* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Ch2Fall2012
Routhian mechanics wikipedia , lookup
Newton's laws of motion wikipedia , lookup
Biofluid dynamics wikipedia , lookup
Centripetal force wikipedia , lookup
Equations of motion wikipedia , lookup
Classical central-force problem wikipedia , lookup
Reynolds number wikipedia , lookup
Rigid body dynamics wikipedia , lookup
058:0160 Jianming Yang Fall 2012 Chapter 2 1 Chapter 2: Pressure Distribution in a Fluid 1 Pressure and pressure gradient In fluid statics, as well as in fluid dynamics, the forces acting on a portion of fluid, control volume (C.V.), bounded by a control surface (C.S.) are of two kinds: body forces and surface forces. Body Forces: act on the entire body of the fluid (force per unit volume). Surface Forces: act at the C.S. and are due to the surrounding medium (force per unit area: stress). In general the surface forces can be resolved into two components: one normal and one tangential to the surface. http://www.biomedsearch.com/attachments/displa y/00/19/77/26/19772652/1743-8454-6-12-1.jpg 058:0160 Jianming Yang Fall 2012 Considering a cubical fluid element, we see that the stress in a moving fluid comprises a 2nd order tensor. Since, by definition, a fluid cannot withstand a shear stress without moving (deformation), a stationary fluid must necessarily be completely free of shear stress (σij=0, i ≠ j). The only non-zero stress is the normal stress, which is referred to as pressure: 𝑝 = −𝜎𝑥𝑥 = −𝜎𝑦𝑦 = −𝜎𝑧𝑧 , taken positive for compression by common convention. Chapter 2 2 058:0160 Jianming Yang Chapter 2 3 Fall 2012 (one value at a point, independent of direction, p is a scalar) n p x = p y = p z = pn = p i.e. normal stress (pressure) is isotropic. This can be easily seen by considering the equilibrium of a wedge shaped fluid element Force balance: Fx 0 pxbz pnbs sin Fz 0 pz bx pnbs cos 12 gbxz From geometry of the wedge s sin z s cos x therefore p x pn pz pn 12 gz In the limit as the fluid wedge shrinks to a point z 0 px pz pn p 058:0160 Jianming Yang Chapter 2 4 Fall 2012 Note: For a fluid in motion, the normal stress is different on each face and not equal to p. σxx ≠ σyy ≠ σzz ≠ -p By convention p is defined as the average of the normal stresses p 1 1 xx yy zz ii 3 3 The fluid element experiences a force on it as a