ABER gnabramovich; the theory of. turbulent Jets translation by scripta technica technical editing by leon h. schindel uh THE WLLT. PRESS » MASSACHUSETTS INSTITUTE OF TECHNOLOGY « CAMBRIDGE, MASSACHUSETTSPreface This book is a monograph on turbulent jets and considers the problems of free jets, wakes behind bluff bodies, jets confirmed by solid walls, and other cases, The first monograph written by the author on this subject came out in 1936 and dealt solely with the free submerged jet, Since that time, the theory of the turbulent jet has been further developed in many works published both in the USSR and abroad; ithas been enriched witha large amount of experimental material and has been applied in different branches of engineering. Two summaries of the rapidly accumulating new information on the theory of the free jet and its applications have been published by the author in 1940 and 1948, But it has only been over the last ten years that substantial progress in this fieldhas been made, and that it has been possible to go beyondthefree submerged jet and solve, in particular, the problem of a jetina stream of fluid, to take into ac- count the interaction between the solid walls and the jet, to ascertain the relationship between the contour of the jet and the ratio of its density to the density of the surrounding medium, to establish the characteristic features of a super-sonic jet, and so forth, ‘The present monograph contains the results of further research by the author and his colleagues, some of which have been reported in journals, as well as the results of a critical reappraisal of the more important theoretical and experimental data published by a number of other authors, The first part deals with the theory of a turbulent jet of incom- pressible fluid, It gives a systematic analysis of numerous experi- mental data on velocity profiles, temperature, and constituent concentration, as well as the boundaries of the turbulent mixing zone, A great deal of attention isgivento a comparative analysis of different theories of free turbulence, A study is presented not only of submerged jets but also those propagated through a stream of fluid (in coflow and counterflow), The theory of a coflowing jet de- veloped by the author does not require the addition of any extra experimental constants (compared with the submerged jet) and in addition agrees well with experimental data, The theory takes into account the effect of initial nonuniformity of the velocity, tempera- ture, and admixture concentration fields, and it has been possible in it to do away with some earlier uncertainity in the choice of an empirical constant depending on the structure of the jet. The effect of a longitudinal pressure gradient on the nature of the spread of the jet is considered. The theory of a coflowing stream is extended to , cover the case of a turbulent wake behinda body; as a result, theoy ): retical equations are derived for the thickness of the wake vvi Preface velocity on its axis, applicable both in close proximity to the body as well as at some distance from it. The second part of this book sets forth the theory of turbulent gas jets, including strongly preheated and supersonic jets, The theory of free turbulence ina gas, suitable in principle for any degree of com- pressibility, is revised, and equations are derived for the motion and heat transfer in the boundary layer ofa jet at very high temperature; it is pointed out that in this case the effect of molecular viscosity may not only prove commensurate with the effect of turbulent mix- ing but may even be predominant, In addition, the section describes tested methods for practical calculation of supersonic nonisothermal gas jets, valid both in the isobaric case as well as at unequal pres- sures in the initial section of the jet and in the surrounding gas stream; it has been possible to take into account the direct effect of the compressibility of a gas on the thickness of the jet, This part of the book also studies gas jets during combustion with diffusion and normal flame fronts. The third part of the bookgives the solution of several problems of the spreading of jets in finite and semifinite space: 1. Flow directly behind a bluff body in a channel of constant sec- tion (detailed study is made of the region of back flow of the fluid), 2, A jet entering a channel closed at one end. 3. A protective jet separating a stream of fluid from a solid wall (“homogeneous film’), ‘The fourth part of the book describes various applications of the theory of jets, many of which have heen developed for the first time or have been revised (jets in a lateral flow of fluid, convective jets, two-phase jets of gas mixed with drops of liquid or solid particles, the mixing chamber of injectors, and so on), Several chapters of the book have been written by the author's colleagues and have been edited by the author (Chap, 7 was written by O, V, Yakovlevskiy, Chap. 8 by V.S, Avduyevskiy, Chaps. 10 and 11 by I, P. Smirinova, and Sec, 4 in Chap, 13 by A. Ya, Cherkez), 0, V. Yakovlevskiy rendered great assistance in preparing the material for the press, The author takes the opportunity of thanking all those involved.Contents Part 1 TURBULENT JETS OF INCOMPRESSIBLE FLUID Chapter 1, General Properties of Turbulent Jets. . . 1.1, Fundamental Concepts . . 1.2, Submerged Jet... ‘3, Velocity Profiles in a Submerged Jet). || ‘4. Spread of a Turbulent Submerged Jet... . 5 6. Lines of Constant Velocity in a Submerged Jet . Velocity Variation Along the Axis of a Submerged Jet eee cece cece neces La. Heat Transfer in a Submerged Jet. 1,8, Diffusion of Constituents in a Submerged Set 1,9. Velocity, Temperature, and Concentration Pro files in a Turbulent Jet Spreading into an External Stream of Fluid... ... eee eee eee eee : 1.10. Spread of a Turbulent Jet into a Coflowing 0 or Counterflowing External Stream 1.11, Turbulence Characteristics in a Free Jet. Bee Chapter 2. Theory of Free Turbulence for the Case of a Submerged ete 2.1, Theory of Free Turbulence ... 212, Prandtl’s Old Theory of Free Turbulence. 2.3, Application of Prandti’s Old Theory of Free Tur- bulence to Heat and Diffusion Problems 2.4, Theory of the Boundary Layer of a Two-Dimen- sional Turbulent Jet of Incompressible Fluid . 2,5, Tollmien’s Plane Turbulent Source 2,6, Tollmien’s Axially Symmetric Turbulent Source . 2.7, Distribution of Temperature and Constituent Concentration in the Main Region of a Jet Ac- cording to Prandtl’s Old Theory of Free Turbulence... .......2.05 : 2.8. Taylor's Free’ Turbulence Theory’ and ‘its Application ed 2.9, Prandtl’s New Theory of Free Turbulénee and its Applications. 66.2... ees ee cece eeees vii Page 36 43 50 50 52 56 61 68 76 85 89 103viii 2.10, 211. Chapter 3, Theory of Turbulent Wake Behind a Body . 3.1. 3.2, 3.3. 3.4, 3.5. Contents Reichardt’s Theory of Turbulent Mixing and its Application ec et Determination of the Temperature Profile in a det on the Basis of the New Prandtl- Gértler Theory of Turbulence and Reichardt’s Theory. . Flow Paitern of Fluid in a Wake ‘Two-Dimensional Wake, Schlichting’s Theory, - Investigation of the Thermal Wake Behind a Heated Body on the Basis of Taylor’s Tur~ bulence Hypothesis........ pee Calculation of an Axially Symmetric Wake Ac- cording to Prandtl’s and Taylor’s Theories .. . Calculation of the Wake Behind a Body According to Prandtl’s New Theory of Turbulence and Reichardt’s Theory .......eeeeeeeeee Chapter 4, Turbulent Mixing at the Boundary of Two Fluid Streams Moving in the Same and Opposite Directions. . . poonded Kinematic Diagram of the Boundary Layer. . Geometrical Characteristics of the Boundary Doyen eee eee Mixing Zone of a Two-Dimensional Jet of Finite Thickness in an Infinite Stream Turbulent Mixing Zone in the Case of Variable Dresemre eee ete ee Chapter 5. Jet of an Incompressible Fluid in a Coflowing External Stream. ...0. 00. 00000 . Initial Region of a Two-Dimensional Jet... . Initial Region of an Axially Symmetric Jet . . General Relations Characterizing the Main Region of a Jet with an External Flow..... Main Region of a Two-Dimensional Jet .... Main Region of an Axially Symmetric Jet. . . Transition Region of a Jet. Turbulent Wake Behind a Body The Influence of an Initial Nonuniformity of the Stream on the Initial Region of the Jet ‘The Influence of an Initial Nonuniformity of F onlthe|Main Dart(ofe\set eee eee . Comparison of the Theory of a Jet with Experi- mental Data........... - Approximate ‘Theory of “Set “tnan’ External Stream, 113 120 126 126 128 133 138 143 148 148 154 162 164 172 172 176 180 189 195 201 208 211 219 225, 230Contents Part I TURBULENT GAS JETS Chapter 6, The Theory of Free Turbulence in a Com- pressible Gas : - Momentum and Continuity Equations . . . Energy Equation Transformation of the Basic Boundary-Layer Equations of a Jet ........4+ ono The Temperature and the Velocity Profiles in’ the Two-Dimensional Boundary Layer of a Sub- merged Gas Jet. an oe 6.5. The Boundary Layer of a Sipersonic isobaric Gas Jet 55 6.6. The Redistribution of Stagnation Temperai a Supersonic Jet... .e.sssseeeeeeceee 6.7, The Gas Jet at a Very High Temperature. Chapter 7. Turbulent Gas Streams ......-00e0s00- Tl. General Properties of Turbulent Jets of a Com- pressible GaS..... 0... eect eeeeee eee 7.2, Turbulent Mixing Zone at the Boundary Between Parallel Nonisothermal Streams of Gas...... Zone of Turbulent Mixing on the Boundary Be~ tween Parallel Streams of High-Velocity Gas . . Basic Relations for Computing the Initial Part of a Jet of Gas in an External Stream . . 5 Principal Region of a Jet of Heated Gas. Principal Region of High-Velocity Gas Jet Discharge of a Supersonic Gas Jet froma Nozzle at Off-Design Conditions... ...+.e+eeee — Chapter 8, Turbulent Jets in the Presence of a Flame ee = 8.3, aa 8.5. 8.6. —- 8.8, 8.9, Front (The Plume)...... . Possible Combustion Regimes .......0..0+% Distribution of Total Energy, Concentration, and Temperature in the Boundary Layer of a Jet. .. Physical Pattern of Combustion in a Turbulent Boundary Layer eee Determination of the Com! ion ‘Temperaturé . Calculation of the Two-Dimensional Boundary Layer ......... eee ee Combustion Flame ‘in’ the Main Region of a Tet cere ce rer eeeceves Combustion’ of a Homogeneous Gas in'a Two- Dimensional Boundary Layer.....+ ace ‘Turbulent Jets with Diffusion Flame Fronts . Experimental Studies of Diffusion Flames 237 237 241 24d, 247 254 259 262, 267 267 278 293 302 306 324 335 345 345 347 351 355 358 360 363 373 385Contents Part ill JETS IN FINITE SPACE Chapter 9, Fluid Flow in a Channel Behind a Bluff Body ... 9.1, Flow in the 9.2, First Region of the Two-Dimensional Cireula— lation Zone... .. 9.3. Second Region of the Two-Dimensional Circula— lation Zone. ..... 9.4. General Characteristics ofthe Two-Dimensional Circulation Zone... ... a 9.5. Diffuser Region of a Plane Channel witha Cireu- Jation ZONE... eee eee eee ee eee ete e ees 9.6. Axially Symmetric Fluid’ in’a Channel behind a Bluff Body . . oo 9.7, Diffuser Region an Axi Ly Symmet with a Circulation Zone.......-..05 Chapter 10, Turbulent Jet Discharged into a Dead-End Channel. : 10.1. Formulation oft the Prot lem. 10.2. First Region of the Two-Dimensional Flow Formed by a Turbulent Jet Discharged into a Dead-End Chamnel ......2.4 4 ee Second Region of the Two-Dimensional Flow Formed by a Jet Discharged into a Dead-End Channel oo... see eeeee rn 10.4, General Characteristics of the Two-Dimen- sional Flow Formed by a Jet Discharged into a Dead-End Channel\(e ene Axially Symmetric Fluid Flow Formed bya Jet Discharged into a Dead-End Channel....... 10.6, Influence of the Initial Nonuniformity of the Velocity Ficld oe. ess seeceeececrece 10.7, Comparison of the Results of Calculation and Experiment....... 10. 10. Chapter 11, The Flow Produced by a Semi-Contained Turbulent Jet.......4. 11,1, Formulation of the Problem. First Region of a Semi~Contained Jet Main Region of a Semi-Contained get for Mole ee 11.4, Main Region of a Semi “Contained Jet “for lg eee 391 391 393 402 414 419 428 435 444 445 452 460 461 467 AML 415 45 476 AT8 489Contents Part IV SOME PRACTICAL APPLICATIONS OF THE THEORY Chapter 12, Jets in Unbounded Space. . nae 12,2, 12.3, 12.4, 12.5. 12.6, 12.7, 12,8, cee 12,10, 12,11. OF TURBULENT JETS Calculation of Submerged Jet of an Incompres- eile riod eee Calculation of an Open-Jet Wind Tunnel | The Jet in a Stream of Relatively High ‘Turbulence ...... Concerning the Gontiguration of a Jet ina De- flecting Flow. The Air Curtain. 1111)! Vertical Submerged ‘Turbulent Jet of Heated Gas. Trajectories of Warm and Cold Air Jets, The Two-Phase Jet in Air...... Concerning the Discharge of a into ‘a fimid een ene See Turbulent Heat’ Transfer from ‘Steam toa Water Jet 5 | The Fan Jet Chapter 13, Jets in Finite Space... ...e+eeeseeeee 13.1, 13.2, 13,3. 13.4, 13.5. References.......+ Resistance of a Labyrinth Seal of an Air Blower . External Drag of an Use of the Theory of a Free Jetfor Calculation of the Air Resistance of a Train.......... Calculation of the Mixing Chamber of an Blend Concerning the Decay of Velocity and Tempera- ture Profiles in Turbulent Streams ...... xi 505 505 524 536 B41 556 562 580 586 600 611 621 625 625 628 632 633 654 663Part | TURBULENT JETS OF INCOMPRESSIBLE FLUIDChapter | General Properties of Turbulent Jets 1.1, Fundamental Concepts In many cases of motion of a liquid or gas, so-called tangential separation surfaces arise; the flow of fluid on either side of this surface is termed a jet. The jets may be moving in the same direction or in opposite directions. Tangential separation is ex- perienced by such parameters as flow velocity, temperature, and specie concentration, though the distribution of static pressure proves continuous. As is well known [4], the instability of the tangential separation surface causes eddies on it, which move in disorderly fashion both along and across the stream; this brings about an exchange of matter between neighboring jets; i.e., there is a transverse transfer of momentum, heat, and constituents, As a result, a region of finite thickness with a continuous distribution of velocity, teniperature, and specie concentration is formed on the boundary between the two jets; this region is termed the turbulent jet boundary layer, At very low Reynolds numbers, this boundary layer may be laminar, but we shall not deal here with these comparatively rare cases. The simplest case of a jet boundary layer is found during the discharge of fluid with a uniform initial velocity field (u= const) into a medium moving at constant velocity (u,=const), since here the thickness of the boundary layer in the initial section of the jet is equal to zero, The thickening of the jet boundary layer, which consists of particles of the surrounding medium carried along with it and particles of the jet itself that have been slowed down, leads, on the one hand, to an increase in the cross section of the jet and, on the other, to a gradual ‘eating up’’ of its nonviscous core — the region between the inside boundaries of the boundary layers, Figure 1.1 shows a simplified diagram of the jet, The part of the jet in which there is a core of potential flow is termed the initial region (see Fig. 1.1). . As shown by numerous experiments, one of the fundamental * properties of a jet of this kind is the fact that the static pressure” is constant throughout the flow,* as a result of which the velocity *In certain cases (when the jet interacts with an obstacle), the pressure may not remain constant, but we shall deal with these particular cases separately.4 Turbulent Jets of Incompressible Fluid in the potential core of the jet remains constant. The “eating up” of the jet beyond the initial area shows up both in its widening as well as in variation of the velocity along its axis. At a certain distance from the end of the initial area, the jet becomes similar in appearance to a flow of fluid from a source of infinitely small thickness (in an axially symmetric case the source is a point, and in a plane-parallel case it is a straight line per- pendicular to the plane of flow of the jet). ‘Transitional Initial region iS ly Ih t Texoma nT am t | Fig, 1.1. Diagram of jet. Use is often made of a simplified diagram of the jet in which it is assumed that the length of the transitional region is equal to zero; in this case the cross section in which the main and initial areas come together is termed the transitional cross section of the jet. If the transitional area is taken into account in the calculations, then the transitional cross section is considered to coincide with the beginning of the main area, 1.2. Submerged Jet The type of turbulent jet most studied is one spreading through a medium at rest; a jet of this kind is said to be submerged. If . the yelocity field at the initial cross section of the submerged jet zis uniform, the boundaries of the mixing layers form diverging surfaces, which intersect at the edge of the nozzle (in Fig. 1.2 at the initial cross section of the jet). On the outside, the boundary layer comes into contact with the stationary liquid, and the outside boundary is taken to mean that surface on all points of which the velocity component with respect to the x-axis is equal to zero (2=0). On the inside, the boundary layer changes to a constantGeneral Properties of Turbulent Jets 5 velocity core; hence, on the inner boundary of the boundary layer, the flow velocity is equal to the discharge velocity (11=u0). Fig. 1.2. Diagram of submerged jet. In this description of the jet, it is assumed that the boundary layer is of finite thickness; in certain theories of the submerged jet, it is assumed that the boundary layer is of infinite thickness and has asymptotic profiles of velocity, temperature, and other quantities, Both these interpretations of the boundary layer can be reconciled in practice, since the asymptotic layer can be ap- proximately replaced by a layer of finite thickness*. 