NTAL EQUATIONS OF VISCOUS COMPRESSIBLE FLow 52 FUNDAME! : teal) sstgoatalea faa al 2 sin® 2 ' Avg | vacotd rsin@ do t 1.19 | THE ENERGY EQUATION ; - ‘The energy equation is based on the first law of thermodynamics which is generally referred to ag the principle of conservation of energy . . Let be an arbitrary closed surface enclosing a volume T of homogeneous fluid ( ie. of constant density p). We assume the velocity vector Vand the heat flux vector q acting at a point on - v the surface element as shown in figure (1.20). The vector q represents the heat flux by conduction and radiation . The total rate of heat flow into the surface S is then given by Figure (1.20) Be - {Ja tas Parnes () Work may be done at the boundary of the system by normal and tangential stresses . The rate of total work done by the fluid on the surface S is DW ff Dr JJ uijinyds e) s Now the first law of thermodynamics requires that the net rate of increase of total energy (ineluding kinetic, internal and potential ) into S is equal to the rate at which heat flows into S, plus the rate of internal heat generation , minus the rate at which the fluid does work on its surroundings . Hence we have pE.2 [Noes [Joeamas =. Wornas+ [Porounase [fer @) - oe ve Remember that E , the total energy per unit mass > is given by E = S+Utw; where y is the gravitational potential denoted by gz. Also, q) is the intemal heat generation rate per unit volume. Using Gauss’ Divergence theorem , the surface integrals in equation (3) can be converted into volume integrals and we getUID DYNAMICS viscous FH 53 JIN 90s SES coeayar-- INE eave Ifans or [[[Zeeer were ai ae 8j,)- @ oxo tT Sno vole 1 havi, the ined penerbe eo Hence Seve +3 SPE uy, 3 qi- % Fe vitiid-q =0 2 ~ or Bipeyge (pEuj) = ree Oa “ ‘The stress tensor oj; May be separated into pressure part and a viscous part as, oj) = oi = - PB toy © du, ' 2 where oj) = aus Fotn one a a a bo Hence Sey S (pew) = Bagi] PU) FR; (6) Bu a bi) = aus ) [see page (34) } therefore equation (6) becomes a a ' ' Bx (PMID, UISITG 2 oa - 2 DPED+5, (PEM) ~ “3x, Now by expanding the terms , we get i ou dE, oe ap 8, Ott OP yy 28H OMG of Sou 28] e[S8esS coe] -anyti-Pany Hans on, 78 2,4 But by continuity equation eo slew =o QE aE DE Also FF *Ui5x, 7 Dt and so from equation (7) DE oui aii gig (8) ogee. 7 FOP ax, -u 3 + Yay, q Where @ = o', 24 isthe viscous ~ ipation function . NOx;VISCOUS FLUID DYNAMICS, 55 Then equation (14) becomes q- ~KVT+G, ar or Ka tay, (1s) and finally fom equations (13) and (15), we get BULA at ui Pdr Pn (KE) - Bo taurerBteorg (16) Usually q, (ie. 4, ;) is negligible and q’ is zero xcept for cases involving such things as resistive heating . It is seldom in fuid dynamics that Gi for completeness . ‘Thus energy equation (16) for D pou. * (x32) oui and q’ must be considered . but we have included them viscous compressible flow reduces to Dt Ox, )-Pax,t® (7?) This equation is too complicated to solve analytically and consequently is greatly simplified in most fuid flow problems . PARTICULAR CASES The energy equation for flow: with certain constant physical properties ofthe fluid may be written from equation (17) as follows : (@ VISCOUS COMPRESSIBLE FLUID WITH CONSTANT THERMAL CONDUCTIVITY 'fthe thermal conductivity K is constant, energy equation (17) becomes DU oT. dui Dt” Kox,ax, Pax, t? a (ii) VISCOUS INCOMPRESSIBLE FLUID @ For an incompressible fluid = = 0 and equation (17) reduces to i DU_LdO/(,aT (19) PD ox, (xiZ)+e (ii) VISCOUS INCOMPRESSIBLE FLUID WITH CONSTANT ~— THERMAL CONDUCTIVITY ; For an incompressible fluid with constant thermal conductivity K , equations (17) and (18) DU sh) +o 20) Titeeto ppp = K (Geax 1D () NON- VISCOUS OR INVISCID FLU For non — viscous or inviscid fluid , we have © = 0 arid equation (20) reduces to i (21) PD DKW T0 FUNDAMENTAL EQUATIONS OF VISCOUS COMPRESSIBLE Fig, 1.23 EQUATION OF STATE FOR AN IDEAL GAS - are virtually no mutual IF the molecules of a gas have negligible volume and there rs seas. Mal atacy between the individual molecules , the gas is said to be an ideal ( or pe I lt y= ¥A(p.7) o ay, then dV, = (4) ao+(24) aT Q) 7 , Now Boyle's law states that for constant temperature , the volume is inversely proportional 10 the press, of the gas . That is Vy ; (T constant ) c or Vv, = > GB) where C, isa constant of proportionality . Charles's law states that for constant pressure , the volume of the gas is directly proportional tote temperature. That is Vy« T (p constant ) or Vy" aT (4) where C, isa constant of proportionality . From equation (3) ov, ) c_ Y, —t| o = ( op), ot Pp (5) and from equation (4) (4) -c,-% aT) rT Oy From equations (2) , (5) , and (6) , we have Vs 454 Me dv, = ~p opt ToT dV, dp dT —,—.= 7 MTR UT Integrating , we get InV,+inp = InT+InR or pVy = RT ” or p=eRT @ where R is the gas constant , the value of which depends on the BAS concemed : , tion ot ; . Equation (7) or (8) i elle he equation of Sat for an ideal gas. Any gas which obeys eq (8) is called an ideal gas .8 VISCOUS INCOMPRESSIBL Fu N no general method for solving Navier — Due to the non ~ linearity of the convective terms , Stok a equations is available. However , in most practical problems » reasonable assumptions can be made hig Considerably simplify the solution of the Navier - Stokes’ equations - PARALLEL FLOWS . ie vewso A flow is called parallel if there is only one velocity component i.e. The pracy application of this simple case is the flow between parallel flat plates ( Planes > + GiTCUlar pipes, Concentric rotating cylinders . In such one — component flows , the Navier ~ Stokes’ equations simpli considerably , and infact permit an exact solution . ou The equation of continuity (2) in this case reduces to $=" = 0. Thus for parallel flow us u(y.z.t),v=0,w=0 i.e. u the velocity component in the x - direction cannot depend on x . Furthermore , Navier ~ Stokes’ equations (1) become ou G) Here the first equation in (3) is a linear partial differential equation for u(y ,z,t) and the second ad third equations indicate that p = p(x,t). We can write (3) as op Gs a) du oe. (24 ou ox yo az )-P ar é a (4) and SPL SP _ dy” oz We now discuss some particular problems which can be solved analytically . 