSTEM V )(I-e-) w= (P))(D)2LoGe Slide 2. Acoustical Instrumentation System.

-12160 Lc + MICHIGAN DATA 150 o ARMOUR DATA Z (NOZZLE No. 100, / ASD-TDR-63-326) Li a_ 140 3'o0 0 - 130 z9. am 120 cr a: 110 0 Ito O 1 2 3 4 MASS FLOW PER UNIT AREA SLUGS/(SEC*FT2) Slide 3. Diffuse vs Free-Field Results. 9.6 DO DB BELOW 10 TOTAL POWER 20 -2DB/OCT. FOR 1/3 30 - OCTAVE 30 +9 DB/OCT. BANDS 40 fm 50 I... A. I.. 1/16 1/8 1/4 1/2 I 2 4 8 16 FREQUENCY RATIO fm =SmU/d cps; Sm I 0.2 WA = K oCo U8 d2 WATTS; K - 6 x I0 Slide 4. Description of Simple Jet Noise.

-13120 O loo 100 SCFM'tT> 60 t 1 A00 n. / Z B 0 0.75 cu 90, 80o FREQUENCY C C NOZZLE Slide 5. JetSpectra;Diffe t (FULL SECTION) 70,,*,**.......,,.,, 1 C R 60 -0 - > 6 //A 0.500" 1.500" 40 1(~20~~0.7072 0.750" a: ~50~'' / ~ C 1 ~7.(00 0.375M 11 40!'SM 8' I00 2 5 1000 2 5 10,000 2 FREQUENCY CPS Slide 5. Jet Spectra; Different Nozzles. -,oor ~ 10~~-.60 0j ~~90 1'- -'50 00 0 80 R0 NOZZLE (FULL SECTION) (D~ 70 - ~dd ~d~.~ 40 25 - 25 d:0.500" >70 j ~J z 50 - 0 ac 0. 40.,,,!,,n I00 2 5 I000 2 5 (0,000 2 FREQUENCY CPS Slide 6. Jet Spectra; Various Flow Rates.

-14100 90 80 LL7 o 70 EXPERI MENTA L RESULTS + U8 60 w 507 o 40 -j c 30 2 // 60 70 80 90 00 110 120 BROADBAND POWER LEVEL DB RE 10I12 WATT Slide 7. Broadband Power vs Mass Flow Rate.

REFERENCES lo Barnett, N. Eo and Thrasher, D. B. "Pure-Tone Diffuse-Field Reciprocity Calibration of a Dynamic Microphone." J. Acoust. Soc. Am., Vol. 35 (1963), 774(A) 2. Young, Ro W. "Sabine Reverberation Equation and Sound Power Calculations, " J. Acoust. Soc. Am., Vol. 31 (1959), 912. 3 Sperry, Wo Co, Peter, Ae and Kamo, R. "Fundamental Study of Jet Noise Generation and Suppression." U.S. Air Force Report ASD-TDR-63-326, Vol. I, March 1963. 4. von Gierke, H. E. "Aircraft Noise Sources." Handbook of Noise Control, edited by C. M. Harris, McGraw-Hill, New York, Chapter 33, 1957. 5. Franken, P. A. "Jet Noise." Noise Reduction, edited by Lo L. Beranek, McGraw-Hill, New York, Ch.apter 24, 1960. 6. Cheng, Sin-I. "On the Aerodynamic Noise of a Turbulent Jet. " J. Aerospace Sciences, Vol. 28 (1961), 321. 7. Marvet, B. Ho "Experimental Study of Grille Noise Characteristics. Trans. Am. Soc. Heating, Refrig. and Air Cond. Engrs., Volo 65 t1959) 613 (Paper No. 1689). 8. Parkin, P. H. and Humphreys, H. R. Acoustics, Noise and Buildings. Faber and Faber, London (1958), 282, Figure 96.