1.3. Velocity Profiles in a Submerged Jet A characteristic feature of a turbulent jet, as shown by theory and also by numerous experiments, is the smaliness of the trans— verse velocity components in any section of the jet compared with the longitudinal velocity. Hence, if the x-axis is aligned with the axis of symmetry of the stream, the velocity components along the y-axis prove so small that they canbe disregarded in engineer- ing problems involving jet theory. Figure 1.3 shows curves for the velocity distribution (or to be more exact, the velocity com- ponents along the x-axis) at different cross sections of the princi- pal area of a round air jet discharged into stationary air (Triipel’s experiments) (5). The initial velocity of the jet u=87 m/sec. The radius of the initial section is r=0.45 m. ‘The velocity profiles were determined at the following distances from the nozzle: *=0.6, 0.8, 1.0, 1.2, and 1,6 m, As in studies made by other investigators, Triipel’s experi- ments show a continuous broadening of the velocity profile of the *In this case the “boundaries” of the asymptotic layer are the surfaces on which the velocities (or, for example, the tempera- tures) differ from their limiting values by a small prescribed amount, for example 1%.6 Turbulent Jets of Incompressible Fluid 2, 2/secy 7| 20) a WS Wid is tae 0a ae Fig. 1.2. Velocity profiles in different sections of an axially symmetric submerged jetfrom Triipel’s experimental data [5). jet. The velocity profile becomes “lower”? and “wider” with in- creasing distance from the beginning of the jet. This conclusion is reached when velocity profiles are plotted in physical coordinates (u, 9). More interesting results are obtained when the same pro- files are plotted in dimensionless form. For example, instead of the absolute velocity, let us try to plot the ratio of the local velocity to the velocity on the jet axisuj/u,, and instead of the distance from the jet axis, let us plot the ratio of this distance to the distance between the axis and the point at which the velocity is equal to half the axial velocity u/y.. The dia~ gram obtained (Fig. 1.4) shows the similarity of the velocity pro~ files in all sections of the principal area of a round jet. Thus, at corresponding points on any two cross sections of the main part of the jet, the nondimensional velocities are the same. Obviously, we need not take y, as the characteristic line, but could substitute, for example, the half-thickness of the jet 6. ‘The equality of the dimensionless velocities for corresponding points in the jet ( 9:/6;=¥elb, ) is then expressed in the following way where u; and up are the velocities at corresponding points of two cross sections of the jet; u,, and u,,, are the appropriate velocities at the centers of these cross sections.General Properties of Turbulent Jets 7 = ra aus 250} 4254 a aii 4s0 asta e550 SRI Fig. 1.4, Dimensionless velocity profile in axially symmetric submerged jet from Triipel’s experi- mental data [5]. The conclusions are valid for other types of jet besides round ones. They apply equally to plane-parallel jets discharged from a long slot. In order to see that this is so, we need only take a look at Férthmann’s results [6]. He has studied the velocity profiles of an air jet emerging from a rectangular nozzle 0.03 m high and 0.65 m wide. The jet velocity at the outlet from the slot was 35 m/sec. Figure 1.5 plots the velocity profiles obtained for jet sections at the following distances from the nozzle: x=0, 0.2, 0.35, 0.5, 0.625, and 0.75 m, When they are replotted in dimensionless coordinates (the same ones as for Triipel’s experiments), the velocity profiles of the principal area of the plane jet, just as in the case of a round jet, are similar (Fig. 1.6). = rede oredzu oe reQisu axredsle ornQbisu azediy ZINES ae a7 “DOF v a ge Fig. 1.5, Velocity profiles at different sections of a plane jet from Forthmann’s experimental data {6].Turbulent Jets of Incompressible Fluid <7 da Pal 3 st G75) ‘* Ae % a 450) a & * a B a2 “he orf 8 oS SO Fig. 1.6. Dimensionless velocity profile in plane jet according to Férthmann’s experimental data [6]. i * alee Cane 475|2 28 s*—a50} lo Rt aes, ai 7 w w w yaw Fig, 1.7. Velocity profiles in different sections of boundary layer of axially symmetric jetfrom Abramo- vich’s experimental data [3]. In 1938 the author made an experimental study of the initial region of an axially symmetric air jet discharged from a nozzle 100 mm in diameter at velocities up to 40 m/sec [8]. The velocity - profiles obtained for the boundary layer of the initial area at distances from the nozzle x=0.1 and 0.25 m are plotted in Figure 1.7, The profiles shown in Figure 1.7 are plotted in Figure 1.8 in the dimensionless form o_o be we): in which uo is the velocity in the core of the unperturbed stream (discharge velocity); 4y,=y—y, is the distance between the point ofGeneral Properties of Turbulent Jets . Z is as 0.8 - G @ seas one’ © x=tt0m | q 3 a2 le es a Fe 2 A eC %, Fig. 1.8. Dimensionless velocity profile in the boundary layer of initial region of submerged axially symmetric jet (=50mm) according to Abramovich’s experimental data §B). ae % 10 exoye a ent” 0 aay” Oren" accel" “0 05 at 02 Oy, a Fig. 1.9. Dimensionless velocity profile in boundary layer of initial region of plane jet (t,<12.7mm),from the experimental data of Albertson et al. [7]. measurement and the point at which the velocity is half what it is in the core (1,-0.5); AY, = (Yo.g —Uo,) 18 the distance between the points at which the velocity is, respectively, 0.9 and 0.1 of the unperturbed stream velocity (u/to=0.9 andu/uo=0.1); we shall con- sider this dimension to be the scale describing the thickness of the jet boundary layer; the exact determination of it on the basis of experiments is extremely difficult, The results shown in Figure 1.8 indicate that in the initial area at small distances from the nozzle (4< 3) the boundary layer10 Turbulent Jets of Incompressible Fluid velocity profiles at different cross sections ofan axially symmetric jet are similar. Figure 1.9 shows dimensionless velocity profiles obtained by Albertson, Dai, Jensen, and Rouse [7] in the boundary layer of the initial area of a plane-parallel air jet issuing into stationary air from a slot-shaped nozzle 1 inch high (2b:= 25.4 mm), at different distances from the nozzle exit (G2, 4, 6, and 8). As we can see, here, too, the different velocity profiles lie along the same universal curve. 1.4. Spread of a Turbulent Submerged Jet In a turbulent jet the components of velocity at any point can be decomposed into a time-averaged value plus a randomly varying perturbation wat ye, When averaged over some finite time interval, the fluctuation or pulsation components are equal to zero: If the mean free path of a fluid particle (mixing length) in a tur bulent stream is equal to /, when moved in a transverse direction, the particle reaches a layer inwhich the mean velocity differs from what it was in the layer from which the particle separated by the following value: ou ay The loss in individuality of the fluid particle — its merging with the mass of the new layer — should be accompanied by a dis- continuous variation (perturbation) in velocity of uw abu, In other words, perturbations of the streamwise velocity component are of the order + wnt, ar It is usually assumed that the transverse perturbations in velocity v’ are proportional to the streamwise perturbations w’, but have the opposite sign —v' nw’, Le, —General Properties of Turbulent Jets cre The absence in the free stream of solidboundaries that might damp the oscillations of the particles led Prandtl [8] to assume that in this case the mixing length at any cross section of the stream was constant: 1(y) =const. ‘The variation in the mixing length along the x-axis /=!(x) may be established by means of available experimental data, A sufficient basis for Prandtl’s assumption is provided by the similarity of boundary layers in different cross sections of a free stream. As was mentioned earlier, this similarity has been established by a large number of experiments in which the velocity profiles prove universal when plotted in dimensionless coordinates = i(2), on ile,, they coincide for different sections of the jet, The similarity of boundary layers in the sections of the given free stream implies, among other things, the similarity of geometric dimensions. In other words, equality may be expected between the dimensionless mixing lengths for different sections of the stream 7 ves = const. (1.2) be Thus, it is sufficient to establish a law for the growth of the jet as a function of distance along the x-axis in order to define the way in which the mixing length increases. Prandtl (9Jassumes that the growth of the jet (i.e. the rate at which the thickness of the jet boundary layer increases) is controlled by the transverse perturbation velocity (1.3) Because of the similarity of velocity profiles in different cross sections of the jet, it is possible to write ot fn a a4) and, hence, according to Eqs. 1.2, 1.3, and 1.4, at a 73 (1.5) On the other, hand, rate of growth of the jet (1.6)12 Turbulent Jets of Incompressible Fluid A comparison of Eqs. 1.5 and 1.6 provides a solution of the problem of establishing the law governing the increase in thickness of the submerged jet and the mixing length in the direction of flow a =const, 6=x-const, ¢=cx. @.7) ‘The derived linear law for the increase in the jet thickness and mixing length along the stream holds for jets of different shape: the boundary layer of an infinite plane stream, the plane-parallel jet, and the axially symmetric jet. Inall three cases the law applies when the velocity profiles in the submerged jet are universal. 1.5. Lines of Constant Velocity in a Submerged Jet Let a uniform, semi-infinite, plane-parallel jet issue from the . wall AO at a velocity up, and starting from point O merge with the surrounding motionless fluid.* The previously established linear increase in thickness of the boundary layer in conjunction with the universal or self-similar nature of the velocity profile means that along any ray Oz drawn from the origin of the coordinates (which coincides with point O at which the thickness of the boundary layer is zero), the velocity remains constant. Fig. 1.10. Diagram of boundary layer in submerged jet. Indeed, the similarity of the velocity profiles implies the equality of velocities at corresponding points in the stream; i.e., at ls f += const *If the plane Ox in Fig. 1.10 is horizontal, it is not necessary to consider gravity, which has an effect, for example, on a non- isothermal gas jet.General Properties of Turbulent Jets 13 there results = read i const. But according to Eq. (1.7) b= x-const; from which we find that on the ray a const (1.8) the following condition is satisfied const. (1.9) Thus, in a turbulent boundary layer of a plane-parallel sub- merged stream, the rays converging at the point where the thick ness of the layer is zero constitute equal-velocity lines. This result refers both to plane-parallel flow and the boundary layer in the initial area of a round jet, since experiments show that in this case, too, the velocity fields are universal. The origin of the coordinates for the equal-velocity lines in the boundary layer in the initial area of the jet is the outlet edge of the nozzle (when the velocity field in the initial section of the jet is uniform), : It should be pointed out that in the initial area, equal-velocity lines plotted for physical (u) and dimensionless (u/ue) velocities coincide, since the velocity up along the length does not vary in the core of the flow. ‘The shape of these lines in the main sector of a submerged jet depends on the method used to determine the dimensionless velocity, If the dimensionless velocity is obtained by dividing the local velocity by the discharge velocity from the nozzle (u/us), just as for the physical velocity, the lines of the principal area form a. fan, shown in Figure 1.11. For the dimensionless velocity, calculated by dividing the local velocity by the axial velocity in the corre- sponding cross section (u/i,), the lines are straight and converge at the pole of the jet (Fig. 1.12). ‘This result stems from the fact that the dimensionless velocity is a function of only the relative Position of the point in the cross section of the jet os s()- (1.10)4 Turbulent Jets of Incompressible Fluid Fig. 1.11. Equal velocity lines in submerged jet. a | Se +$———_ {--——__ Fig. 1.12. Equal velocity lines (#/u,—const) in main region of submerged jet. On account of the linearity of the thickening of the jet (Eq. 1.1), Eq. 1.10 may be brought to the form: which indeed proves that the equal velocity lines for the dimension- Jess velocity uju,, are rays which intersect at the pole of the jet. The rectilinearity of these lines for a dimensionless velocity ( t/#m) is found both in axially symmetric and plane-parallel cases. The experimental data given below are clear proof of this. Figure 1.13 plots five lines on each of which the dimensionless velocity is constant: u/u,= 0.9, 0.7, 0.5, 0.3, and 0.1, ‘The experi- mental points are taken from the Férthmann experiments mentioned earlier for a plane-parallel submerged jet, discharged from a slot of height2b)=30 mm and width 650 mm, and from Triipel’s experi- ments for a submerged round jet with an initial diameter 2-90 mm.General Properties of Turbulent Jets 15 TT TT Tad Lo Jo Trupet’s experiment (5) ris fe Forthumana's experiment [6] =i tt Sey See 2 ese : 2B ea Pais ba eg 2g & Fig. 1,13, Equal dimensionless velocity lines in