result of the fluid pressure distribution if it varies spatially. Consider the net force in the x direction due to p(x,t). p p dFx p dy dz p dx dy dz dx dy dz x x The result will be similar for dFy and dFz; consequently, we conclude: p p p dFpress i j k dxdydz x y z Or: f press p is the force per unit volume due to p(x,t). Note: if p=constant, f press 0 . 058:0160 Jianming Yang Chapter 2 5 Fall 2012 2 Equilibrium of a fluid element Consider now a fluid element which is acted upon by both surface forces and a body force due to gravity dFgrav gdxdydz or fgrav g (per unit volume) Application of Newton’s law yields: ma F dxdydza f dxdydz a f fbody fsurface (per unit volume) f body g gk g gk fsurface f press f visc p 2S for incompressible Newtonian fluid flow, S is the stain rate tensor a g inertial gravity p pressure gradient 2S viscous and V a V V t This is called the Navier-Stokes equation and will be discussed further in Chapter 4. 058:0160 Jianming Yang Chapter 2 6 Fall 2012 Consider solving the N-S equation for p when 𝐚 and 𝐕 are known. p g a 2S B(x, t ) This is simply a first order p.d.e. for p and can be solved readily. For the general case (𝐕 and p unknown), one must solve the N-S and continuity equations, which is a formidable task since the N-S equations are a system of 2nd order nonlinear p.d.e.’s. We now consider the following special cases : 1) Hydrostatics (𝐚 = 0 and 𝐕 = 0 ⇒ 𝐒 = 0) 1 2) Rigid body translation or rotation (no relative motion: 𝐒 = [∇𝐕 + (∇𝐕)T ] = 0) 2 3) Irrotational motion ( V 0 ) If viscous effects are neglected, the incompressible Navier-Stokes equation becomes incompressible Euler equation: V V V g p t also, V 0 V and if = constant, 2 V 0 0 2 p V V V g p gz 0 2 t t 2 p gz const t 2 (Bernoulli’s equation) 058:0160 Jianming Yang Chapter 2 7 Fall 2012 3 Case (1): Hydrostatic Pressure Distribution p g gk p p i.e. x y 0 and p g z or 2 2 1 1 dp gdz p p gdz g ( z )dz 2 1 r g g0 0 r 2 r g , constant near earth’s surface 𝑟 0 liquids 𝜌 = const. (for one liquid) 𝑝 = −𝜌𝑔𝑧 + constant gases 𝜌 = 𝜌(𝑝, 𝑡) which is known from the equation of state: 𝑝 = 𝜌𝑅𝑇 ⇒ 𝜌 = 𝑝/𝑅𝑇 dp p g g dz RT dp g dz p R T (z ) which can be integrated if 𝑇 = 𝑇(𝑧) is known as it is for the atmosphere. 