BASIC PRINCIPLES OF AERODYNAMIC NOISE GENERATION D. G, Criarrtox Departmen of Applied Mathematical Staies, Universi of Leeds, Leds, England Abernct—This paper gives a simple, unifed. analytical description of a wide range of mechasivons assoclated with the generation of sound Dy Unsteady Muid motion. Topics teste 1 ‘rom compact and aomcompact multipole sources, Lighthill's theary of sound omission ide raison ‘rom free surtwence, effects of source eotvetton,soutd generation irom Mow interaction Wih soli surfaces And inhomogeneties of the medium, and cingular pertrhation aspects of the arrodjnamce sound hoise penerston Es fxeessenoie problem. ordenty stmcte of us bulent ies AERODYNAMIC noise has ¢ history of about 20 othas.a pressing orcblem of technology and as a reputable subject for analytical study. In the ‘arly fifties, Lighthill wa: stimulated by anticipa tion of large-scale commercial jet sir to formulate his successful theory of jet noise whicl has stood the tesis of tine and of far-reaching extension. Flow noise has subsequently become a matter of serious military concern in the operation and detection of ships and submarines, and the aeronautical and naval applications of fow-noise theory between them cover a wide range of inter esting phenomena, None the less, there was some Stagnation in the subject in the later fifties. and most of the significant advances in our understand. ing have come within the last 10 veers. Mans of these resent developments form a large part of this ‘At the outset, itis worth saying exactly what we fow noise to achieve, concerned essentially with by-products of fluid lows whose internal dynamics may generally be regarded as unmodified by the by-product— Which is a small energy drain in the form of acoustic oF structural waves, The hope then is that the flow dynamics wil be more or les decoupled from the wave motion, so that we can attempt to Predict the wave motion from an assumed know: ledge of the flow itself. Now. the prediction and measurement of quantities defining a turbulent ow of, say. the jet or boundary-layer type. s still in its infancy, so that a successful theory of by-products must be formulated in such @ way as to make only minimal demands on the Fow specification. On the other hand, lack of knowledge of features of the ow does not entitle us to make assumptions and approximations except with the greatest care, We are dealing with a small by-product, and approxi mations which may be quite safe as faras the flow dynamics are concerned may be fatal for the by Droduct. Vast erors may easily be incurred in this ‘way through apparently harmless approximations, a tion discusses several area of current interest and imporance. including sonic shear ayers. shallow waver wave simulation of flow noi, the ithe general issues bound up with shear layer and jet Insiablty and the especially in underwater flows, where the Mach ‘numbers occurring are always extremely small, and the acoustic energy is a correspondingly minute fraction of the total low energy: The aim of a theory must be io predict gross features of the induced wave motion in terms of, easily estimated parameters of the flow. Highly accurate numerical predictions are simply not 10 be expected—e reflection of the fect that noise levels usually vary fer more widely than do the parame ters which describe the noise-producing 4ow. For ‘example. if the exhaust speed of azurbojet engine i merely doubled, the noise power emitted is in- creased by 24 decibels, i.e. bya factor of 250. Asa ccond example (ef. Section 8 below, the addition cof small air bubbles ina volume concentration of as litle as 1% to s turbulent flow in water can increase the radiated noise power by 2 factor of order 10° Clearly in suc vis far more essen- tial to obtain the correct functional laws for these effects than to pay attention to numerical coeff cients (which anyway are very rarely believed to hhave universal values in the aerodynamic noise ficid). None theless, experience seems to show that errors involving more than a factor of 10 are not incurred if sufficient cae is taken, and that, in any cease, dimensional trends and scaling laws ore of enough value in themselves, in enabling data to be coordinated and experiments to be designed and interpreted, to justify a theory of acrodynamic ‘Our sim in this paper isto give a simple unified analysis of the way in which sound is generated by unsteady fluid flow. In the main we leave aside the way in which sound propagates as sound around iffracting