sub- merged jet. As can be seen, Férthmann’s experimental points (crosses) and Triipel’s points (rings) lie close along the same ray for each dimensionless velocity (u/«,,— const), and the rays for the five velocities converge at the same point. The angular coefficient of the ray u, =0.5u,, which is roughly the same for plane-parallel and axially symmetric submerged jets, is approximately equal to == 0.097. (1.11) 1.6. Velocity Variation Along the Axis of a Submerged Jet As shown by experiment, the pressure in the jet is virtually invariable and equal to the pressure in the surrounding space. On account of this the total momentum of air mass per second should be the same in all the cross sections of the jet: . e fwdm =f putdF = const, (1.12) Q ‘ in which dm is the mass flowing through an element of the cross section of the jet per unit time; ¢ is the air density; dF is the area of the element. : For a round jet the condition for constant total momentum can be written in the following way: 38 | (y22 = const, | (1.13) a in which u,, is the velocity in the center of the given sections, x is the distance between the given section and the jet pole, y is the16 Turbulent Jets of Incompressible Fluid radius, and r is the radius of the outer boundary of the jet section in question. ‘On account of the universal nature of the velocity profiles, the dimensionless velocity(u/u,,Jat the selected point is a function of only the dimensionless coordinate (y/t) of a ray drawn from the pole through this point, Hence, fle}. We find as a result of the Eq. 1.13 that the velocity in the center of the section of an axially symmetric submerged jet is inversely proportional to the distance from the pole ' __ const fy = (114) For a plane-parallel submerged jet, the fact thatthe momentum is constant gives us the relationship wx 2 oan (zy* = const, (1.15) in which 6 is the semithickness of the section. On account of the universal nature of the velocity profiles, {ef -com. Hence the drop in velocity along the axis of a plane-parallel jet takes the following form: (1.16) The proportionality constants in Eqs, 1.14 and 1.16 are deter- mined by the integrals of Eqs. 1.13 and 1.15, and in order to calculate the latter, we must know the velocity distribution in theeet General Properties of Turbulent Jets Ww cross sections of the jet. Because of the universal nature of the velocity profiles, all we need do is determine the velocity distri- bution experimentally in only one section in the main region of the jet. The theory described below enables us to solve this problem theoretically. A shortcoming of Eqs. 1.14 and 1.16 isthe fact that the distance x is reckoned from the pole of the jet and not from its initial section, In Section 9 we derive more suitable relationships, in which the distances are taken from the beginning of the jet. Equa- tions 1.14 and 1.16, as will be shown later, agree well with experi- mental data, Heat Transfer in a Submerged Jet In engineering we often encounter a submerged jet in which the temperature differs from that of the surrounding medium. ‘The solution of the problem of heat transfer from air at rest to a jet (and back again) is possible only if the variations in tem- perature along the jet and in its cross sections are known. Let us introduce consideration of the temperature changes: a, the difference between the temperature at a given point in the jet and in the surrounding medium (room) AT=T—Ty; b, the difference between the temperature on the jetaxis and in the surrounding space AT mn = Tm —T yi c. the difference between the temperature in the initial cross section of the jet (at the mouth of the nozzle) and the surrounding space - AT, = o— Tae The nature of the distribution of the magnitudes of the tempera~ ture differences in a submerged jet, as shown by experiment, is Similar to that of the velocity distribution, In the constant velocity core of the initial region the temperature is constant and equal to the temperature of the fluid in the initial cross section of the jet. In the main region the temperature boundaries of the jet expand with distance from the nozzle since the temperature difference on the flow axis decreases. Figure 1.14 shows the dimensionless temperature field found by the author [3] for a boundary layer in the initial region of an air jet of round section, 100 mm in diameter, at a distance of 250 mm from the nozzle (white dots); the abscissa isthe same as in Figures 1.8 and 1.9. For the purpose of comparison, the velocity field obtained for the same jet section (black dots) has been transferred18 Turbulent Jets of Incompressible Fluid u te BPot oo peo (© Temperature profile #* Velocity profile 3 aro e Loe “GPUS “0 0 & os Fig. 1.14. Dimensionless temperature difference pro- file inboundary layer of axially symmetric jet according to Abramovich’s data [3]. Te «Temperature profile | Velocity Ite, profile ~ 225} —| IES SE Zz & Fig. 1.15. Dimensionless temperature and velocity profiles in main region of plane jet according to Reichardt’s experimental data [10]. to Figure 1.14 from i igure 1.8, In this experiment the discharge velocity of the jet from the nozzle amounted tou,—25 m/sec, and the temperature difference at the nozzle outlet section was AT 9=35°C, The graph in Figure 1.15 plots the dimensionless temperatures in the main region of a plane jet as 12) Tm in which y is the variable distance between the point and the axis (ordinate); y, is the ordinate of the point (at the same section) at which the velocity is half what it is on the jet axis. The experimental data have been taken from Reichardt [10], who investigated a submerged air jet discharged from a long,General Properties of Turbulent Jets 19 rectilinear slot with sharp edges and 7 x150 mm in size; the discharge velocity was about 50 m/sec and the initial temperature difference 10 to 20°C. For purposes of comparison, Figure 1.15 also shows the dimensionless velocity field for the same jet. The velocities and temperatures were measured by Reichardt at 400 mm from the slot; the temperature fields were found to be con- siderably fuller than the velocity fields. Figure 1.16 plots the dimensionless temperature difference AT/A Ta =[(ylydi the data were obtained by Stark [11] for different cross sections of the main region of an axially symmetric air jet discharged into stationary air. In his experiments, Stark used a nozzle with diameter 2r) 41.5 mm, discharge velocity #)=20m/sec and initial temperature difference AT)~15°C. All the points lie roughly along a universal curve, although they relate to different cross sections at distances from the nozzle ranging from 10 to 70 nozzle radii. For purposes of comparison, Figure 1.15 shows the velocity fields for the same sections. Similar results were obtained by Ruden [12] in his experiments. Figures 1.14 to 1.16 show that the experimental curves for the dimensionless velocity and dimension- less temperature obtained for the same cross section of a sub- merged jet do not coincide. The theoretical laws governing the distribution of temperature in the cross sections of the jet are given in Chapter 2. —+—Temperature profile ———Velocity profile asl a5} L o as Ww 1s 20 = Fig. 1.16. Dimensionless temperature differences and velocity profiles in main region of axially symmetric jet according to Stark’s data [11].20 Turbulent Jets of Incompressible Fluid ‘The distribution of temperature along the axis of the main region of the jet can be established in the same way as for the velocities, except that instead of the constant value of the mo- mentum, the constant heat content of the jet must be used. Indeed, in the determination of the heat content of the jet from the tem- perature difference, the fluid sucked up by the jet from the sur- rounding space does not contribute to the heat transfer, since the temperature difference is equal to zero (\7, =0). In other words, the surplus heat content of the entire mass of fluid passing through a jet section chosen at random is equal to the surplus heat content of the initial mass emerging over the same interval of time from the nozzle, ‘The diffusion of heat through a jet shows up in its transfer across the boundary between the initial mass and the incorporated mass, on account of which, as pointed out earlier, the temperature on the jet axis drops, while the temperature field in the cross sections is gradually “smoothed” out. The constancy of the heat content of the free jet, calculated in terms of the temperature difference, is expressed by the following relationship (see Eq. 1.12), ” F {etan= fo u ATAF = const; (1.17) for a jet of circular section, (aT ey dy Gre ag ee Teams (1.18) il AT é Because of the similarity of the profiles in different cross sections of the jet, the dimensionless isotherms, just as the equal- velocity lines, are rectilinesr rays intersecting at the pole of the jet ah =0 (4) . (1.19) in which z is the distance between the pole and the section under discussion; y is the distance between the selected point and the center of the section where the temperature rise is47,,. Equation 1.19 implies that the integral in Eq. 1.18 is a constant. Hence, consideration of Eq. 1.14, gives the law of the drop in the temperature along the axis of a circular jet: AT = S20, (1.20) Further on, it will be shown that Eq. 1.20 agrees with experi- mental data.General Properties of Turbulent Jets 21 Jet theory enables the calculation of the constant in Eq. 1.20, and its value is not equal to the constant in Eq. 1.14, This is be- cause of the difference in the universal profiles of temperature and velocity in the submerged jet. The problem has been elaborated in greater detail in a special section on equations for temperature distribution along the jet axis and in its cross sections, In a plane-parallel jet the conservation of the surplus heat con- tent can be expressed in the following way: am a4 = const. (1.21) 87m tm BT gy thy % Sane Since the integral in Eq. 1.21 is constant, by taking Eq. 1.16 into account, we can find the drop in temperature along the axis of a plane-parallel jet: ge oe (1.22) Ve To round off the preliminary analysis of the transfer of heat in a submerged jet, let us prove the theorem of the relationship be- tween the mean temperatures and mean velocities, The fact that momentum in the jet is constant implies that the product of the mass per second flowing through a random section of the jet and the mean velocity is a constant: Iles Mall, Milley = Moo » (1.23) ie., Furthermore, the fact that the heat content is constant means that the product of the mass flow and the mean temperature dif- ference does not change, either, from section to section: mATes 87 MAT = mM 4T,, i.e., (1.24) By comparing Eqs. 1.23 and 1,24 we find that the drop in the mean temperature along the free jet conforms to the same law as the drop in mean velocity: Tey tes Me ae (1.28) It is interesting to observe that Eq. 1.25 describing the relation~ ship between the mean values still holds, despite the fact that the temperature and velocity profiles in the submerged jet are dissimilar. To avoid errors in the use of Eq. 1.25, we will give the exact definition of the concepts ‘mean velocity” and “mean temperature,”22 Turbulent Jets of Incompressible Fluid Generally speaking, the mean velocity can be calculated in different ways; here the results obtained are not the same, but depend on the physical meaning attributed to the word ‘mean.”? Most frequently the mean velocity is taken tobe the mean arith- metic velocity, which is represented by the ratio of the flow of fluid per second to the cross-sectional area £ fra fe F (1.26) Meg =. ‘The mean arithmetic velocity is obtained by averaging with re- spect to area. The mean velocity can also be obtained as a ratio of the momentum per second to the mass flow per second: F : fudm [wear z (1.27) S foce Feuer The value found from Eq, 1.27 is termed the mean-square velocity. It is found by averaging with respect to the mass rate flow of fluid. Naturally, the values ., and u,. need not be the same, and the difference between them increases as the nonuniformity of the velocity field increases. In the case of a submerged jet, a8 will be shown later on, the mean arithmetic and mean-square velocities differ very considerably. The field of application of each of these concepts will be shown separately. The mean tem- perature also depends on the averaging method. In practical calculations the average temperature according to mass flow of the fluid is almost always applied; hence, from now on we shall take the mean temperature to be the ratio of the heat content of the mass of fluid per second to the mass per second: = . faran — Sarpuar 2 =-t__. (1.28) STa= F S war 6 Returning to Eq. 1.25, we should note that the equality of the dimensionless mean temperature difference and mean velocity is valid only for the mean-square velocity, i.e., for cases in which the latter is derived by averaging with respect to mass flow.General Properties of Turbulent Jets 23 1.8. Diffusion of Constituents in a Submerged Jet In appropriate sections of this book it {s shown that the dif- fusion of any kind of substance which is sometimes found in a jet in the suspended state (for example, gas mixtures, drops of fuel or dust) has very much in common with the propagation of heat. EPS, a NN. ° F0 ar,| ~ + #-20\ Concentration # ~ 4 £-30| profile a a E40 Velocity profile . ‘Temperature profile \, \ as] i \ \, as | NS o ee ce aa L Fig. 1.17. Dimensionless profiles for temperature difference, concentration of admixtire and velocity in main region of plane jet from data given by Abramovich and Borodachev (+=+/bo). Figure 1.17 shows profiles for the dimensionless concentration by weight of carbon dioxide, obtained by the author in collaboration with V, Ya. Boradachev in an experimental investigation of the main region of a plane-parallel jet of carbon dioxide emerging into pure, stationary air: =f(2- = AGE): in which x= Go /G,o. 18 the concentration by weight, i.e., the ratio Of the content of carbon dioxide by weight to the content of air by weight per unit volume at an arbitrary point in the cross section of the jet; x, 1s the concentration by weight of CO; on the jet axis, y is the distance between the axis and the point of measurement; ye i8 the distance between the axis and the point at which the velocity difference is half what it is on the axis. The nozzle from which the jet of gas was discharged had a rectangular outlet24 Turbulent Jets of Incompressible Fluid section with sides 26-3 mm and 2h,=30 mm, i.e., with a side ratio 1:10. The concentration ffelds were measured in the plane of symmetry parallel to the smaller side of the rectangle at the following distances from the nozzle exit: x/bo=10, 20,30, and 40 for a discharge velocity uj=55.9 m/sec. The concentration of carbon was determined in two ways: by chemical analysis of a gas sample taken from the jet and by processing photographs taken with an interferometer. Both methods produced virtually the same results (the discrepancy was the same as the spread of points in the chemical method, which did not exceed 3%of the local concentration on the jet axis). As can be seen from Figure. 