Preface This is the second edition of the SAE Aerospace Applied Thermodynamics Manual. It is a con- ‘tinuation and an up-dating of the SAE Acrospace Applied Thermodynamics Manual originally pub- lished and initially revised by SAE Technical Committee AC-9 under the chairmanship of W. W. Reaser of McDonnell Douglas Corp. and B. L. Messinger of Lockheed Aircraft Corp. This second edition has been prepared as a continuing effort to provide engineers a single up-to-date source of reference material for use in the field of thermodynamics and those fields commonly associated with itin the aerospace industries. Those related subjects include aerodynamics, fluid dynamics, heat transfer, and properties of materials. Also included are procedures and equations commonly used in the application of these technologies along with design information on the aerospace systems in which they are commonly applied. ‘The Manual is organized into four major sections: 1. Engineering Fundamentals 2, Thermo-Physical Properties 3. Aero Application Engineering 4. Space Application Engineering An attempt has been made to make the manual of practical value to engineers, Each section has bbeen prepared by engineers who are experienced in, and are currently actively engaged in, the field of their assigned topics. An effort has been made to ensure the applicability of the material to both the aircraft and space fields, The responsibilty for the SAE Aerospace Applied Thermodynamics Manual is vested in Commit- tec AC-9, Aircraft Environmental Systems, of the Aerospace Equipment Division of the Society of Automotive Engineers. Within Committee AC-9, the preparation of this second edition has been the assignment of Subcommittee AC-9B, Aerospace Applied Thermodynamics Manual. The actual work has been accomplished by many engineers from throughout the aerospace industry. Primary responsibility for the parts of this Manual were as follows: Part 1A, ‘The Boeing Co, Part 1B ‘The Bocing Co, Part IC United Aircraft Corp. Part ID United Aireraft Corp. Part 1E United Aireraft Corp. Part 2 RL. Berner Martin-Marietta Corp. Part 3A CC, Van Valkenburgh Grumman Aerospace Corp. Part 3B C.C.Van Valkenburgh Grumman Aerospace Corp. Part 3C 1.8. Perlee ‘MeDonnell Dougias Corp Part 3D J.S.Perlee ‘McDonnell Douglas Corp. Part 3E J.S. Perlee ‘McDonnell Douglas Corp. Part 3F —_J.S. Perlee MeDonnell Douglas Corp. D.T. Bowden General Dynamics Corp. Part 3G C.C.Van Valkenburgh Grumman Aerospace Corp. Part 3H. W. Adams Hamilton Standard Div., United Aireraft Corp. Part 31 DJ. Melchior ‘The Boeing Co. Part 3) RC. Kinsell AiResearch Mfg, Co,, Div, The Garrett Corp. Part 3K E,W. Adams ‘Hamilton Standard Div., United Aireraft Corp. Part 44 FL. Guard Lockheed. California Co. J.B. Werner Lockheed-California Co. Part 4B P.B.Cline General Electric Co. Part 4C —L.H. Hemmerdinger Grumman Aerospace Corp. Part 4D E.A. Schumacher Martin-Marietta Corp, Part 4E BH. Rowlette The Garrett Corp. F. Morris Hamilton Standard Div., United Aircraft Corp. R. Bambeneck General American Transportation Co.1 engineering fundamentalsIncompressible Fluid Flow 1. Introduction 1.1 ,score—The fuid flow treated in this Section is isothermal, subsonic, and incompressible. The effects of hea addition, work on the fluid, and va tion in sonic velocity is neglected, The effects of changes in elevation are also neglected. An incompressible fluid is fone in which a change in pressure causes no resulting change in fluid density. The assumption that liquids are incompressible introduces no ap- Dreciable error in calculations but the ‘assumption that a gas is incompressible introduces an error whose magnitude is dependent on the fluid velocity and on the loss coeflicient of the particular duct section or piece of equipment Fig. 1A-1 shows the error in pressure drop resulting from assuming that air is incompressible. With reasonably small loss coeffi- cients and the accuracy that is usually Fequired in most calculations, eom- pressible fuids may be treated’ as in compressible for velocities less than Mach 0.2. At higher velocities and/or large loss coefficients (K, and/or 4fL/D), compressible flow analysis accepted standard symbols which are defined in the master nomenclature list, © = Chord, fin D,_ = Equivalent diameter, fin. Friction factor, Fanning G = Weight low per unit, Ibjsee-t? ‘Area per unit time, Ib/see-ft? Pressure loss coeficient = Length, fin. ~ Static pressure, tb/f, in, He Total pressure, lb/ft, in, He = Dynamie pressure (' p¥*), 16/iC, io. Hy = Volume rate flow, f?/sec, 1 /min ‘Curvilinear distance, ft in, = Velocity, ft/se, ft/min = Weight flow rate, Ib/see = Compressibility factor mB Saat Angle of attack, deg Surface roughness, average, = Proportionality factor = Efficiency Mass density, Ib-sec?/tt* ‘= Function "RNENS Subseripes Average should be used © = Compresibe 12. NONENCLATURE—Thislist sone D> Dra tsns those symbols which may, f0'8 ef Bffecive degre, be used uniguely in this sec: > Frewtteam, ouside tion, ‘There are ditional commonly p= Pressure “Ey I Hs wo3 woz s[es eS ‘ a| 4 é | e* [Tt =| oy foe ——T I] war °o 2 4 6 0 1 4 LOSS COEFFICIENT x, Of 444.0 Fig. 14:1 —Incompressble Row, presture loss error Part 1A 2. Basic Fluid Mechanics 2.1 continuity EQUATION — The continuity equation is generally exe pressed in the form AV = constant and reduces to AV = constant for ideal incompressible low. 22 MoMENTuM EQUATION — The form of the momentum equation most commonly used in uid flow is, (a1) dmv) zr 4a) 1.2) where the summation of the forces is equal 10 the pressure, wall, and frie tional forces on the uid, 2.3. ENERGY QUATION—The en- ergy equation for incompresible fri tionless flow can be expressed in a form called Bernoulli's equation: P+ 4p¥? = constant —(1A-3) nore te St tm eh ate pe tunic pet fet The sam at theme cs al ra, To the pigsal mening toe above prety an ods To ee (2) Ste Prere—This is the pronr nnd pa MRevime vey the ad Paste, nued bya pe Sure st normal. he scm, ("Dynamic Peet is he Dora comnly ose (O' Toul Pe Tis the pesos moving Bld i ery roo oa i apo thon ote atc addy pre, nos equation, Ey a he ta freee by ity tse not een cone ny al the re ced ye staring ul seams, weet a fo beat dean en Be oe4 ditions of laminar flow, this shearing stress is given by Newton's equation zw She a4) where the proportionality factor is the property of the fluid called absolute viscosity and dV/dx is the velocity ‘gradient normal to the flow. The shear- ing stress that exists in a fluid is there- fore proportional to the viscosity and the velocity gradient that exists in the fluid normal to the direction of its flow. In most liquids arid gases the absolute viscosity depends primarily fon the temperature of the fluid. There is only a slight dependency on pressure except in the vicinity of the critical point or in dissociated gases, 2.5 BOUNDARY LAYER FLOW— When. fa real fiuid is in motion along a wall, the fluid particles adjacent to the wall are retarded by frictional forces. These retarding forces cause a velocity gra cent near the wall, thereby zone of flow called the layer.” ‘The thickness of the boundary layer is usually defined by the two extremes of velocity: zero at the wall and 99° (or 99.5%) of local free stream velocity the outer edge. Within this layer the distribution of velocity is dependent on whether the flow is laminar or turbue lent, and thereby it determines. the velocity gradient at the wall and the shear force or “drag” on the wall. In constant area duct this drag force per unit surface area per unit length is balanced and thus is equal to a pressure loss or pressure drop per unit cross- sectional area per unit length. 