058:0160 Jianming Yang Fall 2012 Chapter 2 8 4 Manometry Manometers are devices that use liquid columns for measuring differences in pressure. A general procedure may be followed in working all manometer problems: 1.) Start at one end (or a meniscus if the circuit is continuous) and write the pressure there in an appropriate unit or symbol if it is unknown. 2.) Add to this the change in pressure (in the same unit) from one meniscus to the next (plus if the next meniscus is lower, minus if higher). 3.) Continue until the other end of the gage (or starting meniscus) is reached and equate the expression to the pressure at that point, known or unknown. 𝑝𝐴 − 𝑝𝐵 = (𝑝𝐴 − 𝑝1 ) +(𝑝1 − 𝑝2 ) +(𝑝2 − 𝑝3 ) +(𝑝3 − 𝑝𝐵 ) = −𝛾1 (𝑧𝐴 − 𝑧1 ) −𝛾2 (𝑧1 − 𝑧2 ) −𝛾3 (𝑧2 − 𝑧3 ) −𝛾4 (𝑧3 − 𝑧𝐵 ) 058:0160 Jianming Yang Chapter 2 9 Fall 2012 5 Hydrostatic forces on plane surfaces The force on a body due to a pressure distribution is: F A pn dA where for a plane surface n = constant and we need only consider |F| noting that its direction is always towards the surface: F A p dA . Consider a plane surface entirely submerged in a liquid such that the plane of the surface intersects the free-surface with an angle α. The surface centroid is denoted ( x, y ). dF p dA hdA y sin dA F y sin dA A sin y dA A sin y A by definition , y 1 y dA A A pA F sin yA pA where p is the pressure at the centroid. 058:0160 Jianming Yang Chapter 2 10 Fall 2012 To find the line of action of the force which we call the center of pressure (xcp, ycp) we equate the moment of the resultant force to that of the distributed force about any arbitrary axis. ycp F ydF sin y 2dA A Note: dF y sin dA y 2 dA I xx y A I xx (parallel axis theorem) A A 2 𝐼𝑥𝑥 : moment of inertia about O-x ̅ : moment of inertia w.r.t. horizontal centroidal axis 𝐼𝑥𝑥 2 ycp F ycp sin y A sin y A I xx ycp y I xx yA xcp x I xy yA and similarly for xcp xcp F x dF sin xydA sin x y A I xy A A where the product of inertia A xydA I xy x yA I xy (parallel axis theorem) Note that the coordinate system in the textbook has its origin at the centroid and is related to the one just used by: xtextbook x x and ytextbook ( y y) 058:0160 Jianming Yang Chapter 2 11 Fall 2012 To obtain equation (2.29) in the textbook: 𝑥CP,textbook = 𝑥𝑐𝑝 − 𝑥̅ = (𝑥̅ + ̅ 𝐼𝑥𝑦 𝑦̅𝐴 ) − 𝑥̅ = 𝑦CP,textbook = −(𝑦𝑐𝑝 − 𝑦̅) = − [(𝑦̅ ̅ 𝐼𝑥𝑦 𝑦̅𝐴 ̅ 𝐼𝑥𝑥 + ̅𝐴 ) − 𝑦 =( 𝑦̅] ̅ 𝐼𝑥𝑦 = ̅ sin 𝜃 𝐼𝑥𝑦 ℎ𝑐 ⁄sin 𝜃 )𝐴 ℎ𝑐 𝐴 ̅ ̅ 𝐼𝑥𝑥 𝐼𝑥𝑥 = − ̅𝐴 = − ( ⁄ 𝑦 ℎ𝑐 sin 𝜃 )𝐴 =− ̅ sin 𝜃 𝐼𝑥𝑥 