1.17, the dimension- less concentration fields of CO. obtained at various cross sections of the jet lie along a universal curve that does not coincide with the curve for the dimensionless velocity (broken line) obtained for the same cross sections; at the same time, the dimensionless tempera- ture curve (dots and dashes), which we have transferred from Figure 1.16, passes very close to the concentration field profile. In cases in which there is an admixture in the surrounding medium as well, it is advisable to introduce the concept of the “concentration difference” (4x is the difference between the local concentration of the admixture in the jet and the concentration of the same admixture beyond the jet). Obviously, the surplus con- centration of admixture is similar to the surplus heat content and is the same for different sections of the jet: 5 F Jsxdm= | pduudF = const. é The constancy of the surplus concentration of admixture as well as the similarity of the concentration fields in the cross section enables us to find the law governing the variation in con- centration along the jet axis. In the main region of an axially symmetric jet, under the condition of preservation of surplus constancy of concentration, (1.29) Here 4z,, is the surplus concentration of admixture on the jet axis. In view of the universality of the velocity and concentration distributions in cross sections of the jet, the integral on the left- hand side of Eq. 1.29 has a constant value; hence ff Eq, 1.16 is taken into account, Ax, = St (2.30) that is, the surplus concentration of admixture along the axis of the main region of a submerged round jet is inversely propor- tional to the distance from the pole of the jet.General Properties of Turbulent Jets 25 In the case of a plane jet, the constancy of the surplus of ad- mixture concentration is expressed by the equality ® rl e & = const. (1.31) een The integral contained on the left-hand side of Eq. 1.31, given universal distribution of velocity and surplus concentration, is a constant, hence the difference in concentration of the admixture along the axis of the main sector of a plane-parallel submerged jet varies in inverse proportion to the square root of the distance from the jet pole: ty, = const nO Ve On the basis of the same condition of preservation of the constancy of the surplus of admixture, we find that the dimension- less concentration, averaged with respect to mass flow, is equal to the dimensionless velocity, averaged with respect to mass flow which, as was shown in the preceding paragraphs, is in turn equal to the dimensionless mean temperature difference: (1.32) bre Tea te “he 7 8Ty ee in which Ax, is the difference between the concentration averaged over: mass flow (%,) in a given cross section of the jet and the concentration of the same specie in the surrounding medium (;,). Further on it will be shown that the fields of the dimensionless surplus admixture and temperature difference coincide; this is because the mechanism of the transfer of heat and of admixtures in a turbulent stream is the same, 1.9. Velocity, Temperature, and Concentration Profiles in a Turbulent Jet Spreading into an External Stream of Fluid Numerous experiments show that the profiles of velocity, temperature, and admixture concentration in a turbulent jet spreading through an external stream of fluid flowing in the same eee as the jet (“coflowing”) are the same as in a submerged jet. Figure 1,18 shows a universal velocity profile obtained by Forstall and Shapiro [14] in their experiments in theprincipal region of an axially symmetric jet of air discharged into an air26 Turbulent Jets of Incompressible Fluid stream with the same direction and temperature; the dimension- less velocity differences are plotted against the dimensionless ordinates (1.34) in whieh a, is the velocity of the external stream; y, is the distance between the axis of the stream and the point at which the velocity difference is half its maximum: u.—u, =0.5 (u,, —u,). The experi- ments were carried out at discharge velocities up to 70 m/sec and 2 coflovwing stream velocity of up to 28 m/sec at the following velocity ratios: m=u./u,=0.2, 0.25, end 0.46. The diameter of the nozzle discharging the active jet was 6.4 mm in one series of | | is According tor ary i ' experiment {14} experiment [5] Fig. 1.18. Dimensionless velocity difference profile in main region of axially symmetric air jet spreading through coflowing stream of air from experimental data obtained by Forstall and Shapiro [i4]. experiments and 25.4 mm in the second series; the diameter of the outer stream nozzle was 102 mm. The measurements were taken between the beginning of the principal region and the section 136 nozzle diameters from its exit, For purposes of comparison, Figure 1,18 shows the velocity profile of a submerged jet (broker line) taken from Triipel’s experiments (see Fig, 1.4); the universal velocity profiles both with and without a coflowing stream proved virtually identical.General Properties of Turbulent Jets 27 #4 val According tot Me ° 2, mes a» 4 . 2-030 \Experiment| 5 » . ms (15) oF % a moos?) % == Experiment [6] \ a5 ee oS, N. ns x] é © JOS Oa. 4 240 2 we a i ‘a Fig, 1.19. Dimensionless surplus velocity profile in main sector of plane air jet spreading through co- flowing stream of air according to Weinstein’s data [15] Haxin). Figure 1.19 shows, in the same coordinates as in Figure 1.18, the dimensionless surplus velocities (velocity differences) obtained by Weinstein, Osterle, and Forstall [15] with an experimental set-up which produced a plane air jet of thickness 2b-12.7 mm into a plane air stream in the same direction of thickness 2b;=278 mm; the width of the initial cross sections of the jet and external stream were the same and amounted to 302 mm. The planes of symmetry of both streams coincided, The velocity of the jet ranged from 30 to 44 m/sec, and the velocity of the external stream from 15 to 22.m/seo; the experiments produced the following velocity ratios for the two streams: m=u,/uo~ 0.33, 0.5, and 0.67. ‘The points shown in Figure 1.19 are for cross sections at distances of x/69~30 and 120 from the initial section, For purpose of comparison, Figure 1.19 shows the velocity profile obtained by Férthmann (broken line) for a plane submerged jet; transferred from Figure 1.6. As we can see, the distribution of surplus velocity in a plane jet with a coflowing stream can be expressed by the same universal relationship as in a plane submerged jet.28 Turbulent Jets of Incompressible Fluid Zhestkov, Glazkov, and Gusev investigated the velocity and temperature fields occurring in the mixing zone of two plane- parallel turbulent jets moving in the same direction at different velocity and temperature ratios. The initial cross sections of each of the adjoining jets were rectangular in shape with sides 40 X125 mm, the long side being common to both. The partition dividing the jets up to the point of contact was 2 mm thick. Figure 1.20 shows the velocity fields obtained during this investigation in the mixing zone at a distance x-100 mm from the beginning of mixing, at the following discharge velocity ratios: m=u,/uo 0, 0.23, 0.43, and 0,64, and almost identical initial temperatures (the tem— perature of the jet at the higher velocity varied betweens,~50 and 53°C; the temperature of the second jet was 1,=27 to 44°C), ue 55 Symbol ToTxlelo F Kelvin | 7261925) 723] 323| £0) aa. sec | 108 105 05s} 07| Kelvin) 300) 306| 37 \ar4 Gp asec | 0 \24.qe51585| LA | a \azqaaslase 4 i Le 1 a a0" ale ee ee & Fig. 1.20. Velocity profiles in boundary layer of two coflowing pie jets (initial region) according to Zhestkov et al, The ratio of initful velocities was varied by altering the velocity of the second (slower) jet, while the velocity of the active jet hardly varied and was u=105 to 107 m/sec. Figure 1.21 shows the same velocity fields in the following dimensionless coordinates: 4 Fs). in which Su=u—u, is the au,” Nye surplus velocity in the jet; Au=wo—u, is the initial difference between the velocities inthe jet; dy, -y—y, is the cross distance between the point of measurement and the point at which the velocity Au, =0.5 Atm; Ay, = (Yoo—Yo1) 18 the distance between the points at which the surplus velocities are, respectively, Su:=0.93u»General Properties of Turbulent Jets 29 and 4 w= 0.1419. This graph also plots the velocity curve taken from Figure 1.9 (broken line) obtained from the experiments of Albertson et al. It {s not difficult to see that the dimensionless pro- file of the surplus velocity in the mixing zone of two plane-parallel jets having the same direction is universal and coincides with the profile in the boundary layer of a plane-parallel submerged jet. | [# al | au, a6; | A me mr023 Experiment bj m=043 |Zhestkov ct al m=04) |-—— Albertson et al. (7) ‘According toy peeo ala Lo O05 05 -aF 2B Oa % Fig. 1.21. Dimensionless velocity profiles in boundary layer of two plane coflowing jets of air (initial sector) according to the experimental data of Zhestkov, et al. 4 To Tenip. profile according to Borodachey |-—Temp, profile in sub~ merged jetaceont= *% ng to Stark 4 / / x 5 Ff EE. |---Velocity profile / |. according to as| Borodachey~ *j " fe / | 225] we SOF a Fig. 1.22, Dimensionless temperature in main region of axially symmetric jet spreading through coflowing air stream according to Borodachev’s experimental data, Figure 1.22 shows a universal profile for the dimensionless temperature difference in the main region of an axially symmetric air jet when there is a coflowing external air stream, to) ye30 Turbulent Jets of Incompressible Fluid obtained in Borodachev’s experiments (dissertation) at a velocity ratio m=u,/tc=0,.185 (uo~125 m/sec), and a temperature difference in the initial section of the active jet 7>=320° and a temperature of the coflowing stream of about 20°C; the temperature fields were determined at different distances from the nozzle (up to 40 diameters). For purposes of comparison, Figure 1.22 shows an experimental temperature profile (dots and dashes) in a submerged jet, taken from Figure 1.16; furthermore, Figure 1.22 shows the velocity profile from Borodachev’s experiments (broken line). Evidently, the dimensionless profile of the temperature difference depends scarcely at all on the presence of a coflowing stream of fluid, but differs from the velocity profile, Figure 1.23 shows temperature differences obtained by Zhestkov and others for the plane~parallel mixing zone of two jets, described above, at the following ratios of initial velocity and corresponding initial temperatures m=03, f= 51°C, = 32°C; m=05, t=50°C, t= 38°C. The ordinate plots the ratios AT/A To, in which AT=t—t,, while the abscissa plots the same values of ay,/Ay,as in Figure 1.21, For purposes of comparison, Figure 1.23 shows the temperature curve (dotted line) taken from Figure 1.14 and referred to the boundary layer of the initial region of a submerged axially symmetric jet; furthermore, Figure 1.23 shows the dimensionless velocity curve (solid line) for a boundary layer taken from Figure 1.21. Bi] y'D, | 10 = 08118, oY “ PA Temperature profit faccording to Zhestkov .e same in submerged jet ac-| 22] yfpiing to Abramovieh (3) / elocity profile according to x 1_Zhesthov et ale Wit 08-06 08 00k SCC, % Fig. 1.23. Dimensionless profiles of temperature difference in boundary layer between two plane co- flowing jets of air (initial region) according to the data of Zhestkov et al.General Properties of Turbulent Jets 31 It follows from Figure 1.23 that the dimensionless temperature field in the plane parallel boundary layer is universal and not a function of the velocity of the coflowing stream; furthermore, it does not coincide with the dimensionless velocity field. 4x! ae 4 aE Concentration profile according 4E| fo experiment (14) Sie| —— Temperature proiile (see aA Fig. 1.22) ~~ velocity prose (see Fig. 1.18) as| as | a wa Fig. 1.24, Dimensionless profiles for concentration of admixture by weight (helium) in main region of axially symmetric air jet in coflowing stream of air according to Forstall and Shapiro [14]. Figure 1.24 shows a universal field of dimensionless values for helium concentration by weight, obtained in the main region of an axially symmetric air jet with an admixture of helium spread- ing through a coflowing stream of air (1, : The experiments carried out by Forstall and Shapiro [14] with the same experimental equipment on which were obtained the previously considered velocity fields were made at an external flow velocity u,=0.5% and a nozzle diameter 12.7 mm, The initial helium concentration in the jet was %=0.1, Figure 1.24 shows the dimensionless temperature difference, taken from Figure 1.22, which almost coincides with the concentration profile. ‘The experimental data that we have given shows that when a turbulent jet spreads through a coflowing flow of fluid, the profiles for the admixture concentration in the cross sections of the jet are similar to the temperature profiles and dissimilar to the32 Turbulent Jets of Incompressible Fluid velocity profiles; the presence of a coflowing stream has no effect on the nature of the velocity, temperature, or concentration distribution in the cross sections of the jet. Let us now consider the experimental data available, which are regrettably very scant, on the velocity and temperature dis- tribution in a jet interacting with a counterflowing stream of fluid, for example, in the initial region of flow beyond a poorly stream~ lined body, beyond the base of which thereare reverse stream lines, velocity scale £ Fig. 1.25. Velocity profiles in wake behind plane blunt body according to experiments carried out by Abramovich and Vafin, Figure 1,25 shows the dimensionless velocity profiles behind the base of a two-dimensional semistreamlined body situated symmetrically with respect to the axis of a plane channel of constant section: i =1(4): here y is the distance from the channel wall; i=H—B Is the initial thickness of the jet emerging from a slot formed by the wall of the channel and the surface of the body (H is the half-height of the channel, B is the half-thickness of the body); w is the local stream velocity; 1 is the velocity in the initial section of the jet. Points have been taken from experiments carried out by the author in collaboration with F. M. Vafin; the points were obtained for channel sections located at different distances from the base of the body (x/B=1, 2, and 2.88), at which the relative velocity of the reverse flow varied between —u,/ui=0 to 0.35. Similar experl- ments were carried out at several ratios of the body thickness to channel height: @=B/H~ 0.4; 0.5; 0.6; 0.75; 0.83. In width the body filled the whole channel. Figure 1.26 shows these velocity profiles in the form bu dye mF ( Aye )s in which 4 uj~u,—u,, and u; is the maximum velocity in the jets 42 is the minimum velocity of the back stream in the same sectionGeneral Properties of Turbulent Jets 33 (u:<0); Aye is the distance between the given point in the section and the point with velocity Au,=0,5 Au; Ay, 4s the distance between the points with velocities Au=0.9 up and Su=0.1Au9. Figure 1.26 also shows the velocity distribution curve in the mixing zone of the initial sector of plane submerged jet (broken line), taken from the experiments of Albertson et al. (Fig. 1.9). du lau, fe oe [ss '=246|According to|28—]—*| vy fAbramovich | |Z land Vafin we (0 fleas, Go? lage IS, Fu D fra, Ft igs a ~Submexged jet accords } ing to Albertson, 2 etal. (7} 2. ° bas {sot | a ee ae ae as as % iy Fig, 1.26. Dimensionless surplus velocity profile behind plane poorly streamlined body according to Abramovich and Vafin. The experimental data shown in Figure 1.26 indicates that the dimensionless velocity difference profiles in a plane jet bordering on a backstream of fluid are universal and are practically un- changed from those obtained for a submerged jet or those for a Jet spreading into a coflowing stream (see Fig. 1.21). Figure 1.27 shows the results of similar measurements made by the author in collaboration with F, M. Vafin in an axially symmetric channel with a body of circular section; measurements were made at different distances from the body (x/B=0.67; 1.33; 2.2) at a relative body section area ?—(B/R)?