obstacles and through refracting and scattering media. Such a separation into generation and propagation aspects is easily made at an elementary level, and is convenient here, though there is nothing fundamental in it. Generation and propagation aspects are often inextricably coupled, a fact which is becoming increasingly widely recog. nized as ideas from other branches of acoustics, land fluid mechanics generally, find a place in2 DG. crranron ‘aerodynamic noise. For instance, sound propaga tion (“refraction”) through a shear layer is aot necessarily an energy-conserving process. In- stabilities on the shear layer may be triggered by the sound wave into releasing additional acoustic ‘enerey-—a process which has been observed experi ‘mentally both in the case of the jet exhaust shear layer (Crow," Ronneberger®) and In the case of flow through 2 duct (Dean), For the moment, however, we Jeave aside these situations in which the flow and its sound field aze in a delicate balance, and look first atthe Lighthill theory"* ina simplilee manner, deriving the basic resulis, of Which one is well known, the others Jess so, The application of these basic results is to the scoustie power radiated by free turbulent flows in the absence of boundaries. The internal evolution fof the Turbulent eddies will be supposed to be entirely unmodified by compressibility effects, and at first st will be supposed that there is no mean ‘locity field, Eects of eddy convection by a mean flow field will than be seen to be decoupled from those arising {rom the internal ugitatlon of the eddies themselves, In accordance with what was ssumed above, the Mach number of the turbulent eddies will be taken as small, but we shall Jook at oth high and low Mach number mean flows. Now. many flows of practical interest do net involve jus: & homogeneous fluid flowing in region devoié of boundaries. It is then necessary 10 con sider in general terms what the affects of in homogeneities in the low might be, and then try bY detailed celeuletion in & few special cases to sex whether its possible to give a reliable and reason bly general classification of these effects. Mu recent work has bean done along these lines, both for the case where the inhomogeneity takes 1 form of a solic pocy, and for the case where It is ei by bubbles or vortices in underwater flow "The introduction of the methods and ideas of diffraction theory has proved to baan essential step jn tinderstanging the way in which solid surfaces ‘affect the sound generated by nearby turbulence. Here, of course, we are not concemod with the generation of sound through the preseribed motion ff solid surfaces. which is the basic problem of Classical acowsties. but with the potential which # purely passice surface has for converting local hydrodynamic energy into sound. This has now been recognized as a dominant sound-generation mechanism in a wide variety of technologically important contexts. Tt will also De worthwhile to look briefly at an approach to the problem of asrodynamic noise Which is rather more fondamental than that of Lighthil, This approach is based on singular pertur- bation expansions, and seeks to determine the flow dynamics and the emitted sound field as 1wo largely distinc, but overlapping, aspects of the complete compressible flow problem. Few new results have ‘yet emerged from this approach. through one geis.a penetrating insight into familia results, and 2 salut- ary reminder of Just how restrictive are the condi- sions under which we have any theory of aerodynamic noise, ‘The paper is concluded with « discussion of recent problem areas and specilative and con troversial issues which seem likely to hoid the key to further progiess. In acoustics, we are concerned with small amp- litude oscillations of compressible uid about « stale of rest, The basic equations of acoustics follow from linarizing the full equations of mass fané momentum conservation and the equation of state, In general, its adequate (0 assume the fuid ta be Newionian and to neglect all moiecular relax ation and éifusion effects, so that the stress on any fluid element consists simply of a normal pressure, Then, the mass ang asomentum equations can be in tensor form 25 Neglect of molecular diffusion implies that the ceatropy of a fuid element remains constant, and therefore thet the pressure p is a function of density alone, We thus have a differential relation dp =a°dp, where a is called the local sound speed Now, we linearize (2.1) end @.2) about a state of rest