2.3. STEADY LAMINAR FLOW BETWEEN TWO PARALLEL PLATES Consider the steady laminar flow of a viscous incompressible fluid between two infinite horizon! parallel flat plates . Let x ~ axis be the direction of flow and y — axis perpendicular to the direction of flow. Let the distance between the plates be h and the width of the plates in the 2 — direction be infinite- Here the word infinite implies that the width of the plate is large compared with h so that the flow may be treated as two — dimensional Let the plates be long enough in the x ~ direction for the flow to be parallel. Let the lower plate y = 0 be at rest and the upper plate y = fh be moving with uniform locity U_ in the x - direction as shown in figure vel ‘Then since the flow is steady , we have Qn). 2Y 9, the Navier - Stokes" equations (4) becomeyiscOUS FLUID DYNAMICS op_dP_ atu 3 Bx dx” Pay? (sie $2. 22g “ ° : 32 ~ 0 and P= wo since v= ¥(9) Le U isa function of y ato ne, the L.H.S. of equation (1) can be either a function of y oy constant , But since = 0 ie. p cann OY P cannot be a function of y, it follows that the term ae must be a constant . Integrating equation (1) twice wrt, y, we get x au _1 4p 1 paxyta d o < Fe seg of wdx 278% Q) id B i ich wil . where Aan ne arbitrary constants which will be determined by the boundary conditions ofthe flow problem under consideration . We discuss the following three different types of flows () | SIMPLE COUETTE FLOW The simple Couette flow or simple shear flow is the flow between two parallel plates one of which y = 0 isatrest and the other y = h moving with a uniform ( constant ) velocity U parallel to itself in its ‘own plane as shown in figure (2.2). The pressure gradient in this case is taken to be zero . Thus the flow is entirely due to the snotion of the upper plate . The boundary conditions in this case are = u=0 aty=0 and u=Uaty=h (ieu(0)=9, u(h) = U) Since the pressure gradient is zero , equation (2) reduces to u(y) = AyP Moving plate Using u(0) = 0,weget B = 0, therefore u(y) = Ay Using u(h) = U, we get U = Ah U o ATE n x Thus wy) = By ® 9 Fixed plate ion is Figure (2.2) or innon - dimensional form , the vetocity distributi Bae hate ina fui ing the upper plate at constant This result shows that the velocity distribution induced in a fluid by moving om fates and is shown in figure (2.2) velocity is linear across the BP between the two P! flow is also called a plane Couerte ft . AVERAGE , MAXIMUM , AND MINIMUM VELOCITIES The average velocity distribution of this flow is given by h 2 y 1 u -B|F u =1 fugy=a prey nt 2 tly 4) y 4) °VISCOUS HYCUMIFRESSIBLE Floy 84 oo. ¢ velocity distribution increases lin flow the ¢ of simple Couette value of U at the moving plae wat From equation (3) we see that in the cas maxim from the minimum value of zero atthe fixed plate to th average velocity is given by equation (4) to be U/2- Qe huyy= ng @ SHEARING STRESS: From equation (3) , we have sy “th and the shearing stress distribution of this flow is given by du_ du_ pu 6 Tyx = "by Bay "oh which remains constant across the passage between the plates . (i) GENERALISED COUETTE FLOW It is a simple Couette flow with non - zero pressure gradient . Therefore the boundary conditions are the same as in case (i). These are u=0 at y=0 and u=U ay=h ie, u(0) = 0, u(h) =U Using u(0) = 0, equation (2) gives B = 0 i dp y? therefore u= OG St Ay o Using u(h) = U, we get from equation (7) dph? dx viz uU= +Ah 1 dp 2p ax! of A Substituting the values or A= sic FIK f A and B , equation (2) takes the form (8) which is the equation for the velocity distribution of the Beneralized Couette fi low . Introducing the dimensionless pressure gradient by ru (-#) a--(-2 2nU \"dx PsscouS FLUID DYNAMICS 85 ation (8) bevomes uly x 4. teak(1-¥) «10 ) « velocity profiles for different values of 1h ( im 5 & are shown in figure (2.3) (a) or (b) >O lie. oe < oth i ‘ing ir for a dx <9) © Pressure is decreasing in the direction of flow , hence the fluid velocity is everywhere positive in the direction of x — axis over the entire width between the plates rw a <0( velocity becomes negative ( i.e. reverse flow occurs ) over a portion nearer to the fixed plate dp ) ap. fei dx > 0), the pressure is increasing in the direction of flow , and the fluid Figure (2.3) (a) or (b) shows that this occurs when a < —1, For a = 0 ( ie. the pressure gradient is zero an » the NK eqotion (10) becomes ), the velocity distribution is linear and thus uly ap uU"h This panicular case is known as simple Couette flow or simple shear flow v : AVERAGE , MAXIMUM , AND MINIMUM VELOCITIES The average velocity distribution ofthe generalised Couette flow between two parallel plates can becalculated by the expression h =y fuy “h 0 Substituting the value of u_ from equation (10), the above equati (12) ion becomesVISCOUS FLUID DYNAMics, 87 VOLUMETRIC FLOW RATE The volumetric low rate per unit width in the Q= buy = a(Ss +)u 2~ direction is given by (16) SHEARING STRESS The shearing stress distribution i piven by du_ du _ BU pau ape - ae 2 xT Hoy" Bay = at (1-2 2) a7 ‘The shearing stress at the lower plate is = HU wou wu Cy )ye0 = tp =F Uta) (18) The shearing stress at the upper plate is HU Hau yu (yen > ht =a) as Atthe position y = ¥ the Second term on the R.H.S. of equation (17) is zero , hence the shearing stress is independent of c at the centre of the passage . Recalling the sign convention of the shearing stress if (Ty , )y = is the shearing stress which the fluid exerts on the upper plate , then the shearing stress which the upper plate exerts on the fluid is ~(Tyx)yen- Similarly, if (Ty, )y.o is the shearing stress which the lower plate exerts on the fluid , then the shearing stress which the fluid exerts on the lower plate is ~(Tyx)y=9- Thus the sum of shearing stresses exerted 7 the fluid on the plates is _2nUa (ty dyen- (Cty dye0 Ya) -h* Mota) = i 20) and the sum of shearing stresses exerted by the plates on the fluid BU BU _ 2nUa (Tyxdyeon (Tye dyeh = CFA) He) = en Equations (20) and (21) are clearly in accordance with the requirements of static equilibrium (ii) PLANE POISEUILLE FLOW In cases (i) and (ii) we have discussed the flows between two parallel flat plates , one of which is fixed and the other moving with a uniform velocity U. Ifthe two parallel plates are both stationary , the fully — developed flow between the plates is generally referred to as plane Poiseuille flow . 