2.8.1 Laminar-Turbulent—Inlami- nar low the streamlines are orderly and ‘undisturbed, all shear being developed purely by molecular interaction. In turbulent flow, random eddies exist, causing momentum transfer across streamlines due to the irregular move- ment of finite masses of fluid. 2.5.2 Reynolds Number—In order to predict whether the boundary layer flow will be laminar or turbulent, the ratio of inertia forces to viscous forces must be considered. This ratio is, by definition, the Reynolds number (Nae) and is equal to Vxp/u or Vd-p/u. Flow ‘conditions in a smooth closed channel will usually be laminar if the Reynolds number (based on hydraulic diameter) is less than 2000 and turbulent if above 4000. A “transition” type of flow exists SECTION | between these values. For the case of fa smooth flat plate submerged in an infinite flow field, the transition occurs at approximately Nge = 500,000 to 1,000,000 where the distance along the surface from the leading edge is the characteristic dimension used in the Reynolds number, The growth in thickness of the boundary layer is illustrated in Fig. 14-2. Over a fiat plate (negligible pres- sure gradient), the laminar portion in- creates in thickness as the square root — ENGINEERING FUNDAMENTALS ‘of the distance, and the turbulent region increases as the 0.8 power of the dis- tance, Fig. 1A-2 also indicates the transition from laminar to turbulent flow as well as the limitation (beyond the entry region) of the growth of the boundary layer when the flow is con- fined to a duet or channel 28.3. Friction Drag and Duct Pres- sure Loss Concepts—Ea, 14-4 defined the shear stress that acts on the internal surface of a duct when a known velocity gradient exists at the wall, Down- Boundary Layer 33 (a) FLAT PLATE (o) CIRCULAR PIPE with LAMINAR FLOW Lerner Boundory Tren lever unotectes|siton Dy'epposte wal Bowncory yer —D "Farben? boundary loyer anatectes By ‘opposite wall Laminar Sublover ‘Boarder yer growth eases, Drag or pressure ‘rope constont per tint lenge (e) CIRCULAR PIPE WITH TURBULENT FLOW Fig. 1A:2—Boundory layers and velocity distribution TURBULENT FRICTION FACTOR: 2 4 6 810° REYNOLDS NUMBER, Neg Fig. 14-3—Pipe velocity and friction factorsINCOMPRESSIBLE FLUID FLOW os & o oe aren a1 Brees Pe 888 = oe & aoe] g 2 ox 8 & 2005] a ¥ Soosf- 5 coo 001 20004 Sono conve 001 a pe 234 681 234 6810 20 3040608000 ver DAMETER Fi. VAd—Reltive ough, rel dc Table 1441 Tableor Conon org (ny) lo Urisate nd bjt? b/min-fr? Ib/sec? Ib/sec-in.? {b/min-in.? pst 8.35 x 10" 2.32 x 10° 64.4 Bx 107% 11.20 psi 1.202 x 10"" 3.34 x 10” 9.28 x 10? 0.447 1.61 x 10% in. HpO 4.34 x 10" 1.205 x 10 3.35 x 107 1.615 x 10°? 58.14 inHg 5.90 x 10" 1.64 x 107 4.56 x 10° 0.2196 791 x 107 Toble 142 Toblefor poof Ar Uni of in") jin. of Hg abs Be Be in jabs Lea Ibyin.? abs Ib/ft? abs neat aCe 2 ’ * ae 7 wa? ansF aro tax wt 9x 0h un Tt T T T 5 stream of the entrance region of a duct, the velocity gradient is independent of duct length, as indicated in the lower part of Fig. 1A-2. The shear stress ‘causes a force along the wall propor= al to the area in contact with the fluid, and this force is referred to as a drag or friction force. In classical aerodynamic theory, the rag force acting on a surface A is ex- pressed as Fo = Coga For the special case of a round duct, the surface area is A= xDL ‘The drag or friction force must be over= ‘come by a pressure force corresponding, to the duct pressure drop, so that Fo= (Pa - ra)(t22) ae, (2) (lass) Equating: srs co 24 = oe ae se(EJo= Bee ane In duct flow, Cp is replaced by the sym- bol of the Fanning friction factor, j. As mentioned above, the shear stress at the wall becomes a constant value for pipe flowin equilibrium, and therefore the friction factor is also @ constant. Fig. 14.3 shows the friction factor and the ratio of average to maximum velocity in pipe flow for these equilibrium con- ditions Since the roughness ofa pipe changes the velocity profile at the wall, it also ‘changes the shear stress and the frition factor. (See Fig, 1A-4,) 2.6 TOTAL PRESSURE Loss DUE TO SEPARATION AND VELOCITY PROFILE CHANGE—In the discussion above it was shown how the viscosity caused ve- locity gradient at the fluid boundaries, With its associated friction drag and pressure loss. It may be logically rea soned from this that any change in the fluid velocity profile will cause a loss of energy and result in a corresponding pressure loss, Any change in direction, cross-sectional area, or shape will cause loss in total pressure, In dealing with losses ofthis type, the following equa- tion is used: APE Kiel? = Kg at)6 where K; is the loss coefficient for the change in direction, shape, or other type of loss. It should be emphasized that the loss coefficients are determined for one particular arrangement of duct- ing leading into and out of the device ‘whose loss coefficient is being meas. ured. Therefore, when combinations of devices are used, the loss coefficient of the system will not necessarily be the sum of the individual loss coefficients, but may be either larger or smaller By careful selection of the location of bends, diffusers, and contractions to achieve the minimum change in velocity profile, the pressure losscan be reduced ta minimum, Examples of this would be the use of one 180 deg bend instead of two 90 deg bends separated by a straight section of pipe. Do not diffuse around a bend and avoid large ares ‘changes at duct junctions. In summary, the total losses ina system can be expressed by the equation Lee pag Alecks ys ar = 2K, tev? + 2 Et ov a? Dea? Gas) ‘where the dynamic or velocity pressure C/A) ¥? is evaluated at the reference SECTION 1ENGINEERING FUNDAMENTALS Table 14-3. Coordinates for Reynolds Nomber Nomograph for Gases (Fig. VA6).* Gos x ‘cate Acid 77 ‘Acetone 39 Acetylene 98 ry 0 Ammonia 34 ‘Argon 105 Benzene a5 Bromine 39 Butene 92 Butylene 39 Corbon Diowide 95 Corbon Disulfide 8.0 Carbon Monoxide 11.0 Chlorine v0 Chloroform as Cyanogen 92 Cyclohexone 92 Ethone 91 Ethyl Acetate as Ethyl Alcohol 92 Ethyl Chloride 85 thy Ether 29 Ethylene 95 73 108 ma Freon-21 108 Freon-22 10. y 143 130 us 200 160 24 132 192 137 130 187 160 200 184 157 152 120 145 132 142 156 130 151. 238 151 160 153. 170 Freon-113 Helio Hexone Hydrogen SHAN Hydrogen Bromide ‘Hydrogen Chloride Hydrogen Cyanide Hydrogen lodide Hydrogen Sulfide \odine Mercory Methone ‘ethyl Alcohol a | Niner Chri ee | Oxygen | Pentane ee | pop | Propylene rae ee ie Rel. 234 68m 2 34 REYNOLOS NUMBER, Naw em 2 34 680" Fig. 14-5—Friction factor versus Reynolds number > GasINCOMPRESSIBLE FLUID FLOW area of the loss cosfcient andor the friction factor f. For convenience, the dynamic presse (/)¢ can be evalu= ated from Tables 1Av1 and 1A-2 and the equation Lin S oe pet ae uaa) where: © = Chord A sample usage of Tables 1A-1 and 1A2is B= 2702/7 Ib/? (rom Table 14-2) Teble 1A-4 Coordinates for Reynolds Number Nomagraph for Liquids (Fig. VA)" Liquid ‘Ly geld xy ‘Acetoldehyde 152 48 | Froona2 72 47 ‘cate Acid 100% 121 142 | Freon-113 25 114 ‘catic Acid 70% 95 170 | Glycerol 100% 20 300 ‘etic Anhydride 127 128 | Glycerol 50% 69 196 ‘Acetone 100% 1457.2 | Heptone M1 8 Acetone 35% 79 150 | Hexone 47 70 Ally! Alcohol 10.2 143 | Hydrochloric Acid 31.5% 13.0 166 ‘Ammonia 100% 126 20 | — bobuty Alcohol 71 180 ‘Ammonia 269% 10.1 139 | bsobutyrie Acid 122 14.4 ‘Amyl Acstate 11.8 125 | Isopropyl Alcohol 82 160 Allyl Alcohol 75 184 | Kerorene 102 169 81 187 | tinseed Oi, Row 75 272 Anisole 123 135 | Mercury 1B4 16.4 ‘Avsenie Tichloride 139 145 | Methanol 100% 105 Benzene 125 10.9 | Methanol 909% ns Brine, CaCl, 259% 66 159 | Methanol 40% 155 Brine, NoCl, 259% 102 16.6 | Methyl Acctate 82 Bromine 142 13.2 | Methyl Chloride 38 Bromotoluene 200 