in which the fluid has uniform density ps and sound speed ay, Thus. 1 je approximations 224 M0, aa on Elimination of u shows that p (and also p.m) isfies the homogenecus wave equation Bi 2a |oacp=0 area aie tiec anymore at =o a k=UN a per ean nea ee ee fopmelos ones) ic cud Yes reas A eect re! Writing exp i(le-x + wr) as exp iktke+x + ant) shows PE vine same pedo pn ae More complicated so up from plane waves b [ives is also a solution of (2.5) Sumrteg fray By choosing F(R) su integral, we con genera at exp! where 3 =[p) is dist Likewise, in 1wo HEM(&x) exp ikast) symmetric, and desert verging on some kind wave diverging from sign: for the coeffcien resented by and Hike) HBGs): The solutions contaia factor, an x" or » tion for spreading in th asymptotics of the Har Hix) ~ Hx) ~ accurate for x >4 esse In much of what foll transforms in i, which factor exp (~ lat) for frequency w is regar frequency correspon wavenumber magnitud uA plane wave travell represented by exp i(k metric waves diverging by expi(he: ~at) There is an importanar results, and a salut- trictive are the condi- ave any theory of with a ¢iscussion of speculative and con- Likely to hold the key sed with small amp- ressible uid about a ull equations of mess 1 and the equation of te to assume the uid et all moleealar relax. thatthe stress on any of a romnal pressure. jum € cions can be en sion implies that the ‘emains constant, and Dis a function of differential relation the loral sound speed. (2.2) about state of niform dansity py and ustic approximations ew stp dake pm) ave equstioa| Oe em sions of the form ke and k =k. At ‘e solutions is constant at, ie. over any plane ector k. The phase is rormal 19k provided the wavelength 27/h, present plane waves wave fronts normal t bby the unit vector fe cp ik(E-X= ast) shows, ‘agate in the directions day regardless of | i More complicated solutions of (2.5) may be built up from plane waves by Fourier integrals. Thus, [Fore itenshassan Seana) for any angio tine OD By choosing Fik) suitably and performing the integral. we can generate solutions of the type exp ithe = hast), where x=|s| is distance in three dimensions. ikewise, in two dimensions, we obiain Hj? (kx) exp (= ikast). Those solutions are radially symmetric, and describe waves diverging or con: verging on some kind of singularity at x =0. A wave diverging from x=0 must have opposite signs for the coaffcionts of x and f, and is rep resented by r= aut) 6D HIs\ex) exp (~ hat) Hh) exp (* Ket). ep) The solutions contain, in addition to the phase factor, an x or x" required by exergy conserva. lon for spreading in three or two dimensions. (The asymptotics of the Hankel functions are [ [2] accurate for x =4 essentially) In much of what follows, we shall take Fourier transforms in ¢, which amounts 1o assuming a time factor exp(—lut) for the whole fleld. Thus, the frequency w is regarded as fixed, and to that frequency corresponds a single acoustic wavenumber magnitude wlan, which we denote by 4 A plane wave travelling in the x-direction is then represented by exp i(ker ~ wt) while radially syra- metric waves diverging at infinity are represented by HM) ~ exp ior ~ on) or exp (= fut Ha) ‘There is an important relation between the pres Sure, density and velocity in a plane wave. Suppose the wave is travelling in the + z-direction: then , meRe Aci, p=ReA D=Reaide™™, uema0 tnd s0 6 Basic Principles of Aerodynamic Novse Generation 3 The average rate at which energy is transported across unit area normal to the propagation direction is defined to be the incensity in the wave, T= (pu)= 269 pm)~ $8 D and is proportional to the average value of the square of the density Muctuation, The formulae (2.6) and (2.7) may be applied directly to waves spreading in two or three dimen sions. In three dimensions, far example. we have from (23) and 2.4), p= ReAte+0e7 ee" velocity. The leading terms eatisty the pl relationships (2.6), and give an intensity tional to x” The result of integrating this contribu- tion over a large spherical surface of radius x is then a finite number indopendent of x. while the correction terms O(x") make 2 co of Tt order x” and thus no contribution ta the toral ower radiated across a distant surface. The terms of order x”! quoted above are suficieat to give the ekaet valle for the total radiated power, We have seen some of the properties of solutions fof the homogeneous wave equation (2.5). In the sequel, we shall be concemed with this equation when the right-hand side is non-zero, A way of