ℎ𝑐 𝐴 Notice: ℎ𝑐 here is the same ℎ𝐶𝐺 in the textbook: ℎ𝑐 = 𝑦̅ sin 𝜃 𝐼𝑥𝑥 is the moment of inertia w.r.t. horizontal centroidal axis, which is 𝐼𝑥𝑥 in the textbook ̅ is the product of inertia w.r.t. horizontal centroidal axis, which is −𝐼𝑥𝑦 in the textbook 𝐼𝑥𝑦 because the y axis has a different direction when using this definition: 𝐼𝑥𝑦 = ∫𝐴 𝑥𝑦𝑑𝐴 Therefore 𝑥CP,textbook = − 𝑦CP,textbook = − 𝐼𝑥𝑦 sin 𝜃 ℎ𝐶𝐺 𝐴 𝐼𝑥𝑥 sin 𝜃 ℎ𝐶𝐺 𝐴 058:0160 Jianming Yang Chapter 2 12 Fall 2012 With the equations in the textbook, the origin of the coordinates is defined at CG, 8ft 𝜃 = cos−1 10ft ℎ𝐶𝐺 = 12ft; 𝐴 = width × length = 5ft × 10ft = 50ft 2 Therefore, 𝐹 = 𝛾ℎ𝐶𝐺 𝐴 The Gate is rectangular, from Fig. 2.13a 𝑏𝐿3 𝐼𝑥𝑦 = 0, 𝐼𝑥𝑥 = 12 From Eq. (2.29) 𝑥CP = − 𝑦CP = − 𝐼𝑥𝑦 sin 𝜃 ℎ𝐶𝐺 𝐴 𝐼𝑥𝑥 sin 𝜃 ℎ𝐶𝐺 𝐴 =0 =… 058:0160 Jianming Yang Chapter 2 13 Fall 2012 6 Hydrostatic Forces on Curved Surfaces F pn dA In general, A Horizontal Components: Fx F i pn i dA A Fy Ay p dAy 𝑑𝐴𝑥 = 𝐧 ∙ 𝐢𝑑𝐴 = projection of 𝐧𝑑𝐴 onto a plane perpendicular to x direction That is, the horizontal component of force acting on a curved surface is equal to the force acting on a vertical projection of that surface (which includes both magnitude and line of action) and can be determined by the methods developed for plane surfaces. Fz pn k dA A Az p dAz h dAz Az Where h is the depth to any element area dA of the surface. That is, the vertical component of force acting on a curved surface is equal to the net weight of the total column of fluid directly above the curved surface and has a line of action through the centroid of the fluid volume. 058:0160 Jianming Yang Chapter 2 14 Fall 2012 Example Drum Gate p h R1 cos n sin i cos k dA lRd F pn dA R1 cos sin i cos k lRd A 0 p dA n 1 F i Fx l R 2 1 cos sin d l R 2 cos 0 cos 2 0 2 l R 2 0 4 R2 Rl p Same force as that on projection of gate onto vertical plane A Fz l R 2 0 2 1 cos cosd l R sin 1 sin 2 2 4 0 R 2 Net weight of water above surface lR l 2 2 1 1 F1 2 R 2l R 2l / 2 lR 2 2 / 2 2 2 Another approach: 1 1 F2 R 2l F1 F F2 F1 R 2l 2 2 2 058:0160 Jianming Yang Fall 2012 Chapter 2 15 7 Hydrostatic Forces in Layered Fluids Formulas for plane and curved surfaces apply separately to each layer: compute and sum the separate layer forces and moments. 