15.9 | Methyl Ethyl Ketone Be Butyl Acetore 123 110 | Naphthalene 181 Butyl Alcohol 86 17.2 | Nitric Acid 959% 138 Boryre Acid 121-153 | Nii Acid 6096 79 Corbon Dioxide 116 3 | Nimobenzene 162 Corbon Disulfide 161 7:5 | Nitotolvene 70 Carbon Tetrachloride 127 131 | Octane 100 Chlorobenzene 123 12.4 | Octyt Alcohol aa Chloroform 144 102 | Pentachloroethane 73 CChlorosulfonic Acid 112°18.1 | Pentane 52. Chlorotolvene, ortho- 130 133 | Phenol 202) Chlororoluene, meto- 133 125 | Phosphorus Tibromide 167 CChlorotolvene, para- Phosphorus Trichloride 109 Cees, meto- Propionic Acid 138 Cydohexne! Propyl Alcohol 165 Dibromoethane 8 | Prepyl Bromide 98 Dichloroethane 132 122 | Propyl Chloride 73 Dichloromethane 146 89 | Propyllodide ns Diethyl Oxolote 110 16.4 | Sodiom 9 Dimethyl Oxclate 123 158 | Sodium Hydroxide 50% 3.2 25.8 Diphenyt 120 183 | — Stonnic Chloride 135 128 Dipropyl Oxclate 103 177 | SulurDioxide 152 71 thy! Acotate 137 91 | Sulfur Acid 11096 72 4 Ethyl Alcohol 100% 105 13.8 | Sulfuric Acid 9896 70 248 Ethyl Alcohol 9596 98 143 | Sulfuric Acid 603% yo2 213 Ethyl Alcohol 40% 65 168 | Sulfur Chloride 162 124 Ethy! Benzene 132 115 | Tetrachloroethone 19 157 Ethy! Bromide 145 B81 | Totrachloroothylene 142 127 Ethy| Chloride 148 60 | TianiumTetrachloride 14d 12.3 Ethy Ether 145 53 | Toluene 137 104 Ethyl Formate 142 8.4 | Tchlorosthylene 143 105 Ethyl lodin 147 103 | Turpentine ns 149 Ethylene Glycol 60 236 | Vinyl Acctote 40 88 Formic Acid 107 158 | Woter 102 13.0 Freon-11 144 9.0 | Xylene, ortho- 135 121 Feeon-12 168 54 | Xylene, meto- Freon-21 13775 Reh TT G = w/A_ tofewinjin? € = $8.14 (fromTable 14-1) ‘ Sin. H,0 270P, 8.14) where w isin Ib/min, P,in psia, Tin'R, ‘and 4 inin.* Specific Flow Loss Data 3.1 FLOW Losses, stearcur pucTs —As noted in Par. 2.5.3, the pressure lossin straight ducts can be determined as follows: 4 (1 lye . an- 2 (& y (Wa-10) For most aircraft ducting, the duets ‘can be treated as smooth tubes, but 10 ‘obtain a physical feel for the roughness ‘where actual data are lacking, Fig. 1A-4 is presented for a number of common materials. The relative roughness pre- sented in this figure is the mean rough- ness projection (¢) and is the arithmetic average surface deviation from the mean line of the surface, in inches, Friction factors are shown in Fig, 1A-5 for long pipe lengths where entrance ef- fects are small; nomographs for de- termining Reynolds numbers are shown in Figs. 14-6 and 1A-T together with Tables 1A-3 and 1A-4 for gases tnd liquids In computing the pressure drop from Eq. 1A-10 and the Reynolds number for ducts of noncircular section, an equivalent diameter of the section must bbe used and is defined as De w 4% (cross-sectional area) ‘wetted perimeter (aa) For some of the more commor. shapes the equivalent diameters are Square duct: D, = length of side 2ab Rectangular duct: D, = 220, (a and b are sides) Elliptical duct: 1 Peo 208 ae (a and 6 are major axes) Annular duct: Dz = Dy — Dy + Where D, and D, are the outer and inner duct diameters, respectively.SECTION 1—ENGINEERING FUNDAMENTALS, 7 y= = ae : : S| gay gxto 5 Fig. 14-6—Reynolds number in gasespinby ut sequin spjoukoy—L-yL “Oy i 5: ' we " " * t Ht aw + 4, ‘ ea rt ae , =! H ai ‘o0'a ' ret : u : i aA +f je as f wT 7 fa f i a f of In) on tose x , La , : i age Ute Fg EE He a on al i Ft : ae) 7 M : i ; i e joe ms e om om . = : y * ne ! mm : i i mth E vw 00" a iz 2 a” ite md z ‘sh 08 : Ey w ae 2 ! re 4 4 : : ean ieee Hl "Ne He WW/) :NENIW/AT—ALISODSIA z a ‘9 z 3 40 ‘The following formulae are useful in determining the friction factors of smooth pipes over the ranges listed and approximate the data on Fig. LA-4 i Nae < 2100 (1A-12) 0.0791 Nae = a Bott Na = 200010 10,000 fe ae Nne = 10,000 to 200,000 aaa) where Ng, has fairly good accuracy to Mee = 5x 108 and (a-14) Nie = 3000 to 3,000,000 1 = 40 logie(Nee V7) - 0.40 Sie a o Gass Eq. 1A fits the data over the entire range of experiments and is recom- mended for extrapolating 10 higher Reynolds numbers. For air, the factor 4f versus w/D is presented in Fig, 14-8 to simplify the calculation of AP,, using Eq. 1A-10. 3.2 ELBOW FLOW Losses —The flow Joss ina bend or an elbow is arc= (4y4 + chim) ov* (1A-16) where: C= Correction factor for bends other than 90 deg L = Length along centerline of duct ‘The loss coefficients forall bends and. ‘elbows given in the succeeding sections sive total pressure loss due to turning ‘only. The wall or straight duct friction loss must be added to this loss to de- termine the overall loss of the elbow. Fig, LA illustrates the combined pres- sure loss, turning plus duct friction, in ‘circular duet for various radius ratios. tis seen from this figure that the mini- ‘mum loss occurs at an r/.D ratio of 3. While this particular figure is for a Reynolds number equal to 10°, the principle holds true for other Reynolds numbers. Therefore, considering both pressure drop and fabrication, a bend SECTION 1-ENGINEERING FUNDAMENTALS with a radius ratio between 2 and 3 should be used wherever possible. ‘Note: The loss coefficients for bends and elbows in Figs. 1A-10 to 1A-21, unless otherwise specified, apply 10 elbows followed by a length of duct ‘equal to atleast four duct diameters. If no duct follows the elbow, the loss Lersec-in is to ined, Fig, 14-8 — Friction foctor versus weight flow per unit diameter, : fie In smooth ound ducte 08 coefficient should be corrected as in Par.3.2.1. 3.2.1. Effect of Discharge Duct When the discharge duct at the exit of an elbow is removed, the flow pattern through the elbow is changed and this ‘changes the loss coefficient. For bend angles less than 90 deg, the flow losses "00 015 020025 030 038 040 4 RADIUS, RATIO, +/D Fig. TAP—Bend LossINCOMPRESSIBLE FLUID FLOW 3 s 3 2 10 3 i 1s 10 20. | 30 96) 0 3 ) £ [a See ae a 0 REYNOLOS NUMBER, Ning Fig. 1A:10—Loss coefficiet versus Reynolds number for 90 deg radius bends, circular ducts. (Data for Ne, > 10° from Ref. 1, for Ne, < 10° from Ref. 28) 3 To but & sv 2 5 40 6 ASPECT RETO, a TA12—Loss coeficionts for 90 deg radius bends, elliptical duels, Ney = 150,000. (From Ref. 1, Fig. 3a) Pe 3 E 2 fs 2 g 20] Q 2 ” ie i io 33 8 £ 08] a oa a Hiro ont « oa » rest Se ASPECT RATIO, 204060 60 100 120 140 160 180 BENO ANGLE, # OES Fig. 14-18—Loss coefficients for radius bends, Fig. 1A-11—Correction foctor for bends ether then 90 deg for elliptical ducts, Ng, "= 300,000. (From Ref. ducts. (From Ref. 12, Fig. 16 1, Fig. 3b)— cers. Oss 7T est ASPECT RATIO, 6/y Fig. 1A-14—Lose coeficiente for 90 deg bonds, elliptical ducts, Nee = 600,000. (From Ref. 1, Fig 3c) 0 © 40 60 0 0 BEND ANGLE, @ Oey Fig. 1A.15—Loss coeficient for miter bends, circular ducts, Nm <1000. (From Ref. 12) 115 TO ouCT & SECTION 1-ENGINEERING FUNDAMENTALS K190 CORRECTIONS FACTOR, C= Ky /Kyg0) ys 075, ean REYNOLDS NUMBER, Nae Fig, 1A-16—Lots cooficents versus Reynolds number for 90 deg radius elbows, square duct, (From Ref. 1, Fig. 2; Ref.9) i. ol 0-20 40 60 80 100 120 140 160 160 BEND ANGLE, # Deg Fig. 1A:17—Correction factor for bends ather thon 90 deg, rec: tongular ond square duct (From Ret. 5)INCOMPRESSIBLE FLUID FLOW B K190 ST eal 2 3 6 5678910 ASPECT RATIO, 9/5 Fig. 1A-18—Loss coeficient for 90 deg bends, rectangular (From Ref. 1, Fig. 20) Fig. 1420—Loss coefficients for 90 deg recive. bend, rectonguler ducts, Nu. se TasT 23 8 887830 £00,000. (From Ret 1, i. 2e) ASPECT RaTio, o/b ie ts z 3 a 10 g apt u = ze 7 s 5 4 3 2 2 2 eG ' STesT ZS eS eTet No ASPECT RATIO, De fe BEND ANGLE, @ Deg Fig. 1A:21—Loss cooficient for square mitered Fig. 1A:19—Loss confficiets for 90 deg radius bends, rectongular ducts, 300,000. (From Ref. 1 Fig. 26) ‘elbow. (From Ref. 13) Nie30| ae 5\h 20 == SECTION | ENGINEERING FUNDAMENTALS. 