achieving this within ordinary linear acoustics is to suppose that sources of mass are cistributed ‘Throughout the uid, injecting fuid mass at a rate (Qtx, 1) per uult volume: The efect to aid a term Qtx, 1) to the right side of 2.1) or (2.2), and to make the wave equation (2.5) red 30 a oo ew Two akernative forms of this equation are some: times used. We may recognize that Q is ise, by detnition, a tine derivative aW/3t say’ and write the right side of 28) as ¢°W/o"", Or we may regard the besic noise generating function as @ = 9Q/9% rather than Q, and write the right side of 2) simply as q. The tree forms of (.8) dior only in the number of time derivatives which are expictly displayed on the right side, and corzespord (0 adopting either the source volume displacement (Wipa, of the source volune velocity (Wlps = Jo), oF the source vohime acceleration (Wi q/po) as the prescribed source function. Of course, iT one of these functions were known exactly the others would be also, and all forms of (2.8) would Te problems got Known in detail, and we know only their iocnsional variation “with the most important Teresi he source functions ares D.G. Crcsrox ‘featires of the Source functions must be brought Be Seeger Appendix. If G(x, 1) satisfies 'G = 6(x)8(t),then oe=4 5 2 e=aeG , 4 raat) de) A, eo pct Which is usually referred to as a “retarded poten tial” solution Each volume element dy makes) contribution decaying as the Inverse of distanse Travelled, and thgfensity Auctuations follow those ‘of the source termh Skactn—except for atime delay x= N/a. jy sme wave he ime taken for an acousts wave Sbeger ae Tet us now’ make an estimate of the sound fd generated by this source distrbution, sssuning, with 2 view to later apoicaions in ow nore ‘eof that g varies eppreciably oul over some Seven eee eee SorNenien arte in term of avelociy, I The ‘eto need noe ese pope Teal compared withthe sound speed ds, tht there Is ‘no_upper_timi: _on_the typical frequency _u/T We stiplate, owever, that com 15 are not to enter ino the deserp- tion of g, so thet en eimersional sounds we mast hve a propertonalty path This step cal order of magnitude of the source in (2.9), ‘ands istohe integrated over a source volume of order Some care fs necessary here, for at each pointy the source strength is to be evaluated at a diferent time, and we rust decide whether thsi an impor: ‘tant feature or not. Now, the maximum difference in emission times over a volume [is of order I/ay: ‘LLthis ia small compared with the time Liu_over ‘which the source field chang emission tine ste unimportant. If th ie. iLthe Mach number_m= jae ¢i—then all ‘emission times 1— x= w/a: may be replaced by the same time 1 Jas. with x measured from any convenient ongin Within the source region. A further simplification is possible if x is large; for ‘then, inaccordance with what was ssi earlier. only tems OG) are important, the dependence of [xy] upon y is irrelevant, and we get down to 5 a otun~ztx J a(se-2 = dy, 19) ‘and provided the instantaneous volume integral of q is non-zero we may estimate mais "peter ean "3 es the case m 21, we For. ar (20) show the sounds detenined OFT ‘ime variations of the source. These take place with f typical frequency u/l, which will also te the frequency of the emitted sound, which therefore hhas wavelength cal/u = ie”. When m <1, 1 source field, of typical extent Jie compact relative to the radiated wavelength fm”. Note that we have got this result by integrating a &-function over time + to get (2.9), and this is useful when m , it is clear that differences in re- tarded time over a distance | are crucial, for the source oscillates through many periods !/ during the retarded time diference I/a, acz0ss the source scale | Again, f q were known exacily this would make no difference provided the integral in (28) could be evaluated exactly. That is not usually the case, however, and so we must manipulate (29) into & form in which retarded time diferences (and the resulting cancellation of contributions to the integra) are not an important feature. To do this, wwe return to the full expression for p(ay), simp” lifled only by neglect of terms smaller than x”, ) XB{k-sl—at—2)}de de @.12) (The integration limits are always (~=,+2), but no contribution ever arises from values of + greater than t because the Green's function is made to fy causality condition.) This time, we shall integrate over y, instead of , introducing coordi nates y inthe ditection x, and two coordinates y, in the plane perpendicular to x. Then as a=, = +4042) ea x8{e-adt—2)—y) dydrsdr (2.13) Retarded time differences have now been trans ferred to differences in the space coordinate y- But these differences are now, when m > I. as neslisi- ble as were the time differences when m <1. For the space difference ayr has, over the characteristic time !