058:0160 Jianming Yang Fall 2012 8 Buoyancy and Stability 8.1 Archimedes Principle 𝐹𝐵 = 𝐹𝑉(2) − 𝐹𝑉(1) = fluid weight above 2ABC – fluid weight above 1ADC = weight of fluid equivalent to body volume In general, 𝐹𝐵 = 𝜌fluid 𝑔∀displaced (∀displaced = displaced fluid volume). The line of action is through the centroid of the displaced volume, which is called the center of buoyancy. Example: Oscillating floating block Weight of the block 𝑊 = 𝜌𝑏 𝐿𝑏ℎ𝑔 = 𝑚𝑔 = 𝛾∀0 where ∀0 is displaced water volume by the block and 𝛾 is the specific weight of the liquid, waterline area 𝐴𝑤𝑙 = 𝐿𝑏. 𝑊 = 𝐵 ⇒ 𝜌𝑏 𝐿𝑏ℎ𝑔 = 𝜌𝑤 𝐿𝑏𝑑𝑔 𝜌 ⇒ 𝑑 = 𝑏⁄𝜌𝑤 ℎ Instantaneous displaced water volume: Chapter 2 16 058:0160 Jianming Yang Chapter 2 17 Fall 2012 ∀= ∀0 − 𝑦𝐴𝑤𝑙 ∑ 𝐹𝑉 = 𝑚𝑦̈ = 𝐵 − 𝑊 = 𝛾∀ − 𝛾∀0 = −𝛾𝐴𝑤𝑙 𝑦 𝑚𝑦̈ + 𝛾𝐴𝑤𝑙 𝑦 = 0 𝑦̈ + 𝛾𝐴𝑤𝑙 𝑚 𝑦=0 Solution for this homogeneous linear 2nd-order ODE: 𝑦 = 𝐴cos𝜔𝑡 + 𝐵sin𝜔𝑡 Use initial condition (𝑡 = 0: 𝑦 = 𝑦0 , 𝑦̇ = 𝑦̇ 0 ) to determine 𝐴 and 𝐵: 𝑦 = 𝑦0 cos𝜔𝑡 + 𝑦𝜔̇ 0sin𝜔𝑡 Where the angular frequency 𝛾𝐴𝑤𝑙 𝜔=√ period 𝑚 𝑇= 2𝜋 𝜔 = 2𝜋√ 𝑚 𝛾𝐴𝑤𝑙 Spar Buoy We can increase period 𝑇 by increasing block mass 𝑚 and/or decreasing waterline area 𝐴𝑤𝑙 . http://upload.wikimedia.org/wikipedia/com mons/0/03/Lateral_view_of_spar-buoy.png 058:0160 Jianming Yang Chapter 2 18 Fall 2012 8.2 Stability: Immersed Bodies Stable Neutral Unstable Condition for static equilibrium: (1) ∑Fv=0 and (2) ∑M=0 Condition (2) is met only when C and G coincide, otherwise we can have either a righting moment (stable) or a heeling moment (unstable) when the body is heeled. 058:0160 Jianming Yang Fall 2012 Chapter 2 19 8.3 Stability: Floating Bodies For a floating body the situation is slightly more complicated since the center of buoyancy will generally shift when the body is rotated, depending upon the shape of the body and the position in which it is floating. The center of buoyancy (centroid of the displaced volume) shifts laterally to the right for the case shown because part of the original buoyant volume aOc is transferred to a new buoyant volume bOd. The point of intersection of the lines of action of the buoyant force before and after heel is called the metacenter M and the distance GM is called the metacentric height. If GM is positive, that is, if M is above G, then the ship is stable; however, if GM is negative, then the ship is unstable. 