5 and y = s as shown in figure (2.4) . The x — axis is along Let the plates be situated at y 2 “The fluid is kept in motion by a pressure gradient the lates . aeaicapenitt pressure gradient , there will be no flow. This implies that if there is no externalVISCOUS INCOMPRESSIBLE fq, 88 The boundary conditions in this case are: u = 0 at y = lel plates is given by From equation (2) , the velocity distribution for the flow between parallel p| 2d giving A = 0 and B--- 3 ag 2 and inereore uw = 2 $2[ 1 -4(2) ] 22) Equation (22) indicates that the velocity profile of the fully developed laminar flow between two paral! plates is parabolic . Thus if the pressure gradient , viscosity , and plate spacing are specified , then the velocity distribution can be determined . MAXIMUM AND AVERAGE VELOCITIES The maximum velocity of the plane Poiseuille flow can be calculated from equation (22) by letting y = 0, Thus - wae Umax ~~ y dx (23) ‘The average velocity of this flow is given by hi2 Vay (24)VISCOUS INCOMPRESg) 90 BLE FL, 2.4 FLOW BETWEEN TWO MOVING PARALLEL PLATES Consider the steady laminar flow of a viscous incompressible fluid between two infinite y ‘moving horizontal flat plates distance h apart. Let , the lower plate be moving with a velocity Uy and the upper plate with a velocity U. in the direction Parallel to the direction of flow as shown in % — figure (2.5) . The pressure gradient in the direction I . ap Figure (2.5) offlow is 7. We know that the velocity distribution for the flow between parallel plates is given by dpy? wae yt Aye ay ‘where the constants A and B are to be determined using the boundary conditions . The boundary condition in this case are : v= Ur at y= 0 and w= Uz a y=h. (ie u(0)=U,, u(h) = U3) The first boundary condition u(0) =U, gives B= U,, and the second boundary condition u(h) = Uz gives 1 dp h? - dp Ur ay TtAh+U, or ate Substituting the values of A and B in equation (1), the velocity distribution of this flow is given by 1 dp Ui-U) wT ax (yt -yh)+ y+U, (2) AVERAGE VELOCITY The average velocity of this flow is given by veh foo EE a jo I- El- a * Q)yisCOUS FLUID DYNAMics 1 VOLUMETRIC FLow RATE The volumetric Now rate per unit width ig Biven b Y Q=hu, ~L Deh Uisu, waxes idph (uj+yu yh udx 1% ry SHEARING STRESS The shearing stress distribution is given by =,dt.idp Tyx = Hay "2 dx(2y-h)+y () The shearing stress at the lower plate is U2-U, (ye)yeq = (SEM © and the shearing stress at the upper plate is ~ (Uzi ih dp = ( 2) (ydye @ 25 STEADY , LAMINAR FLOW OVER AN INCLINED PLATE Consider the steady flow of a viscous liquid over a wide flat plate inclined at an angle a with the horizontal under the influence of gravity . The depth of the liquid normal to the plate is h which remains constant . There is no velocity perpendicular to the plate and the pressure at free surface is constant . Assume that the air resistance is negligible so that the shearing stress at the free surface is zero . Since the flow over the plate is parallel and y occurring in the x — direction only , therefore , Y and w and their partial derivatives of all orders are . du zero. The continuity equation reduces to 5 = 0, which means that u is constant in the x - direction and all its derivatives of higher order are zero . Since no flow is occurring in the z ~ direction , u is Not a function of the z ~ coordinate , and the derivatives ofall orders w.r.t, z are zero. Furthermore , the flow Figure (2.6) ‘ a. is steady and the pressure is constant , therefore 57 = 0, op _ap_ ap i * equations for incompressible flow in this case =P LPL nce the general Navier ~ Stokes’ equation: 8x By = Fz 7 0+ Hence thew including the body force ( ie. gravitational force ) reduce to hau O= gsinaty ay?VISCOUS INCOMPRESSIBLE Flow 92 : . or «= SU Pgsina by Integrating w.rt. y, we get . ay au -Pgsing oy u Since the flow is everywhere parallel to the plate , By Oa ynh, This implies A = eee, sina Thus equation (1) becomes “ - an (h-y) Integrating again w.r.t. y, we get . 2 ease (ny-b) ip Q) Using u= 0 when y = 0 gives B = 0. Thus equation (2) becomes pe at a (2hy-y’) (3) which is the vey distribution of this flow . AVERAGE AND MAXIMUM VELOCITIES The average velocity is given by h in a -} Joay= 4 1 ne 0 fw y?)dy sin @ oth - Pgsina | hy 2ph 3 _ Pgh sing 3 (4) The maximum velocity which occurs atthe free surface_y h is obtained from equation (3) as Pgh?sina max = pare (5) From equations (4) and (5) , we get pgh'sina eS or 6)\iscoUS FLUID DYNAMICS 93 VOLUMETRIC FLOW RaTE The volumetric flow rate per unit width of the _ Pah’ sina su o SHEARING STRESS The shearing stress at the wall is given by -(,2 2 (Tyyeo (+53), ..-[e Th (ar-2y)] yno = peghsina (8) plate is given by Qa huay 16 STEADY , LAMINAR FLOW OF TWO IMMISCIBLE FLUIDS () LOWER PLATE FIXED AND UPPER PLATE MOVING Consider the steady laminar flow of two immiscible viscous incompressible fluids between two parallel plates distant h apart as shown in figure (2.7). The lower plate y = 0 js fixed and the upper plate moves at a constant velocity U to the right . The pressure gradient in the x — direction is zero ie the motion of the fluid is caused entirely by the movement of the upper plate . The lower half of the region between the plates (ic. Osys 4) is filled with fluid with density and viscosity 4, and teupper att (ie 8 . ‘The Reynolds number is used to classify the fluid flow as laminar or turbulent LIMITING CASES The Navier ~ Stokes’ equations have been simplified for two extreme cases : (When the Reynolds number is very small (i) When the Reynolds number is very lage (i) CASE OF VERY SMALL REYNOLDS NUMBER I the Reynolds number is very small (( Re ), << 1], then either the viscosity of the fidisvey large or the characteristic length of the body and the velocity of the fluid are very small viscous forces exceed the inenial forces and as a first approximation non Navier ~ Stokes’ equations can be neglected altogether called stow motions or creeping motions . Creeping mot "Navier Stokes’ equations in the limiting case of very small the Reynolds number represents the ratio of inertial to viscous The other