40 ASPECT RATIO“ are RADIUS RATIO, “7D Fig. 1A-22-—ffect of exit duct on 90 deg elbow. (Dote from Ref. 5) CORRECTION FACTOR, C: Fig. 14-23—Cortection fae- tor for bends other then 90 deg for bends without ‘exit duets. (Date from Refs, 5 and 12) Fig. 1A-24—Loss coefficient for ind U-bends, circular ducts, (Reduced from Ref. 3, Figs. 27-32; curcutar is] bucr 2040 G80 1G) 120 180 160 160 BEND ANGLE, 9 DES are increased by removing the exit duct; for angles greater than 90 deg, the flow losses are decreased by removing the exit duct. This can be explained by the fact that for angles greater than 90 des, the removal ofthe exit duct reduces the ret turning angle, and this is the pre- dominant loss effect. The effect of the cexit duct is plotted in Fig. 1A-22 for 90 deg bends and presents the ratio of the loss factor with and without the exit duct, The loss without the exit duct is for the loss just up to the exit end of the elbow. Therefore, an additional loss for the sudden expansion should be added to the values obtained to deter- mine the total loss of an elbow without aan exit duct (par. 3.3). For angles other than 90 deg, Fig. LA-23 is used to cor- rect the loss factors obtained from Fig, 1A-22 for bend angle. Although the data used to obtain Fig. 14-22 were obtained from rectangular duets, they ccan be applied to elliptical and circular ducts in lieu of specific data for these duets 3.22 Compound Elbows —Figs. 1A- 24 and 1A-25 for circular ducts plot K, versus r/D for compound U bends ‘and 50 deg offset bends, respectively Fig, 1A-26 plots the same values for Z bends. 3.2.3. Improving Bend Efficiency 323.1 TurningVanes—Turning vanes are used in sharp bends to im- prove the exit velocity distribution and reduce the pressure loss. All the turn- ing vanes in a particular elbow are of the same size and shape as opposed to splitter elbows (par. 3.23.2). Eqs. 1A 20 and 1A-21 were obtained from Ref, 1, pages 10 to 13, and the performance data for thin vanes were obtained from Refs. 1, 9, 13, and 15; data for thick vanes are from Ref. 29. The geometry referred to in the following paragraphs is shown in Fig. 14-27 4a, A: (ren) ‘The actual gap, s, between vanes for all types of vanes is determined from the equation avery 1+(@ ~ (A\cos@ 4 s-—"tat) aria o*) (a) Design (U) Thin Vanes, Circular Profile—Due tothe ease of manufacture and installa-INCOMPRESSIBLE FLUID FLOW 15 tion, a thin circular are vane is recom- ‘mended for general use. These vanes are equally spaced within the bend. A. vane is considered thin when its thick: ness is less than ‘Jy the bend radius (7). ‘The vane geometry is determined from the duct width 6, the inside bend radius r,, the bend angle , and the area ratio A,/43. The inside bend radius ‘should be made as large as space will permit 0 Fig. 1A-25—Loss coeficient for 90 deg offset bends, circular ducts, (Reduced from Rel. 3, Figs. 39-44; Ret. 8, Fig. 32) Vane bend radius: r= 1, Chordlength: ¢ = 2rsin etticont => vu [F4]-1 ann For bends of equal inlet and outlet dimensions, this equation reduces to 2 (im 4) =1 (1A-20) rae In Fig. 14-28, nis plowed as a function of r/b for equal inlet and outlet dimen- _ sions. ‘The angle of attack is presented, in Fig. 14.29, 5 20 30 Q) Thin Vanes, Noneireular Profile— iD For use in bends with equal inlet and Fig. 1426—Loss cooficints for Zbends, crelor duct, (Reduced outlet dimensions, a noncircular thin from Ret 3 igs. 39-38,Ref. 6 Fig. 31) vane more efficient than the circular are vane has been developed, However, since the noncireular arc is more dif- cult to fabricate, it should be used only where the decreased pressure drop is, necessary. The coordinates for this type of vane are shown in Fig. 14-30, and the angle of attack is presented in Fig. 1-31. ‘At the upstream end’ of the vane, x= 0and y= 0, For this type of vvane, the shape of the inside and out side corners should be the same shape as the blade instead of having a radius 11. Therefore, a chord length e should be selected as large as the bend inside- ‘corner clearance will permit. The blade and comer blade locations in the duct are determined from c and the angle of attack, « (Fig. 14-31), ‘The number of blades is determined from c, the s/c value from Fig, 1A-31, and the follow- ing equation: ‘SCARF LEADING AND TRAILING EDGES f THUS! WEN POSSIBLE 1 (A21) " efeyeon 7) ~ Fig. 14:27 — Turning vanes geometryre 6 @) Thick Vanes—Two types of thick vanes are presented and will be desig nated as type A and type B in Fig. 1A~ 32. Both types are very similar and vary only in the contour of the leading, edges. ‘Number of vanes for both types of “He 80 60 a2) 40| 10] 08 08 04] 02 ' 2 4 6 810 SECTION | Chord length: Type A has the follow= ing chord length; type B starts with a type A vane and is shortened by virtue of the leading-edge contour change: e e=2rsin (14-23) where: re = Inside chord radius, (b) Pressure Loss—The following 20. 40 606010 200 NUMBER OF VANES, n Fig. 14-28—Number of vanes for thin circular are vanes, € = 2rsin#/2, n= (Sb/e fin 8/2) ~ 1. (From ef. 1) 388 8 ANGLE OF ATTACK, 0 0 20 30 40 50 GTO 80 50100 [BEND ANGLE, 0 Det Fig. 1A-29—Angle of attack for thin circular are. vanes. (Colevloted from Ref. 13) ENGINEERING FUNDAMENTALS pressure loss coefficients are based on the upstream q immediately preceding the vaned bend, The Reynolds number is based on duct velocity and chord length. (1) Thin Vanes, Circular Profile—Figs 1A-33 and 1A-34 plot loss coefficients and gap-to-chord ratio versus K,, respectively, @) Thin Vanes, Noneircular Are—Fig. 1A-35 plots K, versus loss coefficient (@) Thick Vanes—The pressure loss versus Reynolds number for thick vanes is presented in Fig. 1A-36. The pressure drop data were determined for 90 deg bends with a spacing s equal to Lin, and a radius r equal to 0.743 in, Some of the data were determined with fan open duct downstream; therefore the pressure loss data may be a little low under some flow conditions. The recommended number of vanes as de- termined in Par. 3.2.3.1 (a) was de termined from good turning vane prac- tice and is believed to be good for extrapolating the data to other angles and vane radii, For other angles the logs coefficient can be determined with reasonable accuracy by applying the bend angle correction factor of Fig. 1A-17 for rectangular elbows. 3.2.3.2 Splitters—The pressurelossin ‘arectangular elbow may be reduced by dividing the elbow with a number of flow dividers called “splitters.” These splitters divide the larger elbow into a ‘number of smaller elbows, each having a larger radius-to-duct width ratio and therefore a smaller loss. coetficient. When the splitters are positioned in a manner such that all smaller elbows have the same radjus-to-duct width Fig, 14-0—Thin vones, noncireular profile, coordinates. (Dato from Ref. 13)COMPRESSIBLE FLUID FLOW 0 ratio, the loss will be a minimum for a me Particular number of splitters. This fe ‘means thatthe ratio of the inside duct m radius to the outside duct radius ofall oe smaller elbows will be the same and is S equal to soz 5 cass 25 zo were: (r/ro)a = Ratio of inside re 2 dius to outside rae © = dius of any of sub- divided elbows 6 ita = Ratio of inside to 2030 40 80 60 70 80 90 0 outside radivs of BEND ANGLE, @ Deg original duct before Fe 1491 tae engl “sel gan tudo rein, ‘The Giciioney Woy. exbleo splitirs are added antock curve (shown by the broken diagonal line) is apparently esa of dierent = Number of smaller methods ved in developing the vane profiles. (From Re. 1), elbows, that is, one splitter, = 2 ce To cos 9/2 TYPE @ (@LUNT: Leaping eoce ‘oa TYPE & (Shap: LEADING EDGE aOUS= Galen) ‘The pressure loss of the new elbow ‘with splitters is