=0.36 (B is the radius of the body and R is the radius of the channel); the relative velocity of the backstream varied along the length between —t,/4;=0 to 0.4. The dimensionless surplus velocities obtained in these experiments are shown in Figure 1.28 which also plots an averaged velocity curve (broken line) for the initial region of an axially symmetric jet, taken from Figure 1.7. Thus, the dimension~ less velocity profile in the mixing zone proves to be universal and equally applicable to jets spreading through coflowing or counterflowing streams as well as to submerged jets. Similar results are obtained for the combustion chamber of gas turbine engines. At the beginning of the chamber there is usually a large region of reverse flow adjoining the axis of symmetry.34 ‘Turbulent Jets of Incompressible Fluid LULL LLLLLLLLL LLL LLLLLOLLML 2 FE Velocity scale a eee Fig. 1.27. Velocity profiles in wake behind axially symmetric poorly streamlined body according to ‘Abramovich and Vafin. 3 |———Velocity profile in il submerged jet Fig. 1.28. Dimensionless surplus velocity profile in wake behind axially symmetric hody according to Abramovich and Vafin. Typical profiles for the axial components of velocity in different sections of a chamber, obtained during “cold blasting” (without combustion) by Mikhaylov [18] are shown in Figure 1.29. Figure 1.30 shows the same profiles, but in dimensionless coordinates (as in Figs, 1.26 and 1.28), and it also shows the corresponding averaged velocity curge (broken line) in the initial region of an axially symmetric submerged jet (Fig. 1.8). In this case as well, the dimensionless surplus velocity profile proved the same as for a submerged jet. Figure 1.31 shows the flow pattern obtained by Vulis for the discharge of an air jet from a 20-mm nozzle into a counterflowing air stream emerging from a 100-mm nozzle; the velocity of the jet WAS tip = 63 m/sec, the velocity of the counterflowing stream u, varied between 24 and 30 m/sec; in this manner it was possible to vary the velocity ratio over the range m= 0.38 to 0.47. Figure 1.31General Properties of Turbulent Jets 35 a Fig. 1.28. Velocity profiles in combustion chamber of turbojet engine (with cold blast) according to Mikhaylov’s data [18]. Section PF + laccording to Wo + [Mikhayloy [18] = 1 / Experimental curve accord, ‘ing to Abram (see dQ ° Fig. 1.8) & er on 8 ares wie ota? ms F wa o T ee 4 Fig. 1.30. Dimensionless surplus velocity profile in initial sections of turbojet engine combustion chamber according to Mikhaylov’s data [18]. corresponds to a case in which m=0.38, Figure 1.32 plots data from Figure 1.31 for the dimensionless profiles of surplus Velocity in the main region of the jet in the form in which Au-u—u, is the surplus velocity in the jet; Suom—u is the difference between the velocity on the jet axis and the velocity of the unperturbed counterflowing stream (u2<0).36 Turbulent Jets of Incompressible Fluid Seales: Velocity 12 _# m/sec Length 242 gem Fig. 1.31. Velocity profiles in different sections of an axially symmetric air jet in a counterflowing air stream according to Vulis’s data (19). Hb FP Big | fe. ZSeetion|Acconting to a a= J Vulis (19) Velocity profile ace cording to Tripel (5] as ~: “58 “AF a's Fig. 1.32. Dimensionless surplus velocity profile in an axially symmetric jet in a counterflowing air stream according to Vulis [19]. The figure also shows an experimental profile (broken line) for a submerged jet, taken from Triipel’s data; the dimensionless pro- file for the surplus velocity inboth cases is approximately the same. 1,10, Spread of a Turbulent Jet into a Coflowing or Counterflowing External Stream ‘The spread of a turbulent jet into a coflowing stream can be assessed in the same way as for a submerged jet in Section 1.4 [20, 21].General Properties of Turbulent Jets 37 Let us assume that the rate of increase in the thickness of the boundary layer is proportional to the perturbation component of the transverse velocity which, in its turn, is proportional to the transverse gradient of the longitudinal (principal) velocity of the flow du ay wt in which / is the turbulent mixing length, In view of the similarity of the velocity profiles in different sections of the boundary layer, the transverse gradient of the longitudinal velocity is proportional to the difference in velocities on the boundaries on the basis of which, of ~t ua) oe But it follows from this same condition (the similarity of the profiles) that the ratio between the characteristic linear di- mensions is constant —L = const. ° Thus, the perturbation component of the transverse velocity is proportional to the difference in velocities at the boundaries of the layer: Uw ny uy and, furthermore, in whichGeneral Properties of Turbulent Jets 39 in which tol ey 1 Joel” ee The constants in Eq. 1.39 can be determined from the results of the study of a jet spreading through a stationary medium (12-0) when the following equality holds b= ex. (1.40) In the general case (u-70), the thickness of the boundary layer is determined on the basis of Eqs. 1.38, 1.39, and 1.40: (1.41) In a particular case of two coflowing infinite jets, the velocities on the layer boundaries have the same sign; therefore, ee ate by a ae ) the minus sign being taken when u0), the angle of thickening of the boundary layer decreases as the velocity of the coflowing stream increases: = Where m= on (1.44) Equation 1.44 is valid only at m<1. Indeed, as has been pointed out earlier, the thickness of the mixing zone is a positive value and is determined by the absolute velocity difference on its boundaries, irrespective of the sign that this difference may have. Thus, in the general case of coflowing motion of two jets, the40 Turbulent Jets of Incompressible Fluid following expression, derived from Eq. 137 and also 1.39 and 1,44, holds: do Pee ES (1.48) Here, the plus sign is taken at u;>u,, and the minus sign at —462°K, obtained by Zhestkov and others. The velocity of the heated jet amounted to about 100 m/sec, while the velocity of the cold jet ranged from 0 to about 60 m/sec; the corresponding velocity ratios m for each photograph are shown in Figure 1.33. _ edie Fig. 1.33. Shadow photographs of a boundary layer between two coflowing plane jets.General Properties of Turbulent Jets 41 We can see that during the increase in m from 0 to 0.34 the thickness of the boundary layer, which is clearly visible in the photographs, drops appreciably, and remains almost unchanged as m is further increased. Figure 1.34 shows a theoretical curve for Eqs. 1.43 and 1.44 for a range of m from ~1 to 0 and from 0 to +1: =1 for m<0 and for m>0, In view of the difficulty of determining the true thickness of the jet boundary layer 6 by experiment, here, as in Section 1.9, we use the distance y, between the points with surplus velocities 4uy=0,9 Aup and 4u;~0.1 44; for a universal velocity profile, 4y, comprises the same proportion of the boundary layer thickness in all cases in which 6, is the boundary layer thickness of a submerged jet and 4y,, is the value Ay, for a submerged jet. nar KH a8 | oAccoriing to a5 xageording wo” 7 ; j— eadesidisy ae Abramovich and x afin a2 8 a aR wat ae Fig. 1.34. Thickness of boundary layer of incom- pressible fluid jet as function of outside stream velocity.42 Turbulent Jets of Incompressible Fluid ‘The experimental points plotted on Figure 1.34 have been borrowed from the experiments carried out by Zhestkov and others, described in Section 1.9, and Yakovlevskiy [21] and others. All the experimental points are close to the theoretical curve in the region -0.40.5, This result, likewise detectable in the photographs in Figure 1.33, can be explained in the following way. When deriving Eq. 1.44, we assumed that the turbulence in the boundary layer was caused only by the velocity difference on its boundaries, and that beyond these boundaries it did not exist at all, ie, we assumed that as m>1, v'>0. But in actual fact there is 2 certain initial turbulence in an unperturbed stream as well. Hence, in cases in which the velocities 1, and 1 are close to each other, i.e., when the intensity of the turbulence in the jet is less than that of the unperturbed stream, the influence of the jet ceases and the growth is determined by the turbulence of the unperturbed stream which is not a function of m; it is natural that in this regime the angle of thickening of the boundary layer is hardly associated with the velocity ratios of the layer boundaries. The results obtained for the mixing zone for two infinite jets also relate to the initial region of a jet of finite thickness spread- ing through a coflowing or counterflowing stream, since in the initial region the velocities remain unchanged on both boundaries of the mixing zone, The outer edge of the main region of a jet of finite thickness with counterflow of the fluid surrounding the jet remains the same as in a submerged jet, because the velocity of the counterflow has no effect on the angle of thickening of the mixing zone. It is a more complex task to determine the outline of the main region of the jet in a coflowing fluid, In this case Eq. 1.45 acquires the following form: tb natn eee? (1.48) in which u,, is the velocity on theaxis of the main region of the jet, 4, is the velocity of the coflowing stream (the minus sign is taken when u,>u,)- Since up varies on the jet axis, i.¢.,1,—/(x), the jet boundary in a coflowing stream must be curved: a var. (1.49) dr In order to determine the jet boundary, the form of the function u(x) must be known, and can be obtained from the condition of conservation of momentum in the same way as was done in Section 1.6 for a submerged jet. The solution of this problem will be given in Chapter 5, which deals with the theory of a turbulent jet in a coflowing fluid.General Properties of Turbulent Jets 43 1.11. Turbulence Characteristics in a Free Jet In the theory of turbulence in general, and in the theory of a turbulent jet in particular, the laws governing the distribution of different pulsation characteristics of the stream and their inter- relationships are very important. The perturbations of the velocity vector (or one of its components), averaged with respect to time, is equal to zero; hence it is usual to determine the turbulence of the flow from the nondisappearing mean-square pulsations in velocity, temperature and other parametersV a? ,V 07, Vr?, and so forth, The interrelationship between the turbulent characteristics is determined by the mean values of their products (correlation) ‘Wv',uf, vf , and so on. The averaged mixing length, which in accordance with Section 1.4, can be determined in the following way, Serves as the measure of the scale of turbulence: ae oy y 7 Below we give the results, obtained by Antonova, of an experi- mental study of turbulence characteristics for a free heated turbulent jet. Jn the’ experiments the mean velocity fields in the sections of the jet were determined, and the velocities were measured both with a T-shaped probe and also ahot wire, The pulsation velocities of the stream and the mean and pulsation temperatures were measured with a hot wire having a single-filament tungsten probe with a filament diameter of 20 microns, and a ETA-1A device developed by P. V. Chebyshev. The measurements were taken a the following conditions and in the following sections of the jet: jet parameters upon leaving nomle Main, Tnitial constton [Paar | Tee | rea tg m/sec re Re. = + | 7 ee i ieee we | aa uf» | 1s | sim | a The results of the measurements of the mean velocities and temperature differences in various cross sections of the jet are shown in dimensionless coordinates in Figure 1.35a,b, and c, Here, the local velocities and temperature differences, respectively,44 ‘Turbulent Jets of Incompressible Fluid are referred to the velocity and temperature difference on the axis of each section: S) aay 713): in which y is the distance between the jet axis and the point in~ vestigated in the main region, and the distance between the nozzle edge and the point investigated in the initial region; 6, b is the half-thickness (radius) of the jet at the given cross section of the main region or the half-thickness of the boundary layer in the initial region. These graphs also show the gradients of velocity and temperature. & PT Section Condition asso % ay) | e st iii) aaafs- X \ zs] iP) _t ae ty oe OS a1 b a a gain UY a7h) PDA | ses atthe oaaens | 0 t/d=225 / 9 xfda425 as\s ‘a(a/a,) ed Oya «sas aE c y Fig, 1.35. Fields of mean velocity, gradient of mean velocity, mean temperature, and gradient of mean tem- perature in jet (a and b are in main region, and c in initial region).General Properties of Turbulent Jets. 45 The results of measuring the turbulence characteristics are shown in Figure 1.36 in the following dimensionless form: Ve STm im co as functions of the dimensionless ratio 2. Here, VEY, V? are the mean-square perturbations of the longitudinal and transverse components of the velocity vector and of the temperature, while are the correlation coefficients between the corresponding turbu- lent pulsations. Boundary layer oral, hea ais} \ 7st r a5| as Ne g 1 *Condition/| { cc a2 t e |: a | ee OS 025 0 ~025 -05 ¥ a a ’ q > ,___ Boundary layer ca Condition? | i Section | | x0 z/da025 | Lo Bhg 06 a/d-22s | a: 9 0 2/d-42s F i oF + | | ay Os ie “OF x © 5 a a Fig, 1.36. Pulsation fields for longitudinal velocity component in sections of nonisothermal jet (a and c ih the main region, b and d in the initial region).46 Turbulent Jets of Incompressible Fluid The measurements show that the mean-square pulsations of the longitudinal component of the velocity vector in the cross sections of the main region of the jet can be represented with a fair degree of accuracy by the curve shown in Figure 1.36a and b. Since the range over which the Reynolds numbers were varied is fairly narrow, no conclusions with regard to the similarity of the mean- square pulsations of the longitudinal component of the velocity vector can be drawn on the basis of these experiments. If it is assumed that the following equality holds: a ay * (1.50) = ly then we can find the value of /,, known as the mixing length. The mean velocity gradient with respect to thickness of the section in the main region of the jet is shown in Figure 1.352. The dimension~ less quantity/, /b, found from the given experiments and from Eq, 1.50, remains roughly constant for most of the cross section and is 0.12 to 0.13. In the initial region of the jet, the distribution of the pulsation of the longitudinal component of the velocity vector, shown in dimensionless coordinates =s(<). can also be described reasonably accurately by a single curve (see Fig. 1.36 b and d). The assumption of the validity of Eq. 1.50 in the boundary layer of the initial region made it possible to de- termine |, /b1, which proved to be approximately 0.1 for most of the cross section of the boundary layer. The mean-square puleations in the transverse component of the velocity vector in the Ccetions of the main region are shown in dimensionless coordinates j/* 7, and y/b in Figure 1,37a and c; they can also be described reasonably accurately by a single curve (Fig, 1.372 end c) over the range of Reynolds numbers under investigation. Assuming the validity of the relation oi we can also find the mixing length /.. The ratio /,/ is more or less constant for most of the section, and is equal to 0.1. The mean-square pulsations in the transyerse component of the velocity vector in the boundary layer of the initial region are shown in dimensionless coordinates 7/0 / ig and y/b,in Figure 1.37 and d. The mixing length of the transverse velocity vector component inGeneral Properties of Turbulent Jets aT Fig. 1.37, Fields for pulsation of transverse velocity component in sections of nonisothermal jet (a and are in the principal region; b