SOUNJD POWER STTSJIES OF MODEL SILENCERS

INThrRODUCTION The experimental studies, which are the subject of this paper, represent another portion of a larger research program related to the silencing of jet engines during ground runupo The preceding paper (Reference 1) contained a descri.ptin of the experimental apparatus and the methods of measurement, The principal difference is that instead of being concerned witr. the noise generated by simple jets, here emphasis was concentrated on the acoustical effects accruing when objects are placed. in, near, or around t:he iet efflux from some particular nozzle configurationo In these studies also, the mass rate of air flow through the nozzle system constituted tI.e primary indication of a nozzle's aerodynamic condition while thf-e radiated sound power determined by the reverberation room method yielded the primary acoustical data, As before, for the purposes of these experiments, directionality of the source or sources was ignored. A rather wide variety of objects, placed in or around. the Jet flow, were investi,.- gated.:hese.Inc!luded plates, tubes, rods, wire -er en;- metal felts, and small glass beads -In all, some 680 sound power spectra were obtained, ln.y te w-ire screens and metal felts provided much`K promise of useful silencirng.

EXPXRIEW AL RESULo.TS WITH SILENCER FKom tfiLs large c`koice of material, I have selected a few results obtained with. model silencers to present. (Probably some of the other results obtained with less complex configurations represent more basic pbys-cs but th.ey are difficult, to organize into a short paper, ) Perh.-aps model silencer is a misnomer but it is used here to apply to a tubular objec+^ conf'lgurat-ion. which. at least superficially resembles some existing runup sl.encers. In this instance, however, no attempt has been made to model i~r- ^interior structure of any existing silencer. Slide 1 shows the arrangement being considered. An extension piece has been appended to the one-half inch diameter smooth-approach nozzle (Reference 1) to move th.e nozzle's exit to a more accessible locat i9on, Tr e model silencer consisted of two pieces of stainless steel tulbing mitered and soldered togethWer to produce the right-angled elbow as tilustrated Sincthe many jet engines Lave a tail-pipe diameter of aboutf; 2C0 inte wt ile %t.Wis nozzle ~is one_-alf inch in diameter, t.he appropriatfe`'inea r sca lg f actor is about'L t-o 40 and. dimensions scaled:.-. amcn.. a. ow on atle sl.I.de L..ially the model muffler consis:ed of just t'he bare stainless steel tub-ng without any internal struct+ure or acousticallyoabsorpt4ive lining~ TrIhether or not it represents. an adequate model of any real si-.lencer is debatable but th:here can?'ard.!y be any objecti.on to i.t as just a laboratory configuratiL~on wh.ich, when tested i.n a certain way, gave the results presented below.:'1+ ~,!':cwetve Ser lide 2 illustrates the acoustical consequences of addi!,?r., te n7ozzle extensri —on to th~e one~ibalf inch diameter smooth~-ampacL

nozzle at two flow rates,!.00 SCFMY anod 50 SCFMT which correspond roughly to Mach one and Macth. one:'haif respectively, T..he results for the smoothapproachi nozzle alone are drawn as a solid line while those for t.he addition of the extension are shLcwn as dots. At the highest velocity, negligible change occurred'but, at t'he lower velocity, a significant increase occurred. and so th:.e hiighber vi-ralues will have to serve as the base reference curve for thle muffl er experiments. Withb th.te extension in place, h.alving the mass flow rate on.ly reduced the broadband sound power by 19.7 db instead of the 24 d.b expected from the description of simple jets or the 23,5 db found for the smooth. approach nozzle alone. Slide 3 sh'ows th.e acoustical results for the bare solid-walled muffler compared t(o the extended nozzle alone. A tremendous increase occurred in trhe sound. power radiated at low frequencies,, At high frequencies and for t..e':.igh-est velocity flow, a small decrease occurred, Superimposed on thtse curves for th::e muffler are some bumps which are indicative of tonal generation corresponding to a pipe open at both ends. T4 ese remain fixed. n frequency as t'-e muffler is moved up or down stream but sh:-ift i.n patIaern if th'.e nozzle end of tihe muffler is closed off to yield. an acoustical pipe closed, at one end,. Tke same. general consequences occurred. for both:' flow rates although the.igh;.est flow rate condition was affected most, There are some theoretical reasons which would suggest that th.e result.s shown 4n Slide 3 were to be expected.; someth.ing along t;he line of'any solid surface in tt;he Vicnity of turbulence will enhance thie acou.stic rad stlon." In order to inves+igate this point, anotlher muffler of ident ial dimensions was fabricated from perforated sheet