058:0160 Jianming Yang Chapter 2 20 Fall 2012 Consider a ship which has taken a small angle of heel 𝜃 1. evaluate the lateral displacement of the center of buoyancy, 𝐵𝐵′ 2. then from trigonometry, we can solve for GM and evaluate the stability of the ship Recall that the center of buoyancy is at the centroid of the displaced volume of fluid (moment of volume about y-axis – ship centerplane) This can be evaluated conveniently as follows: 𝑥̅ ∀= ∫𝑐𝑂𝑑𝑒𝑎 𝑥𝑑∀ + ∫𝑂𝑏𝑑 𝑥𝑑∀ − ∫𝑐𝑂𝑎 𝑥𝑑∀ = 0 + ∫𝑂𝑏𝑑 𝑥 (𝐿𝑑𝐴) − ∫𝑐𝑂𝑎 𝑥(𝐿𝑑𝐴) = 0 + ∫𝑂𝑏𝑑 𝑥𝐿(𝑥tan𝜃𝑑𝑥 ) − ∫𝑐𝑂𝑎 𝑥𝐿(−𝑥tan𝜃𝑑𝑥 ) = tan𝜃 ∫waterline 𝑥 2 𝑑𝐴waterline = 𝐼𝑂 tan𝜃 058:0160 Jianming Yang Chapter 2 21 Fall 2012 ∫𝑐𝑂𝑑𝑒𝑎 𝑥𝑑∀: moment of ∀ before heel (goes to zero due to symmetry of original buoyant volume about centerplane) 𝑑𝐴 = 𝑦𝑑𝑥 = 𝑥tan𝜃𝑑𝑥 𝐼𝑂 = ∫𝑤𝑙 𝑥 2 𝑑𝐴𝑤𝑙 : area moment of inertia of ship waterline about its tilt axis 𝑂 𝑥̅ ∀= 𝐼𝑂 tan𝜃 𝐵𝐵′ = 𝑥̅ = 𝐵𝑀 = 𝑥̅ tan𝜃 = 𝐺𝑀 = 𝐼𝑂 ∀ 𝐼𝑂 ∀ 𝐼𝑂 tan𝜃 ∀ = 𝐺𝑀 + 𝐵𝐺 − 𝐵𝐺 This equation is used to determine the stability of floating bodies: If GM is positive, the body is stable If GM is negative, the body is unstable 058:0160 Jianming Yang Fall 2012 Chapter 2 22 8.4 Roll The rotation of a ship about the longitudinal axis through the center of gravity. Consider symmetrical ship heeled to a very small angle θ. Solve for the subsequent motion due only to hydrostatic and gravitational forces. 𝐅𝐵 = (cos 𝜃 𝐣 − sin𝜃𝐢)𝜌𝑔∀ 𝐌𝑔 = 𝐫 × 𝐅𝐵 = (−𝐺𝐵𝐣 + 𝐵𝐵′ 𝐢) × (cos 𝜃 𝐣 − sin𝜃𝐢)𝜌𝑔∀ = (−𝐺𝐵sin𝜃 + 𝐵𝐵′ cos 𝜃)𝜌𝑔∀𝐤 = (−𝐺𝐵sin𝜃 + 𝐵𝑀 tan 𝜃 cos 𝜃)𝜌𝑔∀𝐤 = (−𝐺𝐵 + 𝐵𝑀)sin𝜃𝜌𝑔∀𝐤 = 𝐺𝑀sin𝜃𝜌𝑔∀𝐤 Note: recall that 𝑀𝑂 = |𝐅|𝑑, where 𝑑 is the perpendicular distance from 𝑂 to the line of action of 𝐅: 𝑀𝐺 = 𝐺𝑓𝜌𝑔∀= 𝐺𝑀sin𝜃𝜌𝑔∀ Angular momentum: ∑ 𝑀𝐺 = −𝐼𝜃̈ 𝐼 = mass moment of inertia about long axis through 𝐺 𝜃̈= angular acceleration 058:0160 Jianming Yang Chapter 2 23 Fall 2012 𝐼𝜃̈ + 𝐺𝑀sin𝜃𝜌𝑔∀= 0 ̈ + 𝑚𝑔𝐺𝑀𝜃 = 0 For small 𝜃: 𝐼𝜃̈ + 𝐺𝑀𝜌𝑔∀𝜃 = 0 ⇒ 𝜃̈ + 𝐺𝑀𝜌𝑔∀ 𝜃 = 0 ⇒ 𝜃 𝐼 𝐼 Definition of radius of gyration: 𝑘 = √𝐼⁄𝑚 𝑘 2 = 𝐼⁄𝑚 ⇒ 𝑚𝑘 2 = 𝐼 ⇒ 𝑚𝑔𝐺𝑀 𝐼 = 𝑔𝐺𝑀 𝑘2 The solution to equation 𝜃̈ + 𝑔𝐺𝑀 𝜃 = 0 is, 𝑘2 𝜃(𝑡) = 𝜃0 cos 𝜔𝑡 + 𝜃̇0 sin 𝜔𝑡 𝜔 = 𝜃0 cos 𝜔𝑡 where 𝜃0 = the initial heel angle, 𝜃̇0 = 0 for no initial velocity, the natural frequency 𝑔𝐺𝑀 𝜔=√ 𝑘2 = √𝑔𝐺𝑀 𝑘 Simple (undamped) harmonic oscillation with period of the motion: 𝑇 = 2𝜋 𝜔 = 2𝜋𝑘 √𝑔𝐺𝑀 Note that large GM decreases the period of roll, which would make for an uncomfortable boat ride (high frequency oscillation). Earlier we found that GM