important application of the |whrication which will not be discussed in this text Im this case the — linear terms in te The flows which satisfy these conditions ax ions can also be regarded as solutions of te Reynolds number i.e. [ ( Re Dy > 0) because forces . theory of slow motion is the hydrodynamic theory of Gi) CASE OF VERY LARGE REYNOLDS NUMBER Ifthe Reynolds number is very large | i.e, (Re), +2] then either the viscosity ofthe Milt very small or the velocity of the fluid is large . it : : From the point of view of practical applications tt limiting case is of far greater importance because the two most important fluids water and air , have 7 smal viscosities . Light oil and water are examples of uid with small viscosity , however, the fow!! air at low speeds is also included here : 2.32 THEORY OF SLOW MOTION STOKES’ EQUATIONS iinuity and the Navier - 1 , . wibby hese ue oN Stokes! equations forthe fow ofa viscous ineompresible v.V=0 w\scoUS FLUID DYNAMICS 173, inthe case of very slow motion or j Motion or in the case of very high 'Y Viscous flui i Re << I )._ In such cases the inenti ia small ( approximately Zero . Thus equations (1) and (2 term or cony, vine ) reduce yo VSStVE acceleration term (V0) Vi 5 Wet : nd By ~pVpryviy Q) (4) These equations are kn jown as Stokes? riled erezping flows. equations for very slow motion . Fl + Flows which satisfy equation (4) Equation (4) represents the following three seal lar equations (@) These i equations together wit with nuit ona the continuity equation (1) represent four scalar neo ations ii ee great simplification in Stokes’ equations is that thes ‘vations vance ly flow , Stokes’ equation (4) reduces to “eatin aero so Vp=yv?v tion (6) can be written in scalar form as ° op o a2 7 HV? w ff view as they are linear Te figs equations are considerably simple from mathematical point 0! as that of full Navier — Stokes’ ‘equations so with the full i at ty Nl equation Nae? Many bone Moreover , their order remain the same Si neaTY conditions may he satisfied with the Stokes! equations 2s €S' equations . that the velocity EQUATIONS J velocity field , pressure equations To obtain the equa! are separated SO ALTE, RNATIVE FORM OF STOKES" tion governing thi aagy,,_ |” the Mprese neat form the velocity and the ields may be established separately »VISCOUS INCOMPRESSip1 hy 9} 14 we proceed as follows . We know that w v.V=0 y @ av 1g oY ely ptev?V at ~pyYP Taking the curl of both sides of equation (2), we Bet v Vv 3 yx = tL oxvprovx(V?V) G) op 2 (vx) =ov?(VxV) Taking again the curl of both sides of equation (3) , we get 2(vxvx¥) = ov?(vxvx¥) @) But from vector analysis , we have viv = 0(V.V)-vx(VxV) = -vx(vxV) 6) From equations (4) and (5) , the pressure free form of Navier — Stokes’ equation becomes -2(9°9) = -992(v79) av = or vot = ov © To obtain the equation governing the pressure , take the divergence of both sides of equation (2), weet! ev_4 = veo ~pVeVptov.v? v 1 aos or ail¥-V)=-F¥%pry92(y. 7) Using continuity equation (1), equation (7) becomes Vp=0 ” / (8) : ise, for very slow motion the pressure p satisfies Laplace's equation and is therefore a harmonic fu” refore 2.33 STEADY FLOW PAST A spy This problem was first solved by st ae (STOKES: FLow) we ea “® ind i often referred to as Stokes” flow or Stoke A . fh i - ‘ direction of the z — axis . Let the centre of the sphere — Stream U flowing steadily 8 the ae be in the direction of the uniform st the origin of the coordinate system» Lt? agi a alt ream fay Point from the origin of coordinates, -*Y Om the sphere and tbe the disance of”yisCOUS FLUID DYNAMICS 235 SOLUTION: Since [j~ = 1, therefore the displacement thickness is given by s 1 1 bp = 8 S(r-t)ans Ja-nan 0 0 toate 1 |n-3n |, 9(1-4) = 48 The momentum thickness is given by bu = 0 ‘The energy thickness is given by 1 1 be = 8 fe [-(&) Jon =8 fncienyan 0 0 = 65 Jorn yan = 0 3.20 BOUNDARY LAYER EQUATIONS FOR TWO - DIMENSIONAL INCOMPRESSIBLE FLOW We know that the flow of a real fluid is fully described by Navier ~ Stokes’ equations . Due to the Complexity ofthese equations , exact solutions can be found only in few cases . Therefore there is a strong need to make some simplifying approximations to the full equations and hope that these can be solved In many cases it is extremely difficult even to determine approximations . Prandtl simplified the Navier~ Stokes’ equations to yield approximate solutions for flows with very small viscosity or very high Reynolds number. For simplicity , we consider the two dimensional flow of an incompressible fluid ¥ound a body which has a smooth boundary. The result of his analysis, the boundary layer equations ‘re simpler to solve than the Navier ~ Stokes’ equations . , We first consider the two ~ dimensional flow over a plane surface Coincident with the direction of the oncoming flow. Choose the x — axis of Wall ofthe body and the y ~ axis normal to it as shown in figure (3,28) . Cie, flat wall ) not necessarily the coordinate system along the23 BOUNDARY LAYER THEORY 6 Figure (3.28) ‘The Navier — Stokes equations for two dimensional incompressible flow without external forces are ; Bue y2u, 28. 120, (224,28) W at Yax* Yay ~~p ax *\axt tay? Bry yBtay2e e122, (24,2%4) > at *Yax* ay ~~p ay**Laxttay? @ and continuity equation is : au ov ax*ay ~° 6) ‘We know thar the fluid velocity in the boundary layer changes from u = 0 at y = 0 tou = Us(x,t) aty=6. The mathematical procedure for reducing the Navier — that of finding the first order approximation of the form very large . We rewrite equations (1) , (2) and (3) int the free stream velocity U Stokes equations to the boundary layer equations is fer equations when the Reynolds number (Re). he dimensionless form , by referring all velocities © sand by referring all linear dimensions to a characteristic length L of the wall, We define the dimensionless variables as follows : veh y yViRey =p ye. : U; pu ‘Substituting these into equations (1) and (2) , we get ' 2 , Us ou! Us au! US pau! (eu =. ae (2U,)arw fou (es) 2 Day h Mox bay pT Jee (sus), (evscnen)s L oy ’ 2 , Up_av, Us Us ay LyY(Re), ar LYCRe), * ax LViREY, Yay! 2 : UsV(Re) ap’ gy zy Vere, 2, -- (Year op sav ou vt a + ay“ L?V(Rey, 92>VISCOUS FLUID DYNAMICS 237 ; Ugh Since