nmtal. The solid area was reduced to 65 per cent by 1/8 inch diameter holes located in a triangular pattern, 0o185 inch on centers. Slide 4 illustrates the results. Th'r e solid curve is the reference curve for the extended nozzle alone. The dashed-curve is for the solid muffler and duplicates the results from the previous slide. The solid dots represent; the acoustical results ifor the perforated msuffler shell. it can be seen that these results lie intermediate between the extremes for the bare nozzle and for the solid s'helli In order to investigate the role of absorption without introducing acoustically opaque surfaces, th-e perforated muffler shell was wrapped on the outside with a layer of fine fiberglass. This fiberglass had a very thin coating of neoprene (I think) on the outside surface onlyo The acoustical results for this condition are also indicated on Slide 4 as open circles. This fiberglass (some that was lying around thJ.e laboratory and Ir h-ave lost track of Its original designation) would be expected to show appreciable absorption and transmission loss at high frequencies but negligible absorption and transmission loss at low frequencies,,Reference 2). Tnde.d, tLese expectations are borne out Reduced radiation is observed at h.ighr frequencies but the low-frequency behavior is reminiscent of a; larger percentage of solid surface. T do not think that the muffler walls are resonating to any appreciable extent,; The increased low-frequency radiation is both at tFo low a. frequency and too broadly distributed for such to be th're case, T`i;Ls- aspect of the problem can be investigated in another way, The solid muffler?~ad impervious, fairly rigid, massive walls. Mr, Melvin Roquemore of te UD So Ai.r Force s Systems Engineering Group, who inherited project

monitorship of this research, supplied a flexible-walled model muffler having approximately the same interior dimensions as the metal shells. It consisted of a double-walled structure of rubber-like material and was inflated with water. When filled with water, the model weighed about 17 51 kg. Thus it would seem to constitute an impervious, limps walled massive structure. The acoustical results are shown on Slide 5 as open and closed dots and are essentially identical to those for the solid muffler (dashed curve). Only the detail of the low-frequency tones has shifted somewhat. It would seem that if one attempts to fabricate a jet muffler using a solid shell, the first consequence is a large enhancement of especially the low frequency radiated sound power. Superimposed on this is the tonal generation to be expected for that particular geometry of enclosure. Possibly wall structural-resonances also will be superimposed if they occur but such. have not been identified in this research. Then if one adds acoustical absorption inside the muffler shell, some reduction from t'he new higher base line of noise is to be expected. If one is clever enough, and uses effective enough absorbing materials, some real reduction in sound power below tih.at for the bare jet itself may accrue particularly at high frequencies but tote low-frequency power is apt to remain larger than without a muffler shell0 I have one more slide to present. In tf, is case, we started with the idea of introducing screens and metal felts wtile avoiding the introduction of solid surfaces in so far as possible. That was tlhe starting point, however, Mr. Philip Kessel, who was conducting these particular measurements, began to make improvisations on the original

theme using tt= re screens metal felts and other objects associated with our experimental program, Slide 6 shows the results obtained with one of his conglomerations. 7~t did not raise the backpressure at the nozzle. T"..e radiated sound power was reduced by an appreciable amount at all frequencies- The broadband power was reduced by 39 db or to about llO000 of its original magnliu.de wt-Ale some individual bands were re~ duced by 44 dbo I do not think that tI.'is combination of materials is necessarily anywnere near optimummo Too little researchK. time remained to take this combination of objects apart and:.l- attempt to optimize the composite for eachn addition. ThI"Ie origna l idea of avoiding solid walls got lost somewhere and the rising trend at the upper frequency limit is slightly disturbing alt~rloug.- presumably it could be controlled with conventional absorption. Probably a more compact geometry is possible withlout compromLse of tkhe acoustical performance although scaled up 40 times, it would only be 13 feet in diameter by 8 feet longo JutA woty and'tow t;his mulfiler operates is not at all evident. By t;at mean, we d'- not posse's a Ce+adi —:e 1 p. yial underst-anding of the ailencing action of certain screens, metal felts, and otr~er objects placed in the jet flow; mucr less understand.; ow a combinatrion of suchV obj cts functions or?:Kow t-ey should be combI.ned rationally to meet, a specific silencing criterion~ Neverthkeless, the results shown on SAlide 6 demonstrate beyond ques-ion that significant sound power reduction can be achi.eved. Th<at 2 prys7 al pro eses extst by wmcta large reductions can be actsieved. ~I;>Lee re ul s aire orders sf magnirttude more effective than t~he resultjs for aony sahrlgle screen or metal felt tested; t*..us it~ appears that some