and d in the initial region). the boundary layer of the initial region is approximately equal to 0.088,. The distribution of pulsations in the longitudinal and trans- Verse velocity vector components are close together and identical in nature. ‘The mean-square pulsations in temperature in the sections of the main region of a free heated jet are plotted in Figure 1.384 sad b in dimensionless coordinates. The considerable spread of the experimental points in Figure 1.38 suggests that there is no universal curve for the temperature pulsation profile in different cross sections of the jet. The correlation coefficients for the pulsations in the longi- tudinal component of velocity and the temperature, measured in cross sections of the jet, are shown in Fig. 1.39. Similar experi- ments for determining the turbulence characteristics in the main region of a free nonisothermal jet are described in References {9 and 55]. The dimensionless pulsations given in these publica- Hons coincide with those described above, except for the correla- tion coefficient between the pulsations in the longitudinal component48 Turbulent Jets of Incompressible Fluid Boundary layer vet Condition Condition | ca +0 6 See Saad Be loa acess 3 seed ° xo 2/da1n0) eae az el|. +o ge {8 %,| e wee slo nop te [°o a e omod fe 7 ° | ay a a a a e b Fig. 1.88, Temperature pulsation fields in sections of nonisothermal jet (a is principal region, b is initial region). Kat fas | gla fb JRE af e te 4a : | Conaigon | Section exfda226 | Key oafda$zs | asb— toe + rar 1 8 of 6 | ° a 4, i e ae f | "OF ar 7 ar wOF z ’ g Fig. 1.39. Variation in coorelation coefficients Kis, Ky in sections of nonisothermal jet (a—principal region, b—initial region).General Properties of Turbulent Jets 49 Initial region (Condition 11) he i Boundary layer 4 T | Condition It | Section ! x0 zid-a25 | i oo ayie225 { as| { Ob zd=4.25 } | 7 q OF ~ a 06 rtok7s x0 z/d=100 Fig. 1.40, Variation in correlation coefficient Ky in sections of nonisothermal jet (a - principal region, b ~ initial region). of velocity and the temperature K,,. In Antonova’s investigation (at x/d = 7.5 to 10) Ky, ranges between 0.6 and 1; in [39] ( x/d= 15 to 20), it varies between 0.3 and 0.4; and in [55] ( x/d= 150 to 300), it ranges from 0,3 to 0.6 and attains approximately 0.9 in some sections. The difference in X,, arises from the difference in the methods used. In [39 and 55) if was actually the correlation coeffi- cient between the pulsations in the velocity vector and the pulsations in temperature K,, that was measured, rather than K,,, as in Antonova’s work. ‘The correlation coefficients Ky, Ky, and K, (Figs. 1.39 and 1.40) are expressed by universal curves; i.e., at corresponding Points in any two sections the correlation coefficients are corre- spondingly identical, Furthermore, within the range of Reynolds numbers investigated, the correlation coefficients Ky», K,,, and K,, do not vary in magnitude within the main and initial regions of the jet, respectively,Chapter 2 Theory of Free Turbulence for the Case of a Submerged Jet 2.1. Theory of Free Turbulence At the present time several theories of jet or so-called free turbulence are known: Prandtl’s old theory [8], Taylor’s theory [22], Prandtl’s new theory [23], Reichardt’s theory [10], Mattioli’s theory [25], and so forth, We shall consider only the first four, which are most widely used and successful, Prandtl’s “old”? theory of free turbulence was published in 1925, and Tollmien applied it in 1926 to solve the following three problems concerning the propagation of free mixed jets of an incompressible fluid [24}: 1, The boundary layer of an infinite plane-parallel jet, 2, Parallel jet issuing from a very narrow orifice, 3, Axially symmetric jet issuing from a very smallorifice. In 1929-1930, Swain [26] and Schlichting [27] extended Prandtl’s theory of free turbulence to the case of wake flow behind a stream- lined body and worked out the laws of flow in axially symmetric (Swain) and plane (Schlichting) wakes applicable at flow cross sections, remote from the body, The work of Tollmien and Schlichting included experimental investigations of velocity distribution in streams, By using only one experimental constant they were able to establish agreement between these theories and the experimental results, This agreement confirmed Prandtl’s theory of free turbulence, and its application came in a number of papers, both theoretical and applied, In particular, Kuethe [28] worked out an approximate method for the -computation of the velocity profile in the initial part of a round jet. The investigation of free turbulent flow is of great interest because it is the simplest case of turbulence least dependent on effect of viscosity. The study of free turbulence may be regarded as a preliminary stage in the study of turbulent flow. The successful development of the mechanics of turbulent jets stimulated application of the theory of jets to the solution of heat problems: the problem of temperature distribution along the axis of the jet and in its lateral cross sections, and the problem of heat diffusion from the jet into the adjoining space, and so on, In addition, a direct conclusion of Prandtl’s theory for the cases of the free jet and of the wake behind a body is the complete similarity between 50Theory of Free Turbulence for a Submerged Jet DESSS 51 u : Ways FE velocity and temperature fields, To check.this result, Fage and Falkner [22] in 1932 experimentally investigat i ~ yelocity of flow and temperature of a gas in the wake behind long circular and lens-shaped cylinders, They proved that the theoretical Prandtl-Schlichting velocity fields are well verified by the experi- mental data, whereas no similarity between the velocity and temper- ature fields occur, and heat transfer from a wake in streamline flow is two times greater than predicted by Prandtl’s theory. Taylor, after noticing a contradiction in Prandtl’s theory of free turbulence, assumed that the turbulent shearing stress in a stream must be determined by the lateral transfer of vorticity and not, as Prandtl assumed, by momentum transfer, Thus, Taylor removed an inconsistency of Prandtl’s theory, namely, its neglect of local instantaneous pressure gradients that substantially influence mo- mentum exchange but do not affect vorticity transfer. On the basis of this hypothesis and using, as Prandtl did, only one experimental constant, Taylor obtained velocity and temperature profiles in the wake behind a long cylinder, which are confirmed by the experiments of Fage and Falkner, Whereas Taylor’s theory gave the velocity profile no more accurately than the Prandti theory did, Taylor’s theory is built on more justifiable assumptions, This circumstance allowed retention of all the results in the field of jet mechanics based on Prandtl’s theory, and forced at the time [3] preference for Taylor’s theory in further development of the problem of the free jetand the wake behind a body, especially in cases that involved temperature profiles and heat transmission, In 1942 Prandtl published a new theory of free turbulence, in which Newton’s law of viscous friction was used, as Bussinesq had suggested in the nineteenth century, to determine the turbulent shearing stress in a jet. He assumed the coefficient of ‘turbulent viscosity” to be proportional to the product of density, velocity difference, and jet thickness, Using Prandtl’s new theory, Gértler [23] solved anew two problems of the theory of the submerged turbulent jet: the problem of the boundary layer of an infinite parallel stream, and the problem of a plane jet with an infinitely harrow initial cross section, If Prandti’s old theory of free turbu- lence led Tollmien to an understanding of a jet of finite thickness, then by applying Prandtl’s new theory, Gértler obtained asymptotic velocity profiles (Gaussian) at the lateral cross sections of the jet, characteristic of an infinitely thick jet. The main advantage of the new Prandtl-Gértler theory of the flow of a jet compared with the old Prandtl-Tollmien theory is that in the former, velocity profiles are described by relatively simple analytic formulas, while in the old theory they were determined only numerically, Transforming to the problem of heat transfer without the introduction of an additional experimental Constant is just as impossible in the new Prandtl-Gdrtler theory as in the old Prandtl~Tolimien theory; in this respect Taylor’s theory maintains an advantage over both of them, All these theories of free turbulence are discussed below, Qa ae, ES 99GB52 ‘Turbulent Jets of Incompressible Fluid 2.2. Prandtli's Old Theory of Free Turbulence. ‘The theory of the boundary layer of an incompressible fluid can be obtained as a special case of the theory given below for a com- pressible gas, if the density gradients are taken as zero. It is more convenient and descriptive, however, to begin with the simplest case. For two-dimensional flow of an incompressible fluid, the equation of motion in the x-direction in the boundary layer has the following form (Fig. 2.1): Hy My My eed (2) where wand » are the instantaneous values of the velocity com- ponents; p, p, and » are the instantaneous values of density, pres- sure, and kinetic viscosity, respectively. Fig, 2.1, Boundary layer of a jet. The equation of continuity for the two-dimensional flow of an incompressible fluid has the form on oy Multiplying by u, we obtain the equation : ae i ae te a0, which can also be written in the following form: pecs Ovee ee (2.3) Grieroeemea vee ater After substituting Eq. 2.3 into 2,1, we have oe ado a fey Oe oe : (2.4) or ay oxtheory of Free Turbulence for a Submerged Jet 53 ‘As was shown in the first chapter the velocity components of turbu- lent flow and also the pressure can be decomposed into a time average and a fluctuating component u=itu', v=o+0', p=pty, where the time averages of the fluctuating components vanish: (2,5) In general, however, this cannot be extended tothe square of the fluctuating values and their products, Letus replace the instantane~ ous values in Eq, 2.4 by the mean and fluctuating values, viscosity and density being assumed constant, After averaging with respect to time and taking Eq, 2.5 into consideration we obtain ae | at ate oe (2.6) ey When a low-viscosity fluid flows around a solid, the flow can be divided into two regions: a comparatively thin layer of fluid located close to the solid walls—a boundary layer where viscosity, however small it may be, must be taken into account, and the other part of flow in which viscosity can be neglected and hence the flow can be described by the laws of an ideal fluid, The boundary layer of the turbulent flow itself is assumed to consist of a very thin sublayer bordering directly on the wall, with pure laminar flow and the remaining (turbulent) part of the boundary layer, in which the effect of viscosity is not significant. Thus studies of flow around bodies, fluid motion through tubes, and in general viscous flow of a fluid in the presence of solid boundaries, in principle, do not permit the complete neglect of viscosity, This is a great obstacle in developing a theory of turbu- lent flow, A special characteristic of turbulent free jets is the absence of solid boundaries of flow, and consequently also of a laminar sub- layer, which makes it possible to neglect completely the influence of viscosity in all cases of free turbulence and also explains the self- Similarity of jets, independent of the Reynolds number over a very broad range, Free jets, expanding into an infinite region, filled witha quiescent fluid, and also wakes surrounded by an infinite undis~ turbed stream, as indicated above, possess such small pressure gradients that in most cases they may be neglected. The equation of motion for the two-dimensional free flow of an incompressible fluid, stationary with respect to mean velocities (Gr =» can be written in the following manner: (2.7)54 Turbulent Jets of Incompressible Fluid After a change to mean and fluctuating velocities in the equation of continuity (Eq. 2.2), ia One Oe nor. ety tata (2.8) de dy | Oe It follows from Eq, 2.5 that with a time average the last two terms of this equation vanish, that is, the conventional form of the equation of continuity is valid for the mean velocities: Hy Le (2,9) ox oy Multiplying Eq. 2.9 by1, gives - di = 88 ae tay 79 from which follows an equation analogous to Eq, 2.3: a dus _ = Ot ye li ae ae tay x ay Substituting this expression into Eq, 2,7 leads to the following equation for the two-dimensional turbulent boundary layer of a free jet: a ae pe poe ae tay (2.10) Ox a au® a It is not difficult to show that the term —— can be neglected, because the velocities and fluctuations of velocity change much more slowly along the flow than in the transverse direction, whereas, the magnitudes of u’ and v’ are of the same order. Thus, au au +0 2,11) ates (2.11) For simplicity, the bars over the mean values of velocity are dis~ carded, that is, u and v signify the time-averaged velocity com- ponents, It is well-known that the equation of motion for the two- dimensional steady isobaric flow of an incompressible fluid can be represented in the following form: gy Oe 1 Oey 2.12) " ox aeos ea” ( where +,, is the shearing stress acting in a plane perpendicular to the Oxy plane,Theory of Free Turbulence for a Submerged Jet 55 In the case of a pure viscous laminar motion, we obtain from a comparison of Eqs. 2.12 and 2.1, when 2 , the well-known formula wae, where p=» is the dynamic viscosity coeffi- cient, If we compare Eq, 2,12 with Eq, 2,11 for pure turbulent (non- viscous) motion, we arrive at the original relation of the Prandtl theory for the so-called apparent turbulent shearing stress: (2.13) fey = — PHD In turbulent mixing of neighboring layers of fluid, the product m= pv’ corresponds to the fluid mass passing per second through a unit area of the surface separating the neighboring layers; since u’ is the fluctuation of the longitudinal velocity, Eq, 2.13 indicates that in Prandtl’s old theory of turbulence the turbulent shearing stress is defined by the lateral transfer of the longitudinal component of momentum, As already indicated in Section 1.4, Prandtl assumed that the transverse fluctuating velocities are proportional to the longitudinal fluctuating velocity o’~Fu’, The signs of the fluctuating velocities are opposite when S50 and the same when 2 <0, If a >0, then y iy particle transfer when v’>0 is accompanied by a negative longitudinal fluctuating velocity, that is w<0, and when o'<0:u’'>0. This discussion for ee< 0 leads to the same signs foru' and vo’. y The longitudinal fluctuating velocities are determined in turn from the fact that during a transverse displacement of a mole of fluid it preserves its initial velocity u until its transverse dis- Placement ceases, after which the velocity of the mole jumps to the velocity of the new layer of fluid; for a continuous velocity distribution, the change over the distance of a transverse mole length equals ou Yt] ——]5 2. w 4 a (2.14) the mean transverse mole length ! was called by Prandtl the “mixing length,” From Eq, 2.14 we obtain the absolute value of the velocity fluctuation, In accordance with Eqs, 2.13 and 2.14, we obtain the following well-Inown Prandtl formula of turbulent shearing stress cas %y ou is (2.15) ty = tel Here the sign is plus when 5 >0, The proportionality factor be- tween the fluctuation components of velocity is contained in the magnitude /, Substitution of Eq, 2.15 into 2,12 establishes the56 : Turbulent Jets of Incompressible Fluid dependence of the mixing length / on the coordinate, The essence of Prandtl’s old theory of free turbulence isthe following: Because of the absence of walls (in the case