calculated by using a loss coefficient based on the radius-to- duct width ratio of the subdivided el- Trees 2 AND 6: RAOUS 000% Ae FLow Fig. 14-32—Thick vanes, (Colevlated from Ref. 29) Fig. 14-34 —Effect of gop to chord 4 rotio on thin eiteulor ore vanes. (From Refs, 13 ond 15) 3 (0? Ela lO® 1 ie ores 45 67? € 910 fatogeril. GAP TO CHORD RATIO, sf 4 19 es eg 2 6 = Nga = 4410" 3, 2 ° © 10 20 30 40 90 60 70 80 90 100 0 0 20 30 40 50 60 70 80 90 100, BEND ANGLE, 8 Deg BEND ANGLE, # 065 Fig, 14:93—Losscoeficions for thn ieulor orc vones Fig. 1A-35—Losecoaficient for thin noncircular are vanes. (From Ref. 1)18 bows. See Fig. 14-18-20 for rectan- gular elbows. The Reynolds number should also be based on the subdivided elbow. It should be mentioned that turning vanes may be lighter and show an advantage in Reynolds number if large number of duct divisions are used, Ref. 27 contains additional in- formation on losses in elbows of this type 324. Special Elbows 32.4.1 Transition Elbows — The loss coefficient for 90 deg rectangular transitional elbows may be obtained by using the correction factor of Fig. 1A- 37 and the loss factor for 90 deg con- Mant area rectangular elbows (Figs. 1A-I8 through 14-20). 33. FLOW Lossés, DUCT AREA 33.1 Sudden Expansions or Con- tractions—The accompanying Bgures represent expansion (a) and contraction (©) in iretion ofthe flow, for which a ae Paar one (cs ee eee AP=Kq (ars) where q is the dynamic pressure or velocity head in the smaller area sec- tion. The loss coefficient K; is given in Figs. 1A-38 and 14-29, 33.2. Gradual Contractions—Table LA-5 gives the angular displacement versus pressure loss coefficient, as dia- ‘grammed in the accompanying sketch and expressed by Pe (« +4rp\a “) 1026) = mean equivalent diameter ate 8 2 oe Table 14-5 Angular Dxplacement rus Loss Coefficient = 20, 600 K 0-30 0 20-45 0.05 45-180 Sudden contraction from Ay 10 Az-as shown in Por. 3.31 Ret. 27, 9.15. SECTION 1—ENGINEERING FUNDAMENTALS 2 7 = = 0 2 a er REYNOLOS NUMBER IN 1,000 Fig. 14-96-—Lows cosficiants for 90 deg thick vone bends, (From Ret. 29) Ky OF CHANGING ARES BEND CconRECTION FAcTOR, ¢ = @ . 2 ee WioTH OF ENTRANCE WIOTH OF EXIT + B\/b— Fig. 14-37—Presture lots correction factor for transiionol elbows. Note: Radivs ratio ‘of constant area bend = t,/b) + 1/2 Fig. 14-98-—Lot cooficint of 0 oO Se SST 89 10 Sudden expansion. (From Ref. AREA RATIO, 8g 24)INCOMPRESSIBLE FLUID FLOW Eve Guane comer SEAS Ee AREA RATIO, A/2 Tora vos GoeFricienr, degli tits ox, $4 eae enero EXPANSION ANGLE, EXPANSION FACTOR, 00 2 30 40 50 60 10 Fig. 14-39 — Loss coofi- ent of © sudden con traction e210 Fig. 14-40 — Total loss coefficient versus. expen: sion angle 2 20 009 80 90 100 110 EXPANSION ANGLE, 28 Deg Fig, 1Ac41—Difuser exponsion factor. Curve A—stroight wall, conical and square: Curve B—straight wall, wo dimensioncl, Corve C—curved wall, conical “symmetrically abou 19 333 Diffusers Subsonic) 333.1 Geometry—Whenever pos- sible, a diffuser should be designed axis and with an expansion angle (28) between 7 and 12 deg. Nonsymmetrieal flow or too large an expansion angle may eause the flow to separate from the diffuser walls aand thereby cause an increase in pres- sure loss. When space limitations necessitate @ large expansion angle (26 > 13 deg), separation may be pro- longed by use of a curved wall or truncated diffuser, by the addition of ‘Buide vanes (Refs. 19, 22, and 24), or in extreme cases by the use of boundary layer control (Ref. 23) The ype of intersection (sharp oF rounded corner) between inlet duct and diffuser walls appears to have little ef- fect om pressure loss characteristics ex- cept at high entrance Mach numbers (fy > 0.7), (Ref. 21.) ‘The optimum expansion angle is a function of area ratio, surface rough- ness, and Reynolds number. For ex- ample, for a given area ratio and sur- face roughness, the expansion angle ‘with minimum pressure loss coefficient (as determined from Table 1A-5 and Eq, 1A-26) varies with inlet Reynolds number as shown in Fig. 1A-40, ‘These ‘curves are based on a smooth, straight- led conical diffuser with an inlet-to- cexit area ratio of 0.36 and a length-to- ‘mean diameter ratio of 0.25 (cot, ‘Some types of diffusers are listed here. (a) Straight Wall, Conical—For the fuser shown in the following draw- ing, Dy = 2Ltand + Dy aants® bes aL TAN @+D) (a27) (b) Straight Wall, Rectangular—Re- ferring tothe figure, values are: Two Dimensions a (A282) by = by + 2Ltand — (1A-28b) Square ay =a) +2Ltan@ — (1A-28a) by = by + 2Ltano (1.286)20 =< % : =e i Ca” ARE ‘Wo DiensionaL souARe ae cereal Ne ops bral TAN A bee beak TAN (© Curved Wall—At large expan- sion angles, diffusers of the shape shown (Ref. 1, Fig. 10) appear to have lower pressure loss than do straight wall diffusers. The equation for the ‘curved section is y= 3 z 2 fi + G/DVadan) =| | (129) curve secTion-~ STRAIGHT SECTIONS se < 29 < 40° 28 veg Fig. 14-42—Equivolent expan sion engle for a difwser fol- lowed by o resistance unit, (From Ref, 26) ACTUAL EXPANSION ANGLE, EXPANSION FACTOR, © Fig. 143 — Expon- sion factor foro. dit fuer followed by o resistance unit, (From ‘10 SECTION 1 (@) Truneated—This type is identi cal to a straight wall diffuser except that there is a sudden enlargement into the downstream pipe. This is useful where space is prohibitive and a curved wall diffuser is t00 expensive to fabri- hit 7e<28et 3.3.3.2 Pressure Loss—Pressure loss through a straight or curved wall diffuser may be expressed as the sum of ar-[ew+ 4] a Dn C = Exyunsion itor (Fig 1-41) = Und ASG or ola say Das natu guid eas fy 24 oe al Ag, Ry f te ° 010 20 30 40 80 60 70 EQUIVALENT EXPANSION ANGLE, 28 Deg Ky pe7e: 230 40 8060 70 80 90 Ref. 26) EQUIVALENT EXPRNSION ANGLE, 28 Deg ENGINEERING FUNDAMENTALS. Dit Da = (A30) ‘The loss through a truncated diffuser is equal to that of a straight wall diffuser plus the loss in the sudden enlargement at the exit, Based on the velocity head at Aj, the pressure loss equation becomes - +e apa lcKu+ FF 41 +kalT) a Asn where K 42 is the loss coefficient for a sudden expansion from 4 30 A The pressure loss determined from the preceding values will be less than the actual loss if the diffuser is directly preceded by a bend or a long (L/D > 10) straight section of ducting; it will be higher than the actual loss if the dif- fuser is directly preceded by some in- stallation that produces rotational flow 3.333. Pressure Recovery Equa tions—The equations given below are applicable to incompressible flow only. (a) Exit Static Pressure -»] (A-32) (b) Dilfuser Bfficieney—In some re- ports diffuser performance is expressed as a” dimensionless term called efli- ciency (n): Fa= Fa = Pa qi= 42 gill ~ (1/427) Wasa) (©) Loss Coefficient in Terms of Ef ficiency — This is expressed as ce —nfi-e] aan 33.34 Dilfuser Followed by a Re- sistance Unit—The addition of a re- sistance unit, such as @ heat exchanger, immediately downstream of a diffuser decreases the tendency of flow separa: tion and thereby reduces the pressure losses in wide-angle diffusers, Ref, 26 recommends the shape shown here and gives corresponding pressure loss characteristics for a con cal diffuser. This information may also bbe used as a guide for nonconical dif fasers.