should be positive if a ship is to have transverse stability and, generally speaking, the stability is increased for larger positive GM. However, the present example shows that one encounters a “design tradeoff” since large GM decreases the period of roll, which makes for an uncomfortable ride. 058:0160 Jianming Yang Chapter 2 24 Fall 2012 9 Case (2): Rigid Body Translation or Rotation In rigid body motion, all particles are in combined translation and/or rotation and there is no relative motion between particles; consequently, there are no strains or strain rates and the viscous term drops out of the N-S equation. ∇𝑝 = 𝜌(𝐠 − 𝐚) from which we see that ∇𝑝 acts in the direction of (𝐠 − 𝐚), and lines of constant pressure must be perpendicular to this direction (by definition, ∇𝑓 is perpendicular to 𝑓 = const.). For the general case of rigid body translation/rotation of fluid shown in the figure, if the center of rotation is at 𝑂 where 𝐕 = 𝐕0 , the velocity of any arbitrary point 𝑃 is: 𝐕 = 𝐕0 + 𝛀 × 𝐫0 where 𝛀 = the angular velocity vector, and the acceleration is: 𝑑𝐕 𝑑𝑡 =𝐚= First term = Second term = Third term = 𝑑𝐕0 𝑑𝑡 + 𝛀 × (𝛀 × 𝐫0 ) + 𝑑𝛀 𝑑𝑡 × 𝐫0 acceleration of 𝑂 centripetal acceleration of 𝑃 relative to 𝑂 linear acceleration of 𝑃 due to 𝛀 Usually, all these terms are not present. In fact, fluids can rarely move in rigid body motion unless restrained by confining walls for a long time. 058:0160 Jianming Yang Chapter 2 25 Fall 2012 9.1 Uniform Linear Acceleration ∇𝑝 = 𝜌(𝐠 − 𝐚) = −𝜌[(𝑔 + 𝑎𝑧 )𝐤 + 𝑎𝑥 𝐢] = const. 𝜕𝑝 𝜕𝑥 1. 𝑎𝑥 < 0, 2. 𝑎𝑥 > 0, = −𝜌𝑎𝑥 𝑝 increase in +𝑥 𝑝 decrease in +𝑥 𝜕𝑝 𝜕𝑧 = −𝜌(𝑔 + 𝑎𝑧 ) 1. 𝑎𝑧 > 0, 𝑝 decrease in +𝑧 2. 𝑎𝑧 < 0 and |𝑎𝑧 | < 𝑔, 𝑝 decrease in +𝑧 3. 𝑎𝑧 < 0 and |𝑎𝑧 | > 𝑔, 𝑝 increase in +𝑧 Unit vector in the direction of ∇𝑝: 𝐬=| ∇𝑝 ∇𝑝 =− | (𝑔+𝑎𝑧 )𝐤+𝑎𝑥 𝐢 1⁄2 [(𝑔+𝑎𝑧 )2 +𝑎𝑥2 ] Lines of constant pressure are perpendicular to ∇𝑝. Angle between the surface of constant pressure and the 𝑥 axes: 𝜃 = tan−1 𝑎𝑥 𝑔+𝑎𝑧 In general the rate of increase of pressure in the direction (𝐠 − 𝐚) is given by: 𝑑𝑝 𝑑𝑠 = ∇𝑝 ∙ 𝐬 = 𝜌[(𝑔 + 𝑎𝑧 )2 + 𝑎𝑥2 ]1⁄2 = 𝜌𝐺 𝑝 = 𝜌𝐺𝑠 + const. = 𝜌𝐺𝑠 gage pressure . 058:0160 Jianming Yang Chapter 2 26 Fall 2012 9.2 Rigid Body Rotation Consider rotation of the fluid about the 𝑧 axis without any translation. 