of a jet), near which the mixi length usually decreases, the mixing length is assumed constant in the transverse direction 1(y) = const. (2,16) In the longitudinal direction, to account for the similarity of velocity profiles, the mixing length must vary in proportion to a characteristic length, that is, for example, the width of the jet, Thus //o=const. But as was shown in Section 1.4, the width of a submerged jet is proportional to the distance from the pole of the jet. So, if the flow of the jet is directed along the x-axis, then lecx, (2.17) which upon substitution into Eq, 2.15 gives tym ee S| % I, (2.18) where c is the single empirical constant of Prandtl’s theory of free turbulence, Substituting 2,17 and 2.15 into Eq, 2,2 givesthe equation of two- dimensional motion of Prandtl’s old theory of free turbulence: 22 Ou ou eeeren ut 4 om + 2c : ay aoe (2.19) from the solution of which Tolimien [26] obtained in 1926 the velocity distribution in the boundary layer at the edge of the initial region of a plane-parallel jet, The Tollmien theory is explained in Section 2.4, 2.3. Application of Prandtl’s Old Theory of Free Turbulence to Heat and Diffusion Problems. To determine the law of temperature distribution in the boundary layer of a parallel free jet, we will use the differential equation of heat transmission in the boundary layer of the flow of a fluid: oT oT oT X ar x ‘ou\? Figg Figg Fed OT ize 2,20) Cyr yo On + wsalS)> oo where 7 is the temperature of a fluid particle; 1 the coefficient of molecular thermal conductivity; cy the heat capacity of the fluid; g the acceleration due to gravity, The right side of Eq, 2,20 contains the heat equivalent of the mechanical work of the viscous force (molecular friction) . ( ju y Begg \ ayTheory of Free Turbulence for a Submerged Jet 57 and the molecular heat flow gradient (2.21) where W is the amount of heat transported by molecular thermal conductivity per unit time through a unit area, Next we shall mul- tiply the equation of continuity for the instantaneous velocity by the instantaneous value of temperature: au a0 Teste a ty After manipulation of the resulting expression, we obtain QT) 4 80T) yyy T Sts ox a ay Using this equality, we modify the left-hand side of the Eq. 2,20: OT, WT) 4 WOT) dT yy pany ee ize lis)* a ‘dx ‘oy Papo) ay (2.22) In Eq, 2,22 we replace the instantaneous values of temperature and velocity by the mean and fluctuating components; we perform a time averaging and assume the flow to be steady relative to the mean values of velocity and temperature, As a result, at constant values of 4, c,,p and v, we obtain aT. oT ox + ay (2.23) -St+aa( pee ay? § $icpe Molecular thermal conductivity and the transformation of the work due to viscous forces into heat may be neglected, because they are small in comparison with the turbulent transport of heat, just as in the dynamical equation (Eq. 2.7) where we neglected viscous friction in comparison with turbulent friction, Consequently, we have the following equation for the thermal boundary layer of the plane-parallel jet: eG), aD, a@r) , avr) ty tae tae (2.24) ‘Thus, heat transfer in a turbulent flow is a result of the mixing of layers of fluid with different temperatures, that is, because of correlation between fluctuations of velocity and temperature, Multi- plying the equation of continuity for a stationary average flow by T, we obtain (2.25)58 Turbulent Jets of Incompressible Fluid With use of this equation it is easy to prove that —@ ,.F _ 3GT a@7) ot ao tay (2,26) Combining Eqs. 2,26 and 2,24 and discarding the second-order aut term (ee : (2.27) The equation of the thermal boundary layer of a turbulent jet (Eq. 2.27), when compared with the corresponding equation for stationary flow or. fg ee leo, a tees (2.28) allows establishment of the law of heat transmission for turbulent mixing: W=—c, gpvT’. (2.29) ‘Thus the flow of heat in two-dimensional turbulent flow is deter- mined by the mean product (cross-correlation) of transverse velocity fluctuation and temperature fluctuation, The magnitude gp v=gm’ is the amount of matter flowing per second in the transverse direction during turbulent fluctuations, and the quantity c,7’ is the instantaneous value of the heat content difference transferred per unit weight of substance during transverse turbulent motions, ‘The temperature fluctuation is its change at the instant of loss of individuality by the fluid particle after transverse transfer over the length [: oie (2.30) a = cx, consequently we have* = au moreover, ¢’~ oy totes el Fe keal/m*.sec (2.31) where 7 is the specific gravity of the substance; the plus sign corresponds to the case # >0. Substituting Eq, 2.31 into Eq, 2.28, we obtain the equation of the thermal boundary layer, which arises at the edge of a nonisothermal plane-parallel jet ofan incompressible fluid (2.32) *The bars over the notation will be omitted, as was done in the study of the dynamic layer.Theory of Free Turbulence for a Submerged Jet 59 In the course of turbulent mixing, particles of fluid are trans~ ported from place to place, carrying heat and foreign constituents, ‘The temperature and admixture concentration appear in similar form as scalar properties of the fluid particles, Therefore, as will be seen below, the admixture distribution must be subject to the same Jaw as the temperature distribution. We shall setup the equation of the diffusion of constituents in a two-dimensional fluid flow, in which we isolate an infinitely small element whose main direction of motion is horizontal, Let the admixture concentration—its content by weight in 1 kg of fluid—be denoted by the quantity x, Then the difference between the amount of admixture carried in and out by a flow across the vertical boundaries per unit time equals 2 (yuxdy)dx, and across the horizontal boundaries oa (yoxdx dy. The difference between the inflow and outflow of the admixture caused by diffusion in a direction perpendicular to the principal direction of flow amounts to a dy dx, The rate of change of the admixture content in the volume element equals - Lm dx dy). Thus we obtain the following conservation equation for the admixture: (wr), atv) 1 0G a = Le : 2.3 ox oy 1 oy at Ce) From the equation of continuity, oe a0 ox tay it follows that Sg DO oe ay and Dur) 4 O00 yO yy Be Gh OO ma (2.34) Replacement of the left side of Eq, 2.33 with the result of Eq. 2,34 leads to the diffusion equation (2.35)60 Turbulent Jets of Incompressible Fluid which has the same form as the equations of motion (Eq. 2.12) and of heat transmission (Eq. 2.28). It is known from molecular physics that diffusion caused by the motion of molecules is expressed by the relation ox Gamer : (2.36) analogous to the laws of molecular friction and molecular heat conduction, the coefficient of molecular diffusion ¢ depending on the type of fluid and its physical state (mainly temperature), Substituting Eq, 2,36 into Eq, 2,33 we obtain the diffusion equation in the following form: 1 | Oto) , O60) _ tx 4 a a 230) Replacing the instantaneous values of concentration and velocity by the average and fluctuating values, we perform a time average and treat the case of stationary flow relative to the average velocities and concentration (3 = 0) . As a result we obtain for ¢ =const and 7=const: a Gua) GA, aw) 5 oF) 2, oe oy a oy (2.88) v ay Neglecting molecular transfer of the admixture in comparison with turbulent transfer, we have: (2.39) We note that the equations of admixture diffusion (Eq, 2.39), of heat exchange (Eq. 2.24) and of motion (Eq. 2.7) for the two- dimensional boundary layer of a turbulent jet are identical, Turbulent transfer of the admixture is realized by the transport of layers of fluid with different concentration as a consequence of correlation between fluctuations of velocity and of concentration, In order to display in explicit form the correlation terms in the diffusion equation (Eq, 2.39), the latter can be transformed with the help of the equation of continuity of the mean flow (Eq, 2.9), just as the equations of motion and of heat exchange were transformed: dt 9 5a sod Ete WER 0 and consequently, coe oo ala a Oey et o% (2.40) ayTheory of Free Turbulence for a Submerged Jet 61 Substituting this equality into Eq, 2.39, we obtain ode (2.41) a ox or We discard the second-order term(%*), Then the diffusion equation for the admixture in turbulent flow acquires the following form:* a a. Baha trag (2.42) The fluctuation of concentration is its change of magnitude during transfer of a fluid particle from one layer of fluid to another by a distance of one mixing length (: In addition, from we have (2,43) Comparison of Eqs, 2.43 and 2.35 for steady flow( 3 5 9), allows establishment of the iaw of turbulent diffusion: ou oy G= on du | a Be gy cayt| OH] 2.44) Bn sree| FF oa 2.4. Theory of the Boundary Layer of a Two-Dimensional Turbulent Jet of Incompressible Fluid. ‘The equations of motion (Eq, 2.19), of heat exchange (Eq. 2.32), and of admixture diffusion (Eq. 2.43) were obtained with the aid of Prandtl’s old theory of free turbulence, which is based on an hypothesis of the similarity of the mechanisms of turbulent trans- fer of momentum, heat, and concentration, The solution of these equations makes it possible to construct a picture of the fluid flow, and also of heat and diffusion processes in the boundary layer at the edge of a two-dimensional submerged turbulent jet of an incom- pressible fluid, *The bars over the mean values of concentrationare discarded, as in the solution of the dynamical and heat problems,62 Turbulent Jets of Incompressible Fluid It was shown in the first chapter that the boundaries of such a boundary layer and the equal-velocity lines within it are straight, All these lines intersect at the single point 0, which in the case of a uniform initial velocity field coincides with the boundary of the initial cross section of the jet (Fig. 2.1). If the problem of a free plane boundary layer is solved in coordinates saat, (2.45) the velocity depends only on 1, that is, u= uf (n). (2,46) In order to eliminate the experimental constant from the equation of motion (Eq, 2.19), we assume 2 =a? (2.47) and we choose the coordinate system (2.48) where a is an empirical constant characterizing the structure of the flow of a jet. Then the dimensionless velocity turns out to be a function of only one coordinate* w= ty F' (9), (2,49) where F’(¢) is the derivative of a certain function F(¢), which is proportional to the stream function - 9 = Judy = axuy fF" dg ari F (g). (2.50) By differentiating Eq, 2,50 the transverse component of velocity can be found: om — = ity (y F' — F). (2,51) z Substituting Eqs. 2.49 and 2.51 into Eq, 2,19, taking into account 3 a= oe : aaa on x oy ax leads, after a number of transformations, tothe fundamental equation of the boundary layer of the two-dimensional flow of a jet of an incompressible fluid, when F’+ 0: y PU +F=0. (2.52) *The quantity uw is the velocity outside of the boundary layer in the unperturbed flow region,Theory of Free Turbulence for a Submerged Jet 63 ‘This equation was first obtained by Tollmien [24], On the right side’ of Eq, 2,19 the plus sign is taken because, as Fig, 2.1 shows, a >0, The characteristic equation for Eq. 2,52, 1+ 4=0,has iy three roots: k=l, b The complete integral of Eq. 2,52 equals F(q)= Ce? + Gye cos (YS ¢) + ce? sin( YE 5). (2.53) 9) To calculate the three constants C;, Cz, Cs, andalsoto find the values of ordinates 9 and 2, which determine respectively the inner and outer edge of the jet boundary layer, the following five boundary conditions can be used: At the inner edge of the boundary layer (» = 9; ) we have a, the velocity component along the x-axis (u) equals the unperturbed flow velocity (uo), that is, according to Eq, 2.49, Fe) =1; (2,54) b, the velocity component along the y-axis (v) equals zero, that is, according to (2,51) and (2,54), F@=ei c. the gradient of the horizontal velocity component (3°) equals zero, that is, % (2.55) F'@)=0. (2.56) At the outer edge (¢ = ¢,) we have d, The velocity component along the x-axis equals zero, that is, Fe) = 0; (2.57) e, the gradient of the horizontal velocity component equals zero, that is, F* (g) (2.58) ‘Using the boundary conditions, we obtain:* r= 0.981; 92=— 2,040; C,=—0.0176; C= 0.1337; Cy= 0.6876. aie subject is discussed in greater detail by the author in64 Turbulent Jets of Incompressible Fluid Substituting these numerical values of the constants into Eq, 2.53, we obtain the following equation for the desired function Va + a + =—0.0176e"? + 0,1337e7 cos( - ey + 0,6876e7 sin( ¥3 (2.59) According to Eq, 2,49, the dimensionless velocity component along the »~axis in the boundary layer of a free plane stream equals the first derivative of Eq, 2.59: ¢) = 0.0176e"*-+ 0,6623e" cos( a V3 2 +0228" sin ( (2,60) The transverse velocity can be calculated from the known expression (Eq, 2.51): a = 7 Ser ). (2.61) 2 {Experiments of Albertson other: (7) agg 4 Fr “OF 7 OF Te Fig. 2,2, Velocity profile in the boundary layer ofa submerged jet G = x/). The values of basic and auxiliary functions which enable us to calculate dimensionless velocities, dynamic pressure, velocity gradients, and so on, are given in Tables 2,1 and 2.2, In Fig, 2.2Theory of Free Turbulence for a Submerged Jet 65 ‘Tollmien’s theoretical velocity profile (Eq, 2,60) is compared with the experimental data of Albertson and others [7], which are taken from Fig. 1.9. The value of the empirical coefficient of the structure of the jet is taken to be a= 0.09 and leads to good agreement between theory and experiment, For this value of a the thickness of the jet boundary layer is b= ax(9;—9) = 0.27%; the angle of inclination to the direction of flow of the outer edge of the layer (¢2 =—2,04) is oy are tan(2t) = are tan (az, = — 10° The angle of inclination of the inner edge of the layer (¢, =0.98) a, = are tan (a7) =—5°, The streamlines can be found from Eq, 2.59, A picture of the streamlines in a plane boundary layer of a jet is given in Fig, 2.3, The edge that separates the initial mass of the jet from the mass associated with the surrounding fluid corresponds to a null stream function },=0, i.e, F(g)=0. With the aid of Eq, 2.59 the ordinate of the surface of separation is calculated: 9 (F=0)=—0.185, which gives the angle of inclination c= are tan (a9,)— 12 From Eq, 2.18, 2.47, and 2,49 we obtain a formula for shear stress at the boundary layer 3 alr al Fig, 2.3, Streamlines in jet boundary layer. The temperature distribution in the boundary layer of a plane Jet is found from a solution of the heat exchange equation (Eq. 2,32), It was established in Chapter 1 that the temperature profile, just as the velocity profile, in a jet boundary layer is universal, and consequently the isothermals at the straight edges ofthe layer must be straight lines, that is, temperature depends only on the dimen- Sionless coordinate g=y/ar, The temperature of the fluid at rest66 ‘Turbulent Jets of Incompressible Fluid which adjoins the jet is not equal tozero, therefore it is convenient to introduce the temperature difference instead of the absolute temperature into Eq, 2.32: AT =T—Ty, (2.62) where 7, is the absolute temperature of the adjoining medium, The temperature difference inside the jet (beyond the limits of the boundary layer) is AT, =To—Tr, (2.63) where To is the absolute temperature in the undisturbed core of the flow. In the following it is assumed that the edges of the thermal and dynamical boundary layers (2, ¢-) coincide, The temperature difference in the boundary layer of the jet can be expressed as follow: AT =AT,4 (2). (2.64) Substitution of Eqs. 2.64, 2.49, and 2,51 into Eq, 2.32 and appropriate transformations lead to the equation ce er ca ae (2.65) Table 2.1, Basic Functions of the Boundary Layer of the Jet - | etre fe| 0.98 0.568 | 0.520 0.93 0.510 | 0.519 | 0:88 O.514 | 0-516 0.83 07489 | 9.311 0:78 01463 | 01506 be ova |. 