+ tere! INCOMPRESSIBLE FLUID FLOW De | In this illustration 2 ¢ is the actual ex- pansion angle between walls (Fig. LA- 42) and 20 is the equivalent expansion angle. Recommended values of 2 ¢ for various area ratios and equivalent ex- pansion angles are given in Fig. 1A-42. ‘The pressure loss through this type of installation may be estimated from Eq, 14-31 for diffuser pressure loss and the ‘expansion factor C from Fig. LA-43:, Lee prs Ky, = Pressure loss coefficient of resistance unit following diffuser Data for other area ratios are not avail- able. 3.34 VenturiThe irrecoverable pressure loss through a venturi may be considered to be the sum of the con- traction, friction, and expansion losses With normal venturi design the con- traction losses may be neglected and the pressure loss equation becomes ar-(D Ys cra carom where: B(4 L/D) = Sum of friction toss coefficients in en- ance, throat, and diffuser sections. (See Eqs. 14-30 and 1A+ 30) CK, = Expansion loss co- efficient (Eq, 14-31) A rapid approximation of this pressure loss may be made from Fig. 1A-47, where the pressure loss is given as a function of the differential static head, 335 Orifice in a Duct—The ir- recoverable pressure loss of an incom- pressible fluid through an orifice (see sketch) may be calculated from the equation When 4, = 43, K, may be estimated from Fig. 1A-44. A rapid approxima tion of AP, may be made from Fig. 1A-47, where A P,is given asa function of the differential static head. When 4) = 4s, the loss coeificiont K,may be taken as that of an abrupt, ‘expansion from the vena contracta (A)) to As, based on the velocity head at the vena contracta. Correcting the loss coefficient at 4, conditions to 42 conditions results in AP, = Kegs 26 24 22 20] 2 7 1 a? sone =|] Ae coat Ar Kean be determined from Figs. 1A-45 and 1A-46, knowing A2/Ay and Aa/ As % 2 To AREA RATIO, Ag/) Fig, 1A:44—Losscoefcient for on orifice ina duct: Ay = As. (FromRef, 45, Table 11) a oe ak se AREA RATIO, iA Fig. 1A-45—Loscoofi- lent for an fice in Pro duct Ky = (1/C) (AafAa)]?n 13.4 FLOW LOSSES IN DUCT BRANCHES —The pressure loss due to mixing in ‘converging and diverging branch ducts AP = Kia ‘The loss coefficients discussed below do not include wall friction. 3.4L Diverging Branches—The loss coefficients (K,) are given in Fig. 1048, 3.4.2 Converging Branches — The Joss coefficients (K,) may be estimated from the equation le At ska ha feat a) b=") ‘where: | Ks = Less coefficient from 1 to 3 based on gi (Fig. 14-49) = Function of @ (Fie 14-49) 24 capt = Fupon of snd = Ay/ As (Fig. 1A-49) 242 cos’ = Function of 8 and A2/ : As (Fig. 14-49) 3.43 Manifolds—The symbols used in this paragraph pertain to the fol- lowing figure: ore ses eT ARER RATIO, A\/By Fig. 1A-46—Contraction coffiient of an orifice. (From Ret, 45, Tables 8 and 13) SECTION ‘Three of the variables that influence the rate of discharge from the ports of a manifold are the manifold cross-see~ tional area, port area, and port spacing. Flow regulation may be accomplished by varying either one of these while keeping the other two constant, or by any other combination. These var- iables may be estimated by solving the succeeding equations, It should be noted, however, that values of K, and 4fAL/D are only approximate and exact sizing or flow distribution should be determined by test. ‘The total pressure loss from any one port of « manifold through the next port is equal to ane fer fede where: K, = Entrance loss coefficient into the port (Fig. 1A-48) 4 = Friction factor based on the Reynolds number in the manifold (Fig, 1A-4) AL = Center-to-center distance between ports D = Manifold equivalent di- ameter q= Velocity head in the manifold between ports This relationship is applied as fol lows to determine the flow through the 00 90 20 70 60 s0| 1/6 ~Pea) | ENGINEERING FUNDAMENTALS. various por First Port (Ref. 66): Pam Rae Kude Second Port: r= ta-[fo oe] We= Wi + Ws ‘Third Port: asa. ra to 5 at Ko [tae We = Wy + Wa We Wen wit wrt + want he ‘These equations may be solved simul taneously of by substituting various as- vewruai wire Worst tae ee oe sen panio 22 «THROAT DIAMETER DIAMETER RATIOS’ "Duct DIAMETER Fig. 1A-47—Approximate pressure loss through several ASME low metering devices: Py, = upstream stotic pres: ure; Pia = throot static pressure. (From Ref. 35, Fig. 5)INCOMPRESSIBLE FLUID FLOW ° U 2 3 4 3 RATIO OF WEIGHT VELOCITIES, G/G, Fig. 14-48—Diverging branch, loss coefcint, (From Ref. 7, revised at G,/G <1) 2.0 it le a 60, 20 BOR B, beg ° 20 40 2 sumed values of v,. The quantity w, and/or P,, may then be calculated and ‘compared with the known conditions. the fow is determined by assuming values of w., values of A’, may be de- termined from Fig, 1A-48. If the equa- tions are solved simultaneously, as- sumed values of K, may be used. A recommended value for assumption is Kyat Ge/ Gy = Arasioi/4 poe 10 an open-end manifold this assumption ives reasonable results along the en- lire length of the manifold; however, in 4 closed-end manifold the weight ows will be in error near the closed end, Ref, 66, pages 120-122, makes the fol- lowing assumptions for calculating ‘manifold characteristics: (a) Equal weight flow through each port (wy; = Ws = Wp) (b) Equal static pressure in all ports Cn» Pa Pa (©) Equal friction loss coefficients be- oveen ports 4fAL,_ 4faly 4fabs Puen © D (d) Constant manifold —cross-sec- tional area(4_ = 4), «The various port areas may then be calculated from the following equation; 720) 740 | Fig. 14-49—Converging branch, functions for computing st efficient. (From Ref. 7)a torsbe(a > heme \ sone {es 4 GAL/DIU/OY + Dy + 1D) as thatis, » = 4 for(n ~ 4)th port, y = (x ~ 1) for first port (continuing for all ports) where mis the total number of ports Ifthe port areas are proportioned ac- ‘cording to the preceding equation, the total weight flow or upstream total pressure may be calculated from [sed Ke er Ae Py LLL VASA L/ DY n/OKn—YOn— 1) Pat Fe 4 Ue Pu ‘A manifold may be treated as an i finite reservoir (Pie = Py = ~~» Pu) when the velocity head in the manifold SECTION 1 ENGINEERING FUNDAMENTALS, is small compared with the static pres- sure. q/ P, < 0.01 3.5 FLOW LOSSES, INTERNAL IN- STALLATIONS. 3.5.1 Vales—In most variable ge- ‘ometey components the pressure loss characteristics may vary considerably With size and manufacturer. Addi tional information on valves is con tained in Section 3H, Par. 7. The fol+ lowing loss coefficients should be used only in lieu of the absence of manu- facturer’s data. 35.1. Gate Valves—Refer to Fig. 14-50. 35.1.2. Butterfly Valves—Fig. 1A~ SI gives loss coefficients for a butterfly valve that is fully closed when a = 0. A slightly different valve is one that is fully closed when a equals some angle ‘a, The loss coeficient for this type ‘of valve can be estimated from the same figure cos! £08(a0 + a1) where aig < 20° and a, is the valve ‘opening from the closed position ay. 35.2. Screens and Grids—The pres- sure loss coefficients Fig. 1A-52 (based on upstream velocity head) are for 100 29| «| 190 80 40 69 a Tan IL 40| of 4 5 20] 19] 8 6 10 Z 8 Ba 6 e3 g 4 2 : z ge | z 6 \ 8 4 6 7 4 2 f 2 ° \ ae a ee see ‘0 80 Fig, 1A-50—toss cooficiens, gate valves. (From Ref. 12) Fig. 1AS1— sereens mounted in a duct where the upstream and the downstream areas are the same, The loss coefficients are for Reynolds numbers (based on the sereen open area and velocity) greater than 400. At lower Reynolds numbers the loss coefficient increases rapidly (Refs, 48,53, and 56). ‘The solidity, gis defined as the ratio of the solid (no-flow) area divided by the total (no-flow plus flow) area of the 35.3 Heat Exchangers 35.3.1 Installation Angle—Where space is limited, diffusion and turning may be accomplished with minimum increase in pressure loss by setting the axis of the heat exchanger or similar resistance unit at an angle to the centerline of the entrance duct, as shown here. =z ‘This makes the effective area of the duct at the face of the heat exchanger equal to 4 j/cosa. Fig, 1A-S3 was ob- tained for air but should hold reason- ably well for other fluids. For this BUTTERFLY THICKNESS » DUNE VALVE BORE AREA (+ 0%8) 