the Reynolds number (Re), = 3» we can write these equations as ar, Us (au! | L \ar u; ( oie (2%, Ly(Re). Lar 2 Us Op’ _v atv’ v(Re), atv’ -—Us_|_.., ae (Re), L fx | (Ree TL agt? Ugh ay” © >, * ale ou yOu youl op aty ty fat ax ay ay TR gy ot i vt (5) {=P 1 av, ) dy’ (Rey ax” (RED ay The continuity equation in dimensionless form 0 (6) ax © Inthe Prandtl limit as (Ke ), -» <0, equations (4) and (5) simplify to "ot cau apt ate! au yea yOu. OP aul m ax ay Ox ay md 9 = 22 (8) ey Equations (6) , (7) , and (8) constitute the boundary layer equations , or more precisely , the first approximation to them as ( Re), -> 2. We now return to the physical ( i.e. dimensional ) variables , therefore , equations (7) , (8), and (6) respectively take the form cet Sy? 0 = 2P (10) ata = 0 a The boundary conditions becom: =0; u-v=0, Aty=0; uev =o, Hag, Bee a2) Aty = B(ora); ur Us(xit)y VO Gy= 0, Bye Equations (9) to (12) are generally known as Prandil!’s boundary layer equations . Let y s(X,t) be the velocity in the potential flow ( just outside the boundary layer ) , then at the outer Ou ‘dge of the boundary layer , equation (9) becomes (snc dy" Y"O ays 8)238 BOUNDARY LAYER THEORy 2Ue 4 y 20s 1 oP (13) ot Sax "pdx Using equation (13) , we have from equation (9) 2 Ou, du, du _ OU, BU, aw ) UR toy nts an tay! NOTE: (i) Although the boundary layer equations (9) to (11) are derived for a plane surface , they als hold for a curved surface provided that the curvature does not change abruptly as would be the case with sharp edges (i.e. comers). (ii) From equation (10) we conclude tnat the pressure in a direction normal to the boundary layer is constant ; it may be assumed equal to that at the outer edge of the boundary layer where its value is deternained by the inviscid ( potential ) flow. The pressure is said to be impressed on the boundary layer by the outer flow . It may , therefore , be regarded as a known function as far as boundary layer flow is concerned , and it depends only on the coordinate x, and on time t. As contrary to the case of slow motion , we see that the non ~ linear character of the Navier Stokes’ equations has been preserved . Furthermore , equation (2) i.e. the equation of motion normal to the wall has been dropped completely , and one viscous term in the remaining equation (9) or (14) has also been eliminated . Since the pressure p can be evaluated from the potential flow solution with the aid of equation (13) , hence the boundary layer equations is a system of two simultaneous equations for the two unknowns u and v, instead of the former three variables u,v and p in the Navier — Stokes equations. BOUNDARY LAYER EQUATIONS FOR STEADY FLOW In case of steady flow , the equations (9) to (12) reduce to Stu 1 OP, aty ox” ‘ay ~~ p dx* ay (1s) op ay"? (16) au lav y dx" dy ~ with boundary conditions ; Aty=0; u=v=0 a7) Aty = 6(or®); w= U,(x), 8.0] as) y U, (x) isthe known (prescribed ) potential flow at the ou Equation (16) indicates thatthe pressure p depends ont the boundary layer . Thus from equation (15) ter edge of the boundary layer . yon x i.e, fora given x, p is constant through »wehaveat y = § dUs_ ide sax ~~pax or p+ 4p U; = constant —( Bemoullis equation ) Equation (15) can also be written as uy due y We, atu u5x* Yay Usdx *%ay? 20)BOUNDARY LayER 1 FoR 248 The average drag coefficient is given by 1.328 _ 1.328 _ = ae ey = 0.00171 ViRe), The total drag force on both sides of the plate is given by 2 2 5696.8 L 28 ) (1.25) (860) (3)*(2) _ 25696.8 _ beust (13 a 775 33.2N 2D = 1328 Re), 3.23. FLOW PAST A WEDGE We have discussed the boundary layer along a flat plate with zero pressure gradient . Now we discuss the boundary layer on a surface with pressure gradient . Consider the steady flow past a wedge of angle m as shown in figure (3,31) . The stagnation point of the wedge is taken as the origin O of a coordinate system whose x ~ axis is along the wall Of the wedge and y ~ axis perpendicular to the wall. In this case the velocity of Figure (3.31) the potential flow (i.e. outside the boundary layer ) is proportional to a power of distance along the wall ie, Us(x) = Uyx™ a) where U, is constant. ‘Now introduce the stream function y such that y = & vs = y’ x The boundary layer equation in this case is : éy @y dyaty aU, dy —_ -—> = — dy Oxdy~ Ox dy? ~ Us at -9 TF @) with boundary conditions éy y . ox By 7PM y=0 @G) and oy By ~~Us(X) at y= Equation (2) can be solved if we introduce the variable m+ Us m+i fu mat ney Ve my aS xCH) « Let ou = Usf'(n) (5)VISCOUS FLUID DYNAMICS 249 The stream function w is obtained by integrating the u velocity component as vr-fudy = -Jusromay From equation (4), we can write 2 -(%1 yo Nomi, CH) sothat dy = Vein CP Yon ‘Therefore equation (6) becomes -fu, Vin CCPMYeinan - we CF) Jronen faa CH) fn) ” The derivatives of y wart. x and y are: (6) Also x). 2(2)m an Lay) ay =-u CONTE = -/%} tuLy a en) ay 2a Uj(m+l a ee + RAH) Pen250 BOUNDARY LAYER THEORy yeu, al Yereny- yu, (25+ Jet x tr Ve f(n)- Vain fina aS ra - = , - Vr \ Pe ron-( "5) Vine x TF (n) ey e/ay)_ 2 me axby "5 34) 3x [-Uix™F(n)] . ” én == Uymx™ 'F(n)-Urx™ fn) ZY ” 'U. = -Uimx On) Uae Oda JE (3 “ 1 U eel m =-vime'Co-[ yf fExl ? de (a+ = -Urmx"™'f'(n)- vist) taeton =m ony -(2) Yay Substituting the values of these relations in equation (2), we get , Uy, =1\U » -u onl -m 6 (-(234) Bae (n | NU QU, . [= 4 Us sro fy (82) We ato] \PHun lee «| U uUeiw, v eee (-(8)Y f(n)f" Cn) 2 (AE) S aeae ony = me, (at) een) or mf*(n)- (CE)ame (1) = mo(2 Ne) orf" (ny +FCn) En) - read (f"'(n)-1] =0 C#) xmatgnt f (n) (8) The boundary conditions become : f(n)=f'(n) = 0 a n=0 | fina at n= @ oVISCOUS FLUID DYNAMICS 251 Equation (8) was first deduced and solved rumerically for various values of m — by VM. Falkner and S.W. Skan . Subsequently, solutions of greater accuracy were given by DR. Hartree (1937) and the velocity profiles for i * tin) several values of m are shown in figure (3.32) . . ° For m > 0 (B > 0), equation (8) has a unique solution , while for m < 0(B <0), the solutions are not unique . For m = -0.0904 (B = ~0.199), f(n) = 0 at n= 0 sothat (3) y=0 ne (cm [T, for all x and separation of flow will occur . This is FAR the smallest value of m for which a complete Figure (3.32) numerical solution is available PARTICULAR CASES @ For m = 0 (8 = 0+), the velocity of the potential flow U(x ) is constant , say U,(x) = U, = Ug and equation (8) reduces to f"'(n)+£Cn)£"(n) = 0 (19) This is the case of the flow over a flat plate at zero incidence . Equation (10) differs by a factor 2 in the ” i - U first term from the equation 2 f° ()+f(m)f (1) = 0 which was obtained by taking ae fz Ifthis value of n is replaced by n = \/5y'. then 26" (n)+ fC)" (nm) = 0 will take the form f(y) +f(n)F'(n) = 0. (i) For m = 1(f = 1), equation (8) becomes 2A" (nye tenye" (n)-f (nytt = 0 ‘and the wedge angle becomes 1. This is the case of the flow in the vicinity of a stagnation point in the Plane , which was considered earlier . 