𝐚 = 𝛀 × (𝛀 × 𝐫0 ) = −𝑟Ω2 𝐢𝑟 ∇𝑝 = 𝜌(𝐠 − 𝐚) = 𝜌(−𝑔𝒌 + 𝑟Ω2 𝐢𝑟 ) 𝜕𝑝 𝜕𝑟 = 𝜌𝑟Ω2 , 𝜕𝑝 𝜕𝑧 = −𝜌𝑔 = −𝛾 and 1 𝑝 = 𝜌𝑟 2 Ω2 + 𝑓(𝑧) + 𝑐 𝜕𝑝 2 = 𝑓 ′ = −𝛾 ⇒ 𝑝 = −𝛾𝑧 + 𝑓(𝑟) + 𝑐 𝜕𝑧 ⇒ 𝑓(𝑧) = −𝛾𝑧 1 2 2 𝑝 = 𝜌𝑟 Ω − 𝛾𝑧 + const. 2 The constant is determined by specifying the pressure at one point; say, 𝑝 = 𝑝0 at (𝑟, 𝑧) = (0,0) 1 𝑝 = 𝑝0 − 𝛾𝑧 + 𝜌𝑟 2 Ω2 (Note: Pressure is linear in 𝑧 and parabolic in 𝑟) 2 Curves of constant pressure are given by: 𝑧= 𝑝0 −𝑝 𝛾 + 𝑟 2 Ω2 2𝑔 = 𝑎 + 𝑏𝑟 2 which are paraboloids of revolution, concave upward, with their minimum points on the axis of rotation. 058:0160 Jianming Yang Chapter 2 27 Fall 2012 The position of the free surface is found, as it is for linear acceleration, by conserving the volume of fluid. Unit vector in the direction of ∇𝑝: 𝐬= ∇𝑝 |∇𝑝| = −𝛾𝐤+𝜌𝑟Ω2 𝐢𝑟 − [ 2 ( 2)2 ]1⁄2 𝛾 + 𝜌𝑟Ω 𝑑𝑧 𝑔 Slope of 𝐬: tan 𝜃 = 𝑑𝑟 = 𝑟Ω2 . (𝜃 is the angle between the surface of constant pressure and the 𝑥 axis) i.e., 𝑟 = 𝐶1 exp (− Ω2 𝑧 𝑔 ) is the equation of ∇𝑝 surfaces. 058:0160 Jianming Yang Chapter 2 28 Fall 2012 10 Case (3): Pressure Distribution in Irrotational Flow Potential flow solutions also solutions of NS under such conditions: 1. If viscous effects are neglected, Navier-Stokes equation becomes Euler equation: 𝜕𝐕 𝜌 ( + 𝐕 ∙ ∇𝐕) = 𝜌𝐠 − ∇𝑝 𝜕𝐕 𝜕𝑡 𝜕𝑡 1 + ∇ ( 𝐕 ∙ 𝐕) − 𝐕 × (𝛁 × 𝐕) = 𝜌𝐠−∇𝑝 2 𝜌 1 = − ∇(𝑝 + 𝜌𝑔𝑧) 𝜌 1 Vector calculus identity: ∇ ( 𝐕 ∙ 𝐕) = 𝐕 ∙ ∇𝐕 + 𝐕 × (𝛁 × 𝐕) 2 2. If = const., 𝜕𝐕 𝜕𝑡 + ∇( 3. Assume a steady flow: 𝑉2 2 𝜕 𝜕𝑡 𝑝 + + 𝑔𝑧) = 𝐕 × 𝛚 𝜌 (𝑉 2 = 𝐕 ∙ 𝐕) =0 ∇( 𝑉2 2 𝑝 + + 𝑔𝑧) = ∇𝐵 = 𝐕 × 𝛚 𝜌 Consider: ∇𝐵 perpendicular to 𝐵 = const., also 𝐕 × 𝛚 perpendicular to 𝐕 and 𝛚. Stream lines 𝐞𝑠 : 𝐕 × 𝐞𝑠 = 0; vortex lines 𝐞𝑣 : 𝛚 × 𝐞𝑣 = 0 𝐞𝑠 ∙ ∇𝐵 = 𝜕𝐵 𝜕𝑠 = 𝐞𝑠 ∙ (𝐕 × 𝛚) = 𝛚 ∙ (𝐞𝑠 × 𝐕) = 0 𝐞𝑣 ∙ ∇𝐵 = 𝐞𝑣 ∙ (𝐕 × 𝛚) = 𝐕 ∙ (𝛚 × 𝐞𝑠 ) = 0 Therefore, 𝐵 = 𝑉2 2 𝑝 + + 𝑔𝑧 = const. contains streamlines and vortex lines: 𝜌 058:0160 Jianming Yang Chapter 2 29 Fall 2012 1. Assuming irrotational flow: 𝛚 = 0 ∇𝐵 = 0 𝐵 = const. (everywhere same constant) 2. Unsteady irrotational flow 𝐕 = ∇𝜑 𝑉 2 = ∇𝜑 ∙ ∇𝜑 ∇( 𝜕𝜑 𝜕𝑡 + 𝑉2 2 𝑝 + + 𝑔𝑧) = 0 𝜌 𝜕𝜑 𝑉 2 𝑝 + + + 𝑔𝑧 = 𝐵(𝑡) 𝜕𝑡 2 𝜌 𝐵(𝑡) is a time-dependent constant. Alternate derivation using streamline coordinates: 𝐕 = 𝑣𝑠 (𝑠, 𝑡)𝐞𝑠 + 𝑣𝑛 𝐞𝑛 = 𝑣𝑠 (𝑠, 𝑡)𝐞𝑠 𝜕 ∇= 𝐚= =[ 𝐞𝑠 + 𝜕𝑠 𝐷𝐕 𝐷𝑡 𝜕𝑣𝑠 𝜕𝑡 = 𝜕 𝐞 𝜕𝑛 𝑛 𝜕𝐕 𝜕𝑡 + 𝐕 ∙ ∇𝐕 𝐞𝑠 + 𝑣𝑠 𝜕𝐞𝑠 𝜕𝑡 ] + 𝑣𝑠 [ 𝜕𝑣𝑠 𝜕𝑠 𝐞𝑠 + 𝑣𝑠 𝜕𝐞𝑠 𝜕𝑠 ] 058:0160 Jianming Yang Chapter 2 30 Fall 2012 Time increment: 𝜕𝐞𝑠 Space increment: 𝐚=[ 𝜕𝑣𝑠 𝜕𝑡 𝜕𝑣𝑛 − 𝜕𝑡 + 𝑣𝑠 𝜕𝑠 