form of casc-ding Is app.licable. Hfowever, these results with small cold nozzles merely affirm the possibility of compact runup silencers having large effect. th tey do not provide a fully-engineered operational design. A considerable amount of basic research needs to be done to fully urderstand the pti ysics of the acoustical processes associated with flutd floTwss and much development engineering rests thereupon to realize optimized, industrially-fabricated. runup silencers.

lo DIAM. NOZZLE (20 ) = 0.062" WALL 3"O.D.(10O) ~( i(~-~"6 I I (l 14" (46.6') CROSS-SECTION OF NOZZLE AND MUFFLER SHELL "SCALE" 1: 40 Slide 1. Experimental Configuration. 120,.. -,. 100 SCFM I10 100 SCFM 3 100 - 50 SCFM - 90 0 0.500" DIAM. 8 80 SMOOTH APPROACH Jr 50 SCFM NOZZLE.:: 70 - ~ NOZZLE WITH 2 LONG TUBULAR,-J 60 _'. EXTENSION. W / > 60z;~ C- 50 - Im 0o 0 lo 40- 5 m 100 2 5 1000 2 5 10,000 2 FREQUENCY CPS Slide 2. Effect of Nozzle Extension.

-25120 - i00 SCFM HF~~ ~ 100 SCFM 50 SCFM cIQ - 90 o - 1/2 DIAM. NOZZLE WITH 2 EXTENSION. LLJ 80 50 SCFM LU 8 /O/oMO MUFFLER SHELL SOLID. c 70 110 o > 60 100 2 5 1000 2 5 10,000 2 FREQUENCY CPS Slide 3. Effect of Solid Shell. 120 110 H~-; - I/2 DIAM. NOZZLE Ho 1 ~ o A ~ -' WITH 2' EXTENSION. 100 (/" / o " -- MUFFLER SHELL SOLID. - o 90 7 wuo 0 * MUFFLER SHELL Mn PERFORATED. Q70 I > 60 z 0 MUFFLER SHELL 5 ABSORPTIVE COVER. o 100 2 5 1000 2 5 10,000 2 FREQUENCY CPS Slide 4. Effect of Different Shells.

-26120 " 100 SCFM I10 - * 100 SCFM.. I -I/2" DIAM. NOZZLE' f_.f. ~ -! WITH 2" EXTENSION. Q< 100 / f J', = * * ~ 50 SCFM C 90 / /' X RUBBER MUFFLER 59 ~~0 9~z Slide 5 Effect of Water-Filled Shell. /,...,......,.. O 70 120 1/2" SMOOTH APPROACH 110- 100 SCFM NOZZLE ALONE I- - 0 / 39 DB / x-O MESH < //'S CREE t soW * 1 <... ~ o,, 0g50o LU,, 4 0- i. o..,. 100 2 5 1000 2 5 10,000 2 - FREQUENCY CPS METAL FELTS Slide. Effect of CoposWaterFilled Shell. 120 60 ci FREQUENCY CPS METAL FELTS Slide 6. Effect of Composite Muffler.

REFERENCES 1. Barnett, N. E. "Sound Power Spectra from Subsonic Jets. " Paper R5 presented at the October 1964 meeting of the Acoustical Society of America in Austin, Texas. 2. Geiger, PO H. and Hamme, Ro N. "Sound Transmission through Aircraft Soundproofing Structures." University of Michigan, Engineering Research Institute report on Project M820 to the Air Material Command, Wright Field, dated October 27, 1949. -27

UNIVERSITY OF MICHIGAN 1III I I 1 1111111111111115111111111111 13 3 9015 02227 1392