324 FLOW IN A CONVERGENT CHANNEL Consider the flow in a convergent channel With flat walls ( sink ) as shown in figure (3.33) . The velocity of the potential flow in this case is given by Uo =-4 ay where U,(>0) isa constant. Figure (3.33)258 BOUNDARY LAYER THEQRy 3.26 THE ENERGY INTEGRAL EQUATION ____ Consider the Prandtl's boundary layer equations slong discussed in the derivation of momentum integral equation .. We kn with boundary conditions which were .ow that the first of these equations con be written as du, au OU, bau ax* Yay ~ Us ax tp ay? Multiplying this equation by u and integrating w.r.t. y across the boundary layer from 0 to 6, we get 8 é du aU f au -£f au, a fo(udt-u,2Us dy+ uvaydy p s%aytt® 0 0 0 Consider the second integral 5 8 foray - fSGe)ev “3 vywey 6 J 0 0 0 6 J O < 2 5 vA (er -Us ay 2 au by subtracting the zero quantity oy. from the integrand . 6 v(utu [3 J (wu?) Say 0 The integrated term is zero at both limits since v = 0 when y = 0 and u = U,aty = 5. 5 6 ou 1 2_y?) dv Therefore frugtay = =3 fc -Us)5yay i) 0 5 2 Likewise J u St dy 0VISCOUS FLUID DYNAMICS 259 the integrated term being zero at both oars Using equations (2) and (3) in equation (1) , we get 6 8 Jo(o8-vsQe)ay+ 3 foo-ub ay -B f(S8) tte ony -2, we get ° 8 -fos(ott gy Bday Joo -ui)tay = 2# f(32) ay i) or or The upper limit of integration could be replaced by y = 2, because the integrands become zero outside the boundary layer. Therefore the above equation can be written as 8 au? or # (vise) -3# (3) dy (4) 0 which is usually referred to as the kinetic — energy integral equation . NOTE: This equation is applicable to laminar flow , but can also be applied to turbulent flow . Inturbulent flows , this equation assumes the form 8 (vise) = 2ft (Se )ay 0 8 d 2D ou « § (ule) = 2P wtore = fx (5f)ay 0 327 APPLICATION OF MOMENTUM INTEGRAL EQUATION FLOW OVER A FLAT PLATE We now use the momentum integral equation to calculate the boundary layer over a flat plate at E10 incidence . Assuming that the velocity profiles are similar we can writeINCOMPRESSIBLE TURD, 326 ‘Te unit of fiton velocity Is 3 fm em m/e m/s that u* isnot an actual velocity, itis only quantity that aM honey. we Note Turtle ow asi 417 PRANDTL MIXING LENGTH THEORY ot Prandtl in 1925 developed the necessary theory a sara shear Se 19 veloiy gradient through the boundary layer This theory isk 7 ai mining eng got ‘The mixing length theory may be applied to two ~ dimensional flows such as pipe flow “ f flat plate , free jet and wake flows. In feet some calculations have been successful ting sy length concept . It remains , however , an empirical concept and does little to explain the funda, Mis turbulent flow . em Since turbulent flow involves transverse motion across the layer ( i.e, in the y direc mixing of the fluid particles in that direction , therefore there is a momentum transfer between a layers of different mean velocities . i Inthe Kinetic theory of gases one molecule before striking anther, travels an avrg kaon the men fee path ofthe gas «Using this analogy , Prandtl ssumed that a ign displaced by the Nuctuation by a length ¢, perpendicular tothe main flow direction, before it re is changed by a new environment . Thus the mixing length ¢ is defined as that distance in vie paricle tavels from a region of one velocity to another region of a differen velocity ini inet neighbourhood . iy Consider the velocity profile as shown in figure (4.2) where two layers of fluid at a distance ¢ ‘apart move with different mean velocities U and +AU. The element which was initially at y and had shal y, | C sy iy the velocity Wis now at y + € and has a higher Velocity + AW. The velocity difference is a measure ‘of the fluctuation velocity in the x ~ direction : W = AU = T(y+e)-(y) Figure (42) If U(y+0) is expanded ina Taylor series upto the linear term only , we obtain vest Beye acy) . ‘? Prandtl further assumed that the fluctuating velocity components u' and v' are of the same order ‘magnitude and that each one of them is proportional to velocity difference Aj. Thus heya pl vada ctrp DYNAMICS a ys Fl se of some adhoc assumptions and physical reasoning , Py with oN turbulent shea stress and the velocity gradient in the + Prandtl suggested the relat yee” RY) Nonship 1 a ptt au) ve ace oy w sean” inthe form we arial 2a ai PF epuv 7 P8 Joy! oy @ eso ve | ai is inroduced to give the shear stress T and ihe velocity gradient the using equation (2), the total shear stress can be written as sg a se |] 8 ») Zivot Joy i y-KARMAN SIMILARITY HYPOTHESIS " aner considering similarity sonditions in a turbulent fuid » suggested thatthe me ov on duldy @) ee Kl paiay’ kK ism empirical constant and has an experimental value 0.4. length € depends only on the velocity distribution . atin that the mixing he turblet shear stress in this case becomes . a (28 iw , (ould . (3i)'- 9 iia we (gra109"Y 7 (aasay?) INTERNAL FLOWS Fluid flows can be divided into wo categories : External flows and internal flows « External flows are the flows around objects immersed in an unbounded fluid and have been Internal flows are the flows completely bounded by the solid surfaces such as turbulent . Except for discussed in chapler qi). flows in pipes » ducts » nozzles and diffusers « Internal flows may be laminar or mal flows of engineering interest are generally vary viscous fds in small meter pipes , intel i turbulent iciples to the flow of vit 418 Wepow apply the basi prin scous incompressible fluids in pipes and ducts . ‘The flow of a fluid ( liquid or gas ) in a closed conduit ( commonly called pipe of it is of circular ros ~ ten ot ‘a duct if it is not circular ) is extremely important because of their wide engineering wis. We se that there is @ wide variety of applications of pipe flow « Such applications range riba hundreds of miles long pipeline that carries crude oil ( or gas ) to the natural systems of | cary blood throughout our body and air into and out of our lungs « Other examples include the2 —= es INVISCID COMPRESsip, LEFLoy, 5.7 FUNCTION OF STATE Let @ be some thermodynamic property (function ) of gas expressible in tems of any yy the state variables p, @,and T. This is always possible if the equation of state of a gas is known ed changing from an initial state A to the final state B , the value of @ depends only on the conditions snd B and not on the paths joining them , we say that @ is a function ofthe state . For example y4 specific enthalpy H_and the interal energy U are the examples of functions of state, sinee i] 2 =H-= H=U+t> or U H-5 5.8 RELATIONSHIPS FOR GASES REAL GAS Incase of a real gas , the specific internal energy U of a simple system depends upon V, and T ie U = U(Vs,T) ou, au a= (Fe) ave (52), aT a) s From the first law of Thermodynamics , 5Q = dU+pdV, : (28) av,-(2) = (2) av,+ dT+pdV, @) (sy, ee Nay, 5 If Vg is kept constant, d V, = 0 and we have from equation (2) 5 au «= (59) = (5) ® wer), If p is kept constant , we can write equation (2) as 8Q_ au (2) [avs at ~ (at), *| P* lav), fat vy T 8a) . (24) au) 7(2% 7 dt) ~\at), *} P*lav, aT ? Vs T Pp where the conversion to partial derivative is based on the assumption that Vg is a function of Thus the above equation can be written as cy ef ( 4) 108), ' Note that C,, is greater than Cy. U andT. Next assume that the specific enthalpy H depends upon T and p ie H=H(p,T) a therefore, dH - (2 ou 38), 4° ar),a7 (3)viscous FLUID DYNAMICS amo H=USPYS pene a = oUt pAvSANSEP = 8Q+V,dP a «= 8Q = SNMP ‘ (3). an+($4), éT-Vsdp * © Ifpis kept constant , ap 0 and rom equation (0), we me (8), “2 @)1@ * (3 - (5), (3) ell (3), ot pia differentiation is used secause p isa funct Mm op hen ion of H and T. Thus the above equation becomes c-eo[()_-“(H), ® IDEAL GAS For an ideal gas ,Joule’s experiment showed that ‘au (e),-° This implies that U = U(T) ie. the internal energy U is a function of the temperature alone - Furhermore, H_ = U+P Vs 9) ‘ = U+RT (since pVs = RT) (10) Since U depends only on T .H is also a function of the temperature alone , ie H=H(T) ql Now cy = (2U au cy 3), = FF [sing equation (9) ] & dU=C,dT . (12) (28) 4H 7 > Siaitarty, Cy é), = Fy [using equation (11) ] © dH=C,dT (13) .‘388, ESOS COMPRESS From equation (10) , we have dH -dU+R4T or = Cyd T = Cd T+RET [esing equations (12) 206 03] o = Cy-Cy = R Equation (14) shows that the gas constant R is Givectly proportional t2 Se SGemer of me oes heats. Now integrating equations (12) and (13). ¥= 8% T u(t) = Jovat=er ° T and H(T) = Jo,ar= eT ° where Cp and C, are assumed constant for an ical £88 From equations (15) and (16) the changes in internal energy ant specific costes Seoesen Ne see and (2) are given respectively by . U,V, = AU CCT TY) £ and Hy Hy e AN OAT = Ty) up NOTE: (1) Equation (14) can also be derived from relations (2) oF (8) oF roel SESS For example , equation (A) reduves to vy Cyn Cy on *) om (# a . eyen(8) (aie pV, RD) This implies that Cy Cy = 8 (WH) For tiquits aind solids Cy and Care very mow equal Poe saane dead how arwdions oe ® weed even though there may be pressure variation in the Mit (Wt) Foran ileal gan the anaterial derivative Of apeviie ternal emega ant qavolle eens oe written fio equations (18) ant (LO) above ax bul. pt De es py bu vt wt Sein FNANIFLE (A Compute Une change UW internal eweegy aut qaidayas af kat kg oe Motte UE Hes Geamper ature f inereavedd Mom TNA ee ake [tate Oy SAR Rae a RDta (Prive) *5g(P ¥2) = 0 a py a Lowy evd tt =o inuit a ar ‘this is the equation of continuity for three — dimensional axially a in — symmetric yindeial polar coordinates Me Mealy compres fa ¢ SPHERICAL POLAR FORM The equation of continuity for three ~ dimensional steady inviscid compressible ow j in spherical poarcordinates i: 18 opty.)4 te 2 (psindvg)+—te & j sg (PY) trend 36 0) + Taine 3g (P ¥4) = 0 0 . a , ror axially ~ symmetric flow, 35 (P Vg) = 0, thus equation (12) reduces to 3 14, . _ AZ (pve) + rein 6 (P8086) = é 18 2pyv cot 6 or gr(P Ye) * 799 (P Ve) * 7 tee = 0 (13) ‘his isthe equation of continuity for three ~ dimensional axially — symmetric steady compressible flow in ical polar coordinates . “se Tipieieo Beewwlte mate pg cut la val i Preshe 521. “BULER’S EQUATION OF MOTION Umnet be F102 Euler's equation of motion is valid for both inviscid compressible and inviscid incompressible flows, Here p= Density, t= Time, p = Pressure VECTOR FORM If V isthe velocity vector and F the body force , then Euler's equation of motion for unsteady Compressible or incompressible flow Wiese ve ar t(V.V)V = (1)406 INVISCID COMPRESSIBLE Foy, In terms of material derivative , equation (1) may be written aS DV su a) Dt “pve i to In case of steady compressible or incompressible flow . ‘equation (1) reduces = oe e By (V.9)0 =F -5vp ) CARTESIAN FORM t If u,v, and w are the velocities and Fx. F y and F , the body forces inthe x,y, ang, directions respectively , then equations (1) and (1A) may be written as au, du, ou, au. p, oP attYax* Yay’ az * pox ov, Ou, Ov, ye. py, bP @ att Yax*Yay* az % POY aw, 2,2, 2% p 1 oP at tax **ay “az * poz In terms of material derivative , equations (2) may be written as Du, 122 Dt~ "*~p ax Dv, 1p 2A) Dt” Fy-p ay (2a) Dwl, 12P Dt 2p az CARTESIAN TENSOR FORM If uj is the velocity and F; the body force in the xj — direction , then Euler's equations of motion (2) can be written as oy, OU at Miax; @) In terms of material derivative , the above equation can be written as Du 12 Duy _, 1 2p iiss C¥LINDERICAL POLAR FORM If v,,¥g,and vz are the velocities and F,,F9,and F, the body forces in the r,0,and directions respectively , then equations (2) may be written as av,, Ov, ve av, Ov, v Ove, OMe ve artYrar*r ae*'! O27” av OVe V9 OY, a 2¥o,, Ove, Ye Ave Va VeVe 1 aut art a0 tt art Fem @) ov, | av a —_ OVz Va OVE ov ant artr oot? oz = F,-409 ges. sco! ous FLUID DYNAM joes of material derivatives , these equations can be written as va py, Yee 1p Bee Pep or Dvp Ye =F Le we r e- pro _ Dy 12 pr Fp a2 p. Bay 2+ O42 whee Dt pt iar™ 7 30° 702 incase of steady compressible or incompressible flow, the corresponding equations can be written from i= equations () by takin a sPHERICAL POLAR FORM If vy,¥o-and YO are the velocities and F,,Fo. 2nd Fe the body forces in the r,@,and, equations (2) may B® written as 4 directions respectively» then ve OMe vit¥) - p 12P Fi-por ave, y Oe wae ve St art Yt oF +7snd 0d av ons vg Ve Ov, Yeve_ weenie 128 5 Setvear F 00 rn a — Fo~prae ° ave avg, 2, vg Ove Mery ns. Fe-= 128 30 *rsin@ 6 r e-prsin6 06 +, a or equations may be written as Interms of material cata the above Dv ney = Fro por Dr Dvy Dt Bos, vave, vovasct 1 oP F r Fe-prsind 06 (A) D_2 Dia, o,%08 Dene 88 ve 2 att ’rar* r 30 rsind 36 sponding equations can be written from Inease of ady —— ‘or incompressible flow » the ct ()by taking 2 = 0. NOTE: E: For gases ‘ation of moti + the external body force is usuall F i ly very small , therefore F can be neglected in Euler's412 INVISCID COMPRESSIBLE Foy, Therefore dp = Kyp""'dp I-Ie ky Jov? ap er ” = Kye yal 1 8) . yt 2 But from equation (6),K 9” = 5 From equations (7) and (8) , we have ig Thus from equations (1) and (9), we have o) 1 +2. 10) ZV7+2*T Tp c (10) In the absence of body forces , equation (10) reduces 10 ty2,1 2. in 2V*771p © an For any two points on a streamline , we have 1 2 y (P2_P fowogh(es)° 2. BBs). vi-v?) Z 7 “ at pi) n2VinVa) «2 524 BERNOULLI’S EQUATION FOR IRROTATIONAL INVISCID FLOW We know that the Bemoull’s equation for irrotational inviscid compressible fo" conservative fores i -2t,1 Lyeas f[B- rey wry! 4 where £(1) is an arbitrary function of time t. The function f(t) at a given instant baste 2% throughout the entre Now Feld. Equation (13) hold for both the incompressible and compress" ISOTHERMAL PROCESS For an isothermal process , we have from equation (2) J2-xinp ? Thus equation (13) becomes oe ~Bedvtsaek inp = (0) a4)atita() = tase ¥x(vxv Jeo : _ as a(ueZ)-ras = Vx(vx¥ 4" 0 From the energy equation , we have vy? a(n+¥) =0 thus equation (7) reduces to or tas = (VxV)xV.dr 8) Since V is parallel to dF along a streamline , and (vx V )x V is perpendicular to soi Ow that (VxV)x¥.d¥ = 0, alonga streamline. Thus from equation (8), TdS = 0 9) from which it follows thatthe entropy i constant along a given streamline - Hovever,entopy ny, from streamline to streamline . go? " 531 KELVIN'STHEOREM qt 4 apt of) Change ” Kelvin’s theorem weno cireulacion ale® LAred BV a7) = a0-P+V.4¥ UMW nei () Integrating the expression in (1) along the closed curve C, we get Diy at dp liso $ P(V.a7)= $ an- $ Bo $ 3a(V NA) c tor C ce eo : br Lee ay Pp Dt $ [od v.v)}$ 7 c & = i dp ~ $ alow ]- $2 @ c c Since ©, V in the integrand in the term on R.H.S. of equation (2) are single valued , it must be zero Therefore equation (2) becomes pt? 6) c| sgoUS FLUID DYNAMICS 421 orabaratropic fluid , we have from Stokes’ theorem pdt vp = ge. $2 at _ § Wiaz ~ § o(2).07 p p e p c c c c 7 JJvxe($).tas -0 a 5 Zar 1d ayo, foran isothermal flow TAS = dH-"” pati op Suthep « Gn an-Tds Gaver Integrating both sides along a closed curve C, we get JHe ba gai 2 gard § 2 - $ (antes) Ane gd ° - 21d 7°R c c (4) = |H|c-T|Slc = 0 since H and S are single - valued functions of T. Hence the rate of change of circulation along a closed curve is zero for both baratropic [p = p(p)] and 4 4 _ 4 for an isentropic fluid . This result is fsathermal fluids . Equation (4) also shows that § c ‘pected as an isentropic process is a special case of the baratropic process . 532, RELATIONSHIP BETWEEN VELOCITY AND PRESSURE FOR AN ISENTROPIC FLOW ALONG A STREAMLINE ‘The energy equation in terms of enthalpy is 1 H+ V? = constant na streamline , and H2 and Vz Hy andy 1 are the values of the enthalpy and velocity at one point o treamline , then energy equation Wthe values of the enthalpy and velocity at a second point on the same s ‘Written as 1 Hisgvi = Hi+gVi * ytya 2-Vi = 2(Hi-H2) = 2(C,Ti-CpTz) -2er[1-(P)] a (since H = CpT)Now FRE (tex zQ sus) RB? ‘re stagnation density is given by y-1 Wa) Pom p[stztma] intts- 7 cs.739[ 1 +435 (08s? ] ey = (8.73) [140.15 (0.55)? ] N°) ygnation density can also be found from _ Po _ 604983 Po RT.” (189) (317) = 10-Lkg/m? = (8.73) (1.045) 2 ” VON ker me ‘The sta suf RELATIONSHIPS FOR MACH NUMBER We know that Toe yy te Ma? qs 1+°3"Ma Mat = (Ay) (#-1) . a) Also, we know that P. -1 y(y-1) Be +t | (r-1)/y P -1 " (%) = 14 Ma? mw LG] @ Finally , we know that [1 y-1 er p + Ma? an f p wo {Po 1 (4) = 14 Ma? | ik Gy[(B)"'] 6) oeINVISCID COMPRESSIBLE FLD 442 Equations (1) , (2) , and (3) are the relationships for the Mach number in terms of temperature, press: and density ratios respectively . For air (y = 1.4), equations (1) , (2), and (3) become mes). oY], sf) Sa MACH NUMBERS LIMIT FOR INCOMPRESSIBLE FLOW We know that for an incompressible flow 1 Po=pt+zpv? Pi pv? 2 fo py OV v or 1+ = (since Bt RT)UID DYNAMICS tor yisCOUS FL! op and 92 gp od ide op dz rap or yp dz Integrating wa r= WE BEE dv, OP ror 2p d2 dv, 1 9P,A hs a are war 2p eee (4) Efe aint B The first Integrating again, V2 = Gy dz where the arbitrary constants A and B are to be determined from the boundary conditions - from the symmetry of the flow which requires that v must be finite on the ev, would become boundary condition is found = 0). It follows that we must take ‘A = 0. because otherwis axis of the pipe (T (4) reduces 10 infinite at r = 0. Thus equation 2 dp vet paz? () ‘The second boundary condition isthe no~ siP condition at the wall ies V2 70 at r=R (6) With this boundary condition, the constant B is obtained from equation (5) as pe eRe ap dz 2) Hagen — Poiseuille ‘equation (5), we get the axial velocity distribution of x Substituting the value of B in flow through a pipe as Rid al (f ? v.7-77a lI lR eng teal J] & ° which has the form of a paraboloid of i ity di ee inemsiene, of revolution . Note that the velocity distribution depends upon the MAXIMUM AND AVERAGE VELOCITIES The maximum velocity in this . a sins mina ty in this case oceurs at the centre of t he pipe where r = 0. This can be dp where gz <0 0)