/u, & maximum velue of order In”, which i ssmall compared with 1 varies. Consequentty, t negligible, and ple, 1) = (4rauxy" which shows the dens spatial structure ofthe direction, Proviged we in 2.14) to vanish we » or for m>1. This is re non-compact case. The hhave the samme spatial ve except for the spatial wavelength and the sow (On the other hand, the such larger than the so might speak of a tem) when > For cases m = O(1) compact and non-comp: nor (2.15) will hold, since for time inevitably play « ure then isto integrate t exampie) in the directicr Js normal co the curves o constant argument. By 1 for variables atthe expe p~a(t Which is uniformly valid and (2.15) in approprist ‘emerge later, in Section 5 used to discuss convecti mean velocity U ‘The reader who is not sional arguments may at concrete example. The f is representative of the } although any other func i! and|=| y+ Ox) ay -y)dedydr 2.13) ~ ant + 007, 32.7) dy, dr 2s have now been trans- 2 space coordinate y. But "when m > 1, ae negligi- ences when m Jes m = O(1) intermediate between the and non-compact extremes neither (2-11) or 2.18) will hold, since differences in either space or time inevitably play a delicate part. The proce dure then isto integrate the éfunction tin (2.13), for example) in the direction in the (47) plane which is normal to the curves on which the Sfunction has sament. By not favouring either the or variables atthe expense of the other we can get 16) Which is uniformly valid in m, and repreduces (2.11) ‘and (2.15) in appropriate limits. This result will ‘emerge later, in Section 5, where the method will be used to diseuss convection of the source field by a mean velocity U. The reader who is not convinced by these dimen sional arguments may at this point like to look at a conereie example. The function alonyn7) = 2H exp {—¥ [s representative of the kind of source envisaged although any other function, of order put'l” for [p|-<1 and |r|-i/u, small outside these ranges, ond ‘With no integral conservation property such 3s Sady=0, will lead to similar (but less elegant) results. For this source the distant fleld ete) [a(yyet Jaa Jtd) oven can be evaluated exactly t0 give (1m! renee )Trm w= zea) soo Pe Thus, when me 1 confirming the result 2.1!) and the idea that the temporal variation of the scurce is reflected in the temporal variation of the sound a time x/ay later When m > 1 -a}, Im agreement with (2:15), and showing how the source and the sound now have the same spatial ‘Suppose now that we regard the source volume velocity, rather than acceleration. as prescribed. To be more precise, we recognize that @=aQ/ct, where Q varies over a length scale { and a time seale I/u. Within those length and time seales Q is of order putll, while at much greater lengths and times Q vanishes, Then when m < 1, formula Q.11 continues 0 describe the radiation field without modification, as is easly seen. When m © 1, on the ‘other hand, the presence of the time derivative in ¢ a significant effect. For if one argues that differences ay in the y, variable are unimportant when m >I, one gets, from (2.14), Geet e~rm )dysde ein Daas] 0 a ge EE Vee 3 ox = easy [ aGs xia-olt=r)=whdidade 2 using the property (A.8) of convolution products to take the 3/a! off Q and outside the integral Now wwe can approximate as in (2.1 es pte.t)= Gmassy S [QC ~o e) In which tis important to note that here 3/4 is not equivalent to multiplication by the typical source frequency ul, but by the typical sound frequency Jl For the 3 :t acts on Q only through the space = ast, and 50 220)6 DG. Casanrow Tn an exactly similar way we find p~po(4) m 221) when m > i we regard the mass displacerent W 2s specified. The differences betneen (215), 229) tnd @.2l)arse from diferent assumptions as tothe rhysial nate of the source—whether one con trols W, W or W (resrestvey). Inthe compact cose there is 10 diferenee between the results, because the sound field controlled bythe source frequency ull Tn the non-compact cas, iti the cand frequency al which matters, andi spon sinl forthe factor m diference between (213), (2.20) and @21), corresponding a each stage tothe Tecopniion of mors Ene derivatives inthe source fonction. The