3040 60 6070 BUTTERFLY ANGLE, @ Deg 2030 Lots coeflcient, butterfly volves. (From Ret. 65){= Loss coefficient of heat ‘exchanger plus turning ‘angle at approach angle fa = Loss coefficient of the re- istanee unit at zero an- ‘ale of approach, row Lossis, Exits 1 General—Exits from dust into compartments are extremely and there are probably no two Exits are usually one of three nondiffusing (grills, slots, and ), diffusing (most circular over- types) and ejector. Grills and Perforations—In ‘of specific manufacturer's data ure loss for grills and perfora- ‘ean be estimated by adding the loss for a screen (Fig. 1-52) Dressure loss for an abrupt ex- from the total area of the exit flow area of the enclosure into ‘the air is discharging. Perforated -will match the square edge data, Screen type grills will match the wire sereen data. Round wire exits produce less noise than ted grills. The noise level of ated grills can be reduced by put- ‘ing round wire scrcens behind them, 3.6.3 Slots and Nozzles —The pres- “sure drop for slot type exits that have 8 88 888 ee ke a SOLiDiTY, © Fig. 14-52-—Loss coeffcionts for screons and grids, (From Ref. 56) ‘well rounded entrances can be de- termined by using a loss coefficient for aan abrupt expansion from the exit area to the flow area of the enclosure (see Par. 3.3.1). When the entrance is not well rounded, the loss can be deter- ‘mined by using a loss coefficient based ‘on an abrupt expansion from the vena contracta ares to the flow area of the ‘enclosure (vena contracta area = free flow area of exit multiplied by dis- ccharge coefficient). Nozzle type exits have the same loss coeflicients as the ‘well-rounded slot type of exits. In all subsonic types of exits where there isno vena contracta, the static pressure at the exit plane of the exit is ‘equal to the static pressure of the en- closure into which the exit is discharg- ing. When a vena contracta is present, the static pressure atthe vena contracta {sequal to the pressure in the enclosure. For sonic or supersonic flow, see Part 1B. 3.6.4 Diffusing Exits — Manuface turer's pressure drop data should be used for diffusing outlets because de- signs of these outlets are so varied. An approximate pressure drop ean be ob- 10 «10. ME’ ROUND WIRE SCREEN 2s tained by applying loss coefficients based on the diffusion angle (see Par. 3.33), 3.65 Velocity Distribution in En closures—To have a comfortable en vironment in an occupied enclosure, the velocity distribution should be such that the air velocity on personnel is ess than 50 ft/min, A velocity between 25 and 30 t/minis desirable. The follow- ing discussion is confined to free air jets except where noted, A free jet is one in which the flow area of the jet is less than one-fifth the flow area of the en- closure. The velocity from an exit into aan enclosure is constant for a distance of about four effective diameters and is ‘equal to the exit velocity, The effective diameter for rectangular slots is equal tothe diameter ofa circular area which is equal to the free flow area of the rectangular slot times its discharge coefficient. For distances from 4 De to 8 Da the centerline velocity ean be caleulated from 010 2030-4050 60 10 80 90 APPRONCH ANGLE, @ 000 Fig. 1A-53—Effect of angle on heat exchangers. (From Ref. 6) 4 200 corm vam. Pie s0 WAG BTL waoTh DUCTING 8 11+ 560 8 sr 2 7 8 8 10 Fig, 14-54—Flow lots colevations26 where: b = Width of jet of air at out- letor vena contracta, ft x = Distance from the face of the outlet, ft V = Initial velocity ofthe jet = Fe CaR jo ft/min V. = Velocity “at x distance from outlet face, ft/min Coefficient of discharge of theexit Ry = Ratio of free area to total area of exit Dg = Effective diameter of the outlet, ft C1 = Constant from Table 14-6 s For distances greater than 8 Dg from the exit face, the following equation can be used to calculate the centerline ve- locity down to a ¥, equal to 150 fy min. When ¥, is between 150 and 50 fA/min, the same equation can be used, but C’ should be reduced by 20% "ETERS amine GVAD x> De SECTION 1-1 where: Q = Airflow, cu ft/min A, = Total area of exit, Aq = Effective area of exit AcCaR jo ft? If we define a distance x at which the centerline velocity is 50 ft/min and call this the throw of the exit, it can be de- termined from Throw = wi ‘The velocity at locations other than on the centerline may be obtained for diss tances greater than 8 Dy from the exit, using the following equation: Zi =33los where: V_ = Centerline velocity at plane ofr r= Radius from the center- line of jet. ros = Radius at which ¥ 05%. The radius rps occurs at an angle equal to one-half the angle of diver zgence, The angle of divergence is equal ENGINEERING FUNDAMENTALS. 10 20-24 deg for slots and nozzle types, and 18 deg for grills and sereens. ‘When a jet discharges along a wall, the jet behaves the same as a free jet that has a width or diameter twice as large as the jet discharging along the wall and whose centerline velocity is atthe wall, The C’ values must also be multiplied by V2. 3.7 EXAMPLE OF FLOW LOSS CALCU raTions—Caleulate the low losses in the example illustrated in Fig. 1A-S4, Table 1A Volues of C' = Constont_ es eS 500- 2000- a ps nia 1000 116100 Openings Round or square S77. Rectongular, large ‘ipectratio < 40 49 60 Annuler sets, ‘axial or adi 3948 Grills ond Grids Free area 409% ‘or more 4757 Perforated Panels Freoarea3-5% 3037 Freearea 10-20% 40 49 Toble 147 Exemple of ofat@] « Location and/or | Description | Pinte, w, | Dor (Gee Fig. 14-54) | inc Hg | 1.°R |lb/in| By. in Beg 40 | 560 | 200 | 6 Difiner | 39.84 | 560 | 200 | 6 Or ® ie elbow | 39.21 | 560 | 200 | 8 One ‘Oto 39.78 | 560 | 200 | 7.50v Or® | ened 39.7 | 560 | 200 | 2.97 30.71 | 560 | 200 | 7 39.63 | 860 | 200 | 7 = oe “Abrupt Ex pansion | 39.27 | 560 | 200 | 7 Total olelale a Avtslet| orexit,| thy | og. | a in® |mninem.?|th/f? fin. HoO | Now 20. | 7.06 |o.ovas| 905 | 7 x 10° 283 | 7.08 Jo.ovas) 9.10 | 7 x 10° soa | 297 loovas| 297 | 53 x 10° 503 | 297 loovaz| 267 |a98 x 10° 49 | 408 Jo.0042| 304 | 2 x 107 49 | 408 |o.0r0| a04 | 47 x 10° 49 | 408 |o.0%39| 3.04 25 | 8,00 Joosse| ... | 84x 10° 49 | 4.08 |0.0990) 3.08 Example colton for @ and Oe For columns 1,2, ond 3 assume typo values © = (35) = 200/283 ~ 7.06 pq ~ 1526P9/T = 0.0948) (Table 142) 2 = GipgC ~ 7.087/0.0748) x 38.4 ~ 9.05 (Table 14-1) 4b = 00125 torw/O = 0.355 (fom Fi. 148) 4/0 = 00123 » 120/6 = 0.245. incolumn 17) BP, = (6) x (18) = 905% 0248 = 222m. HyO"RESSIBLE FLUID FLOW determine the proper number of srcular turaing vanes for the mini= loss in the 7 in, square elbow. ie through the exit screen avery large enclosure (A = -,) calculations and a recommended format are given in Table J. Most ofthe calculations are self= tory. Items that are felt to re- further explanation are: |For areas (2) to (3) in Fig, 14-54, in 9 is determined at the inlet, 16 is based on this Reynolds and column 17 is based on the ic average diameter. For areas (4) to (5), all calcula are based on average equivalent ‘of the inlet and the outlet. For (5) to (6), De is equal to the 0 check the applicability of the (7), the second line of calcu determines the Reynolds num- ‘on the free flow area of the to check the applicability of REFERENCES: 1, John R. Henry, “Design of Power Plantinstallations, Pressure-Loss Char- acteristics of Duct Components NACA Wartime Report L-208, June, 194s. 2. J.T. Higgenbotham, C. C. Wood, and E. Floyd Valentine, “A Study of the High-Speed Performance Charac- of 90° Bends in Circular NACA TN 3696, June, 1956. 3. O. P. Lamb and James S. Hold- hhusen, “Investigation of Aireraft Duct- ing Components at High Subsonic Speeds.” Fluidyne Report WADC TR 56-187, ASTIA Document 97321, Sep- tember, 1956. 