0: 0,438 | 0-499 0:63 | 0.681 07413. | 0.491 0:63 | 0.636 01265 | 0.472 0:58 | 0.589 = 01319 | 0.450 0.53 | 0583 6 =o. 0.3 2 0:48 | 0.198 98 =o 0/234 98 0:43 | 01453 1 7 07196 | 0.365 0:38 | 9.410 8 i o.161 | 0.333 0:33 | 0.268 | 0.838 5 0.198. | 0.300 0:28 | 07326 | 0'8i6 i 0:101 | 0.263 0°33 | 0.28 | 0.793 7 o.orz | 0.295 018 | 0/287 9 at 01056 | 0-189 0:13 | 0209 | 0745 i 0.039 | 0-151 0:08 | 0.172 | 0.720 7 0:02 | 0.113 0.03 | 01137 } 0-695 7 0.017 | 0.074 ~0102 | 07103 | 0.670 = O-Oll | 0.086 =0:07 | 0/070 | 0644 = 0,009 | 0 0112 | 0.038 | 0-618 = 0 0 17 | 01008 | 0.592‘Theory of Free Turbulence for a Submerged Jet 67 Table 2.2, Auxiliary Functions of the Boundary Layer of the Jet ’ 0.990 0:64 0.922 0.869 0,806 0.738 0.666 0.59 0,520 0.449 0.382 0.320 0.264 ong ori71 SSSRERRsVSseszee bbbbbbccossssses from which, according to the Prandtl-Tollmien theory, the velocity and the temperature difference fields in a two-dimensional boundary layer of a submerged jet must be similar: 0@) =F), (2.66) that is, at points with equal values of the dimensionless ordinate e=ylax the dimensionless magnitudes of velocity and temperature difference are identical: “ ar ie Oren (2.67) The diffusion equation (Eq, 2,43) is identical to the heat transport equation (Eq, 2,32), We shall introduce the concentration difference into Eq, 2.43: Ax a (2.68) where x, is the concentration of a given constituent (admixture) outside of the jet, and we find on the basis of existing experimental Gata, given in Chapter 1, that the edges of the diffusion boundary layer, just like the thermal boundary layer, coincide with the edges of the dynamical layer, and the dimensionless concentration difference Profile is universal and is expressed by the relation bn mo o(@). (2.69) Comparing Eq. 2,69 with Eqs. 2.49, 2,51, and 2.43, as should be expected, we reach the conclusion that according to Prandtl’s theory,68 Turbulent Jets of Incompressible Fluid the admixture concentration profiles are similar to temperature and velocity profiles *#.T.+t, (2.70) This conclusion is not confirmed by the experimental data, from which it follows that the concentration and temperature fields are similar, but differ from the velocity field (see Fig. 1,22), 2.5. Tollmien’s Plane Turbulent Source. It was shown in Sections 1.310 1.5 that in the downstream (main) region of the turbulent submerged jet the velocity profiles at various lateral cross sections are similar and contours of equal dimensionless velocity are straight lines; if we extend these straight lines beyond the limit of the downstream region to their points of intersection—the pole of the downstream region—and place the origin of coordinates at this point, itturns out that in the down- stream region the jet behaves as if it emanates from an infinitely narrow cross section, coinciding with the pole, Using this, one may plot the flow over the entire downstream region, The first part of the source, which is situated between the pole and the end of the transition region, must be discarded and replaced by the initial and transition regions of the jet. Fig. 2.4, Plane turbulent source. Tollmien investigated the velocity distribution in the free flow which is formed during efflux through a narrow rectangular slot (Fig. 2.4), The x-axis coincides with the axis of symmetry, the y-axis is perpendicular to the slot; the coordinate origin is super- posed on the origin of the jet. It was established in Chapter 1 that the velocity at the axis of the downstream region of a parallel jet is inversely proportional to the square root of the distance Uq = = (2.71) VeTheory of Free Turbulence for a Submerged Jet 69 Consequently, the velocity at an arbitrary point ofthe jet is 4— f (7), where y=. (2.72) Ve a We introduce the stream function y, Its partial derivatives equal a the longitudinal and transverse velocity components:u: whence p= JudyanVas fenan. Adopting the notation Fea) = Sf) da, we obtain new expressions for the stream function and velocity components yan xFO, (2.73) u= =F’), (2,74) “uo Leg i [prop Fea]. (2.75) A study of the plane turbulent source must result in a determina- tion of the function F(x). The solution of this problem is made possible by the use of the momentum theorem, We shall isolate a control surface in Fig, 2.4 for this purpose and consider the momentum balance inside it, The rate of change of momentum crossing the lower part of this surface is puuvdx*, Across sides of the control surface the momentum is observed to change by the amount —afeuay. 5 The turbulent shearing stress is sen (2 neta (the minus sign is taken because zo as is apparent from Fig. iy 2.4), After elementary manipulations, we obtain the momentum equation in the following form: (2.76) *The length of the surface along a line perpendicular to the Plane of the diagram is assumed to equal unity,70 Turbulent Jets of Incompressible Fluid Using Eqs. 2,74 and 2,75 to transform Eq, 2.76, we obtain 20° (F" J] = FOF’ (2,77) The experimental constant c occurs in Eg. 2,77, This indicates that in the specified coordinate system (x, 7) the desired functions depend on the structure of the jet, We can exclude the influence of the structure of the jet om the function Fix) by means of 2 simple change of coordinate, We introduce the related coordinate +, (2.78) where Fa— tre, (2.80) consequently the fundamental differential equation is written as [FF = Fs) F’(@). (2.81) We introduce the additional transformation InF(z), which corresponds to FQ) : Substitution into the original equation gives After the substitution 7=Z, we obtain a first-order differential equation —z-yZ. (2.82) A mimus sign is placed in front of the quantity 1/Z, because the transverse velocity gradient at ail points ofthe cross section of the jet is less than zero except onthe axis and at the external boundary where it is zero, The solution of Eq, 2.82 has the form aaa 7Theory of Free Turbulence for a Submerged Jet 7 By integration this becomes goc—2 fil Z41)—m[V2—-Va4 5] + age), (2.83) Va We shall examine the boundary conditions that this integral must salty, The first condition is that on the axis of the jet where , the transverse component of velocity v is zero, but the dimensionless longitudinal Wee ae (Wu) equais unity, ‘As a consequence, when ¢=0, F@=e=0, F@)= +V Bare tan hence, (2.84) The second boundary condition states that the transverse velocity does not vanish at the external boundary of flow of a jet when ¢ =p, Whereas the longitudinal velocity equals zero, Therefore, when 9p, F@) e#0, FY) whenee Zea (2.85) We shall calculate the integration constant from Eq. 2.84: Equation 2,85 makes it possible to determine the coordinate of the boundary of the jet 4n op = 2.412. 7 3V3 Equation 2,83 is used to find the value of the second auxiliary function Z=z', The values of the first auxiliary function z and the basic function F are determined from the following two equalities: > z= at Zao, (2.86) * z=In{F (9), Where z is the desired function; ¢ is the corresponding value of the argument (of a dimensionless coordinate); and 2) is a known Value of the function at some point with coordinate ?%.72 Turbulent Jets of Incompressible Fluid It was shown above that close to the axis of flow of a jet Z +0, and z+—o (at 9=0: e7=0), This indicates the uselessness of Eqs. 2.83 and 2,86 in the neighborhood ofthe axis of a jet, We seck a new solution, suitable in this region, starting from Eq, 2.82, At 90, when Z- ©, the small magnitude /Z can be neglected in Eq, 2.82, Thus we obtain whence Zarot and z>In 9 +C). ‘The constant of integration can be determined from the consider- ation that the dimensionless velocity on the axis of flow of a jet equals unity, a condition which gives C,=0, Consequently, the first auxiliary function that we introduced to lower the order of the differential equation (Eq, 2.81) tends to In 9, In general, moreover, z can be represented in the form of a series whose first term is Ing, Solving the problem by the method of successive approximation, we find:* 7 0425+ 0.012 +..., (2.87) whence ah 0h gt (2.88) ‘The asymptotic approximations in Eqs, 2,87 and 2.88 enable us to calculate valuesof the functions z’(¢), 2(¢), F(g)=e" and F'(~)=ze? at points with small values of the argument. At other points of the transverse cross section of the plane jet, Eqs. 2.83 and 2.86 should be used, Performing these calculations and using Eqs. 2.79 and 2,80, we obtain me TP oF) — FA). @ Um We shall compose a table and plot a graph of longitudinal and transverse velocities in cross sections of the downstream part of the plane jet (Figs. 2.5 and 2.6). Velocities u and v are given in Table 2.3 and Figs, 2.5 and 2.6, in fractions of the velocity at the jet axis Un. ‘All the resulting functions of the turbulent source of a plane jet are suitable only in the downstream part of the jet when BD Kye *Detailed calculations are given in a paper by the author [1].‘Theory of Free Turbulence for a Submerged Jet cm in a7 ast INN oi Ww i 2 Oo Fig. 2.5. Longitudinal component of velocity in the downstream part of a plane jet (Tollmien’s solution), #&, a2 a as Fig. 2.6. Transverse velocity component in the downstream part of a plane jet (Tollmien’s solution), Table 2.3 La ar =F'@)| —— | . Ol Tim [OTe ° 1 12 | 0357 | 0 0.598 O14 0.0%9 | 0,990 | 1:3 | 0.300 | —0.056 | 07518 02 0,081 | 0:970 | 114 | 0-249 | —0'099 | 0499 0.3 07120 | 01947 | 1's | 0.200 | —0.160} 0.447 04 0.151 | 01918 | 1.6 | 0.165 | —0.212| 07405 25 ol165 | 0.885 | 1:7 | 0.125 | —0.250| 0.354 o6 0.168 | o'si9 | 18 | 0.095 |—0:318 | 0.208 0.7 0.166 | 0.812 | 1.9 | 0.057 | 0.356) 0.259 08 O.i5I | 0.777 | 2:0 | 0,016 | —0.402) 0.214 a9 01120 | 0:73 | 2:1 | 0.030 | 0440} 0:170 10 0.091 | 0.689 4 2:2 | 0.020 | —o-469 | 01141 ia 0:09 | o:641 | 2:3 | 0.009 | —0.490 | 0.095 24 | 0 =01498 | 0 7314 Turbulent Jets of Incompressible Fluid It is apparent from Fig, 2.7 that the Tollmien velocity profile in the downstream part of the plane jet is verified by Férthmenn’s experimental data, transferred from Fig, 1.6, The axial velocity of a plane jet itself is a function of distance from the pole of the jet (see Eq. 2.79): a, wx const VE We shall write this expression in a somewhat different form, for which we use the conservation of momentum ina free plane jet: ep of wtdy = pugs, = const, a where Yr» is the ordinate of the jet boundary, bois one half the thickness of the initial cross section of the jet; and u isthe velocity of efflux, y Un) a + 275) a50) O25, “GS-20 MS DEE BD ashe Fig. 2.7. Comparison of the theoretical velocity profile in the downstream part of the plane jet with Férthmann’s experimental data [6]. After a change of variable (=y/ax replaces y), we obtain Uy 7 V 7 fe (Fae We shall calculate the integral according to Table 2.3, intro- ducing into the preceding formula 28 { F’?de = 0,685. @‘Theory of Free Turbulence for a Submerged Jet 75 Replacing the integral by its numerical value, we obtain the final formula for the axial component of velocity in a parallel jet: Bes L Up b From this we find the distance to the cross section at which n= Uo (the dimensionless distance from the pole of the jet): axq/by = 1.44. Comparison of the curve of velocity decrease along the axis of the plane jet with Férthmann’s experimental [6] at a= 0.11, Proskura’s (31) at.a=0.12 and Turkus’s [32] at a=0.09 is given in Fig, 2.8, where the distance s from the nozzle exit is plotted along the abscissa instead of the distance x from the pole, Good agreement between experimental and theoretical values of the axial velocity is achieved under the condition that the pole of the downstream region is inside the nozzle at a distance # | I I 10) © Férthmann's expts. [6] —| NR 6 Proskura's expts, [31] 28}-—+ @) Turkus's expts, [32] t tm! —Theory " I oat el oy} el a2 129 eae @ Fig, 2,8, Variation in velocity along axis ofthe plane jet, 2 041. We note that a determination of the constant a can be effected not only by using the empirical curve of velocity decrease along the axis of flow of a jet but also by the empirical dependence of the Jet thickness on the abscissa; in the latter case other values for a and s) are obtained. For instance, it follows from the experiments of Forthmann, given in Fig, 1.13, that the half-velocity line in the downstream part of the plane jet has a dimensionless ordinate 4Yelt= 0.097, whereas the value of the ordinate ¢,= y, /ax=0.96 Corresponds to this line on the Tollmien velocity profile, Hence, %~0.1, i¢., 10% less than that obtained from the axial velocity curve, Moreover, the line y. in the Férthmann experiments76 Turbulent Jets of Incompressible Fluid crosses the jet axis at the point with abscissa as1/bo=—0,), i.e, the position of the pole determined from a geometrical parameter, differs from that found from the velocity curve, Thus, in a calcula. tion of axial velocity according to Tollmien’s theory, strictly speaking, we should take one value of the experimental constant (a) and of the polar distance (s:), and ina calculation of thickness of the jet—the other values (a, 501); however, the discrepancies between theory and experiment in a calculation of axial velocity and thick ness from one pair of values or the other (a, s oF a, S:) are relatively small, 2.6. Tollmien's Axially Symmetric Turbulent Source Applying the same’ considerations as in the examination of the downstream part of a two-dimensional jet, it is possible to analyze the flow in the downstream part of a turbulent submerged axially symmetric jet by examining the flow from a turbulent point source (Fig. 2.9). Fig. 2.9, Axially symmetric turbulent source, Positioning the coordinate origin at the source pole, we note that as established in Chapter 1, the dimensionless flow velocity remains constant along any radial line, drawn from the pole and lying within the limits of the downstream region of the jet sob oe Just as for the plane problem, we shall disregard this result in the upstream part of flow, which is located between the pole and the end of the transition region, and we shall replace it by the proper flow in the initial and transition regions of the jet, : On the basis of the property expressed by Eq, 2,89 and the consideration that the velocity on the jet axis can depend only on the abscissa, we obtain in general form the law of velocity of the down- stream part of a jet with a round cross section a, = Unf (1). (2.90)‘Theory of Free Turbulence for a Submerged Jet 17 We give below an application of the Tollmien’s theory of flow from a turbulent point source, and we deduce a universal relation for the velocity profile in the downstream part of a round jet In the first chapter a law was established for the variation of axial velocity along the downstream part of a round jet (2.91) Equation 2,91 makes it possible to obtain a formula for velocity aa tef (= =f). (2,92) ‘The components of velocity in an axially symmetric flow can be expressed by means of a stream function in the following manner: ay 1 Ho pe Khe iy y oe” Hence, o= faydy =me[f (yada. We introduce the notation Fo) =Jfovadn and we rewrite the expressions for the stream function, longitudinal velocity, and transverse velocity: p= msF (x), (2.93) (2.94) mip 2 om 2 [F + Feo). (2.95) The problem under investigation consists of a search fora function F(4) and its derivatives, The solution of this problem can be obtained with the aid of the momentum theorem, We define in Fig, 2,9 a control surface symmetrical with respect to the axis, and we set up the momentum balance. The momentum transported each second across the inner part of this surface of area 2xydx equals Qxpuoydr. The momentum inside the surface changes at the rate —2npd fury dy, 5