radiated felis are of course independent of direction in ll cases, We have called the cave of prescribed q the “simple source” fel; that as fociated with prescribed Q is sually referred to as monopole” sce fed. Physical soures in fie mechanics generally are mos! closely modelee BS {he monopole «ict type of source ‘We have seen how the presence of time deriva: tives in the source functions greatly affects the found leven the non-compect sas, hog not a ail imthe compact cxse. We shall ee next how the presence of space derivatives in the source fen thon greatly affects the radian in the compact lim, but not in the aop-compact. The arguments are exactly se presented ave, bat with a reversal Of the ros of space. y. and time = Consider the effect of superimposing a negacive source distribution g(a) on the Bel, but with tech negative source imlaced a vector distance A fromthe soresponding postive one. Then the right site of (8) becomes aa.e-an=3.0= ) Letting A=0. a2, so that Aq stays finite and equal to F, say, the right side tends to ~9F;/@x Suchs distribution of sourees which cance in equal fand opposite pairs is called a dipole distribution, fangs characterized by the appearance of 2 single space divergence on the right side of (2.8). Mass is clearly conserved in a dipole fie, for the total mass source strength is [Bar fanas if F venishes outside some finite region. However, the dipole type of source i exactly equivalent to an extemal force distribution acting on the fuid, and the foree per unit volume is precisely F. For, if such tan external force acts on the fuid, (2.2) and (24) have an additional term Fon the right, and the ‘wave equation reads as stated above ‘Actually, it iS not necessary to take the limit 8-0 too strictly; all that is required is that A be femall compared with the wavelength of the radiated sound. Thus, if A€A, two equal and opposite sources are equivalent 10 2 dipole of strength q, and that is equivalent to a point force GA applied to the fluid. ‘Next we estimate the field radicted by a dipole, characterized as before by the scales 1, | (and, of Course, the ambient deasity ps). The typical mag- nitude of Fis pot’ implying naturally that Ais of ‘order 1. The selution to the wave equation can be ‘ritien in several different forms. For example, we right write Se Gnat) ay, |? pls, t) =— pltst)— ‘sing the properties of convolution products. Both ive identical results if evaluated exactly, but dit- ferent rough estimates if care is not taken. For ‘example, we can simplify (24) as previously to get the density fuetuation of a compact dipole dis tribution in the form Fras } oy end, if we casually estimate the integral as put eee 1 we get pls. t)~ po 2s jn (2.11), However, @ gross overestimate has ‘been made here, for i retarded-time differences are neglected, thea ae by the divergence theorem. The “gross overesti- mate” has completely failed to take account of the almost complete cancellation of contributions to (@.24);thas estimated the dipole type of integral as equal to the value pou” of the positive parts of the ipole, muhiplied by the integration volume P, and this inevitably produces the expression (2.11) for the simple source, rather than dipole radiation Thus, if we use ©.24), retarded:-time differences must evidently play 2 critical ole in determining the true value of the integral, and since we can only rake a general estimate of the functions involved, (2.24) jean unsuitable description for our purposes. Equation (2.25), on th from such a delicate bal so that say, where (6, 6s) € the vector x in the obs snd there i, in general here should vanish. Car mates nee plx.t) for the field radiated by tion. Note that the line that of 2 monopole by by Om’). Note also the the compact dipole. Thi value (at given s)ia the and drape to 2ero at rig the opposing sources server and so cancel © Fic. 1. Direct Next, consider the ne shall see now that (2.2: Fer. without neglectin, but only working to 0 Retarded-time difteren S0 we introduce # 8 ps.) afan scary to take the limit is required is that A be wavelength A of the AA, two equal and iwalent to a dipole of sivaleat to point force adioted by a dipole, ¢ scales w I (and, of y go). The typical mag ing naturally that 4 is of 2 wave equation can be forms, For example, we yt —lienl) ae Volusion products. Both Juated exactly, but cif- care is not taken. For 24) as previously to get compact dipole dis- imate bas led-time differences are <)dy-0 ‘The “gross overesti- 4 10 take account of the ‘on of contributions to pole type of integral as the positive parts of the egration volume I’, and expression 11) for than dipole radiation. tarded-time differences cal role in determining ‘and since we can only the functions involved, piioa for our purposes. Basie Principles of Acrodyasmie Noise Generation a Equation 2.25), on the other hand does not sur from such a delicate balance. Retaining only terms O(x"), (2.28) gives, so that say, where (B,, Bs fs) are the direction cosines of the vector w in the observation direction. Hence. Dear Peegee nr cates 26 and there is, in general, no reason why the integral here should vanish, Consequently, making the esti> (a ptx.t)—p(2)m’, e for the field radiated by a compact dipole distri tion, Note that the linear eld here is weaker than that of 2 monopole by a factor O(n}, the intensity by O(n). Note also the characteristic directivity of the compact dipole. The radiation has its maximum value (at given +) in the direction of the dipole axis and drops to 2ero at right angles to this axis where the opposing sources are equidistant froma the ob- server and so cancel completely (see Fig. 1). Oe ic. 1, Directional distribution of inensitv fiom a dipole To 608 & Next, consider the non-compact case, m > 1. We shall see now that (2.28) is the unsuitable formula For, without neglecting retarded-time differences, but only working to Ox) (2.25) gives 1 f2F(, =v ac 3 (tS) ay Retarded-time differences are crucial when m > 1 $0 we introduce a S-function as in (2.13) to get PGs, t) = {x — at ~widydydr Hore, we have just followed the steps leading to (@-14) for the case m > 1. Again. we can argue that the space differences a.r are unimportant when ‘m >, but then the integral becomes as [Bara con integration over 7. Space dilferences therefore ‘must be taken into account, but we cannot do this ‘with only 8 rough idea of F. Thus, we see that (2.25) is suitable when m-< J, but quite unsuitable when m > |, However, (2.28). which failed when m <1, is now suitabl in that it leads to a form susceptible to rough astimation. Introdueing the 5-function and integrating over y, yielés ay. does not 1 any rate, al ponents of F and neglect of the space differes now lead to a zero integral—or though the two terms involving © perpendicular to direction x int are stl leit with a, post Bu-aty.7 This is the estimate of the field produced by a non-compoct dipole distribution. It is identical with that for the non-compact simple scurce distribu tion, as given in (2.15). The reason is simply that when m <1, the opposing sources of the dipole are close together compared with the wavelength. so that they only fail to cancel altogether because of Slight differences in the times at which they must emit in order to be heard simultaneously at x. This neatly complete cancellation is represented by th factor m differeace between (2.11) and 215). On the other hand, as m increases, the source scale ! land wavelength A become increasingly compara bie. so that cancellation between opposing sources of the dipole becomes less important. When opps ing sourees are not close together on a wavelength seale, they are heard more or less independently without large-scale cancellation in all directions (though, of course. in particular directions. they ‘may Interfere constructively or destructively de- pending on precise details of the phases). Conse quently, there is no reason to expect the Feld radiated by a non-compact simple source to differ from that of a non-compact dipole. so that there is no value in emphasizing the multipole nature of a source region if it is not compact. We could go on from a dipole to construct ‘multipoles of increasing complexity, but such mul tipoles have never played a role in classical acous-8 D.G. CmiaxTox tics, They do, however, play a central part in Lighthill's theory of acrodynamie noise, and will be described in that context. Liga eso tres it oe in drawing an analogy between the f hier fw prOHen-and th ea tery of cs ‘al acoustics. The analogy is etact, and itis essen- {ial that no approximations be made unt the exact {orm of the analogy is established. We consider {hud in which a finke region only contains a non steady fluctuating flow. Far {rom that region, We suppose the ud to be uniform, with density py and sound speed ax, and to be at rest apart from the sll motions induced by the passage of sound waves generated by the unstesdy flow [Now, the exsct eastions expressing consery tion of mass and momentum ia & fuld aie provided there ar no external sees of mass and 10 extemal forces acting on the uid. py is the compressive stress tensor. The vector pany is the force exerted across unt area with unt normal n by the uid on the side to which m points on the aid on the other side, Ifthe Suid is inviscid. p, ~ pi 1 the stress rechces toan isotropic pressure For Bui Newtonian ud Fy SPH = PA Duley —Seud), 63) where is the coefficient of viscosity, ¢y isthe rate of strain tensor, = au, aw] tex “an and snd(@2) can be written essentielly in the fo the eivtins 2.) and.) of linear aconstis, by 4s the sum of an “acoustic stress" on 69) Ti = pu +P. ~ ab oo) These are just the equations. of motion of a ficiti- ‘ous acoustic medium acted upon by an extemal stress_distribution Ty(x.1). If the real fuid, in which highly nen-near motions may o:cur, is replaced by a fictitious acoustic medium, in which coaly small amplitude linear motions occur, and if this medium is acted upon by an external stress system Tox,t), then exactly the same density fluctuations will be produced. This is the basis of Lighthil's theory; if we know Ty in some useful sense, the problem of sound generation by uid flows is reduced to that of solving « problem of Toniveational acoustics. Exactly what that problem is follows by eliminating ou. between G.4) and G.5) li high Lichhilsinbomozencous wav eataton_ ‘The (rms on the right contein two space deriva tives, and, therefore, represent neither a monopole nor s dipole distribution, In fact they represent a quadrupole distribution, derived from dipoles in exacily the same way es the dipole from a monopole or a “simple source”, Each limiting recess on opposing sources involving" space Alerivatives leads to a source field invalving (a + 1) space derivatives. A quadrupole is formed by tak- ing the limit of two adjacent equal and opposite dipoles. Thus, there are two characteristic direc- tions associated with a quadrupole—one the direc. Tion of the individual dipoles, the otber the direc- tion in which the dipoles are separated. Since there are three components of each direction, there are nine possitle independent crientations of the quad- rupole axes. These divide into « group of three longitudinal quadrupoles in which the dipoles or ‘monopoles are arranged in line and in which both characteristic directions coincide, and a group of six leteal quadrupoles with mutually perpendicular characteristic directions (Fig. 2 +--+ oe ages -+ (a Fic. 2 (a) Longitudinal quacrupove, (b) Lateral quad Now. two equal and opposite adjacent forces con- stitwie a stress; a pressure type of stress in the longitudinal case, 2 shear siress inthe lateral ease. Hence, © quadrupole distribution acting on an acoustic medium is exactly equivalent to the action of an external stress on the medium. This is exactly What we should expect, since Lighthill's formulae tioa involves writing the full momentum equation in the form (3.8) of an acoustic momentum equation driven by an external stress field Thx, 1) Moreover. t's obvious that any wave equation of the kind derived by Lighthill must contain only quadrupoles and higher-order multipoles fo, in the absence of external sources of mass or exiernal body forces, mass and mo served, i.e. the right side of have 4 double integral van regarding @°T, Jaxx, asa mc {otal monopole strength is, [zhe while regarding @°T,/axé. op /os be ot doe [rea-—[o eae wave equation to take the f derivative (and for T, 10 v2 fas x->=) in ofder that me conserved. Next. we consider the ¢ Lighthill's wave equation § senseless 10 write the ecu unless physical conditions : tions, satisfying the homos do emerge sat sufficiently 2 unsteady flow. If we restri region of unsteady flow, th side that region. For then, take place. which implies that viscous forces are nee and that pressure and dens aida. Therefore, Lighthill ordinary acoustic theory a region of unsteady flow “This migh: seem an obvi of any theory, but it is x theories (eg, that due 10 R differen: representations Ribner's theory requires th the wave equation to oceuy than hat in which the unste 2 region with typical dimer the flow source volume is s whenever the s. recsel At any rate, Lighthil's only non-zero within the fi ‘he Reynolds number is Viscous conteibutions to T; are better reasons, given | ous terms, Pressure and dé related by the local sound if temperatures in the 0! from those outside, the dif Will be small. Then, the onl ‘© Lighthill's strese tens Auctuating Reynolds stress umber is small, p may ensity py in the flow with (Mach number). Thus, in