4. Charles H. McClellan and W. A. Bartlet, Jr. "Investigation of Air Flow in Right Angle Elbows in a Rectangular Duct.” NACA Wartime Report 1-328, October, 1941 5. R.D. Madison and J, R, Parker, ressure Losses in Rectangular Ducts,” Trans. ASME, Vol. 38 (1936), 167-176, 6. Mark R. Nichols, “Investigation of Flow Through an Intercooler Set at az Jax] om] os] aa | an ] ae ] om (vend) any, or20 or "| ck, | ar, ity | c | cx, 44 | 41/0, b+49/0,| in. HO 120 [00123 | 0.246 | 0246) 2.22 10 0.075} oor4s| 11.35]0.0123 | 0020 | 0034 | 0.91 90 [10 |ors | 1256/0013 | 0020 | 020 | os7 60 Jo013 | 0.104 | 0.104 | 0.30 90 : 020 | 061 ‘ 3 0340 | 0:340| 1.03 [16s | 495, 10 | 308 1307 YYorschord © = iam 7 Example Colaloin for fring Ky for vaned bow zpgsind/2 = 298 ‘bfe(in @/2) ~ 1 rom Eq, 1819) 22 ~ Yauin. (Sm = spoce between vanes se = 4 ont From Fgt. 18.33 ond 18.94, ~ 0.20. 2” Various Angles to the Supply Duct.” NACA Wartime Report L-408, April, 1982, 7. Andrew Vazsonyi, “Pressure Loss Elbows and Duct Branches,” Trans, ASME, Vol, 66 (1944), 177-183. 8. John R. Weske, “Pressure Loss ia Duets with Compound Elbows.” NACA Wartime Report W-39, Feb- ruary, 1943, 9. Loring Wirt, “New Data for the Design of Elbows in Duct Systems,” General Electric Review, Vol. 30 (June, 1927), 286-296, 10. W. H, McAdams, Heat Trans- ‘mission. New York: “MeGraw-Hill Book Company Inc., 2nd ed., 1942, p18. 11, John H. Perry, Chemical Engi- neers’ Handbook. New York: McGraw- Hill Book Company, Inc., 3rd ed,, 1950, pp. 370-373. 12. “Airplane Heating and Vent lating Equipment Engineering Data, SAE Information Report No, 23, October, 1951 13. G. N. Patterson, “Note on the Design of Corners in Duct Systems.” R & M1773 British), October, 1936. 14. Lewis F. Moody, “Friction Fac- tors for Pipe Flow,” Trans, ASME, Vol. 66 (1944), pp. 671-677, 15. C. Salter, “Experiments on Thin * Turning Vanes.” R de M 2469 (British), 1952. 16. H. Peters, “Conversion of En- ergy in Cross-Sectional Divergences Under Different Conditions of Inflow.” NACA TM 737, 1934, 17. AN. Vedernikoff, “An Experi mental Investigation ofthe Flow of Air in a Broadening Channel." NACA TM 1039, 1944, 18. Jerome Persh, “The Effect of Surface Roughness on the Performance of a 23° Conical Diffuser at Subsonic Mach Numbers.” NACA CR 3737, 1932. 19. Elliot G. Reid, “Performance Characteristics of Plane-Wall Two- Dimensional Diffusers.” NACA TN 2888, February, 1953. 20, B.H. Litile and Stafford W. bur, “Performance and Boundary~ Layer Data from 12° and 23° Conical Diffusers of Area Ratio 2.0 at Mach Numbers Up to Choking and Reynolds Numbers Up to 7.5 x 106” NACA ‘TR 1201, 1954. 21. Martin R. Copp, “Effects of In- let Wall Contour on the Pressure Re- covery of a 10° 10-Inch-Inlet-Diameter2» Conical Diffuser.” September 5, 1951. 22. Carl A. Moore, Jr., and Stephen J. Kline, "Some Effects of Vanes and ‘of Turbulence in Two-Dimensional Wide-Angle Subsonic Diffusers.” NACA TN 4080, June, 1958, 23. William J. Biebel, ““Low-Pres- sure Boundary Layer Control in Dif- fusers and Bends.” NACA WR L-84, as, 24, G. N. Patterson, “Modern Dif- fuser Design,” Aircraft Engineering, Vol. X, No. 115 (September, 1938), 267-273, 25. Richard Scherrer and Warren E, Anderson, “Preliminary Investigation of a Family of Diffusers Designed for Near Sonic Inlet Velocities." NACA "TN 3668, February, 1956, 26. Charles H. MeLellan and Mark R, Nichols, “Investigation of Diffuser Resistance Combinations in Duct Sys- tems.” NACA WR 1329, February, 1942. 27. “Fan Engineering.” New York: Buffalo Forge Co., 4th ed., 1938 28. C. P. Kittredge and D. 8, Raw- ley, “Resistance Coefficients for Lami- rat and Turbulent Flow Through One- Half-Inch Valves and Fittings," Trans. ASME, Vol. 791951), 1759-1766. 29. AR. Collar, “Some Expe ments with Cascades of Acrofoils.” R & MNo. 1768, 1937, 30. K. J, Bell and O. P. Berges “Flow Through Annular Orifices, Trans. ASME, Vol. 79 (1957), $93-599. 31. E.J. 0. Brandt, "How You Can Measure Airflow,” Product Eng. (June, 1957), 145-150. 32. Edgar Buckingham, “*Notes on Some Recently Published Experiments (on Orifice Meters," Trans, ASME, Vol 78 (1986), 379-387. 33, R. G. Cunningham, “Orifice Meters with Supercritical Compressible Flow,” Trans. ASME, Vol. 73 (uly 1951), 625-638. 34. A. L, Ducoffe, JR. Bennett, and . G. Ray, “Subcritical and Critical Flow Through Straight-Through El- NACA CR 3569, SECTION 1 ‘bow, and Tee A-N Fittings and Sharp- edged Orifices at Elevated Temper tures,” Trans. ASME, Vol. 80 (October, 1958), 1349-1357, FlowMcasurement, 1959.” ASME Power Test Codes, PTC 19.5.4, 1959, 36. H. P. Grace and C. E. Lapple, “Discharge Coefficients of Small Di- ameter Orifices and Flow Nozzles,” ‘Trans. ASME, Vol. 73 (1951), 639-688, 37. H. W, Iversen, “Orilice Coef- ficients for Reynolds Numbers from 4 to 50,000,” Trans. ASME, Vol. 78 (1956), 359-364, 38. A. L. Jorissen, “Discharge Co- ficients of Herschel Type Venturi Trans. ASME, Vol. 74 (1952), 905-913, 39, A.L. Jorissen, “Discharge Meas- urements at Low Reynolds Numbers— Special Devices,” Trans. ASME, Vol 78 (1956), 365-372. 40. A.L, Jorissen, “Discharge Meas- urements by Means of Venturi Tubes,” Trans, ASME, Nol.73 (1951), 403-41 41, A.L. Jorissen and H.T. Newton, “Discharge Measurements by Means of Cylindrical Nozzles,” Trans. ASME, Vol. 74 (1952), 825-835. 42. LO, Miner, “The Dall Flow Tube," Trans, ASME, Vol. 78 (1956), 475-479. 43, J. A. Perry, “Critical Flow ‘Through Sharp-edged Orifices,” Trans. ASME, Vol.71 (1949), 757-768. 4, Miguel Rivas, Jr., and Ascher Shapiro, “On the Theory of Discharge Coefficients for Rounded-entrance Flow Meters and Venturis,” Trans ASME, Vol. 78 (1956), 489-497. 45, J. M. Spitzglass, “Orifice C ficients —Data and Results of Tests,” Trans. ASME, Vol. 44 (1922), 919-974. 46, Joel Warren, “A Study of Head Loss in Venturi Meter Diffuser Sec- Trans. ASME, Vol. 73 (1951), 399-402, 47. Alexander Weir, Ir, J. L. York, and R.B. Morrison, “Two and Three Dimensional Flow of Air Through Square Edged Sonic Orifices,” Trans. ENGINEERING FUNDAMENTALS. ASME, Vol. 78 (1956), 481-488. 48. B. Eckert and P. Pfluger, “The Resistance Coefcient of Commercial Round Wire Grids.” NACA TM 1003, January, 1942. 49. Alfred A. Adler, “Variation With Mach Number of Static and Total Pressures Through Various Screens NACA CR 2007 (23), February, 1946. 50. §. F. Hoerner, “Pressure Losses ‘Across Screens and Grids.” AFTR 6289, November, 1950. SI. W. D, Baines and E, G. Peter- son, “An Investigation of Flow Through Screens,” Trans. ASME, Vol 73 (1951), 467-480, 52, L. L. Repko, “Pressure Drop Across Various Commercial Screens,” ‘Design News January, 1951), 51-53. 53. W. G, Cornell, “Losses in Flow Normal to Plane Screens,” Trans ASME, Vol. 80 (1958), 791-799. ‘4, R, E, Dannenberg, B. J. Gam- bbucci, and James A. Weiberg, “Per forated Sheets as a Porous Material for Distributed Suction and Injection.” NACA TN 3669, April, 1956 35. L. F. G, Simmons and C. F. ‘Cowdrey, “Measurements of the Aero dynamic’ Forces Acting on Porous Screens.” R & M 2276, British R.A.E, 1945, 56. G.B, Schubauer, W. G. Spang- enberg, and P, S. Klebanoff, “Aerody~ namic Characteristics of Damping Screens." TN 2001, January, 1950. 37. P. Jonas, “The Changes Pro- duced in_an Air Stream by Wire Gauze,” Engineers Digest, Vol. 18 (May, 1957), 191-193. 38. A. L. Loeffler, Jr, and Morris Perlmutter, “Turbulent Flow Through Porous Resistances Slightly Inclined to the Flow Direction.” NACA TN 4221, February, 1958. 59. “Flexible Components for Pneumatic Ducting Systems,” Arrow- head Produets, Technical Bulletin $16, 1959, pp. 1-8. 60. C. M. Daniels, “Pressure Losses in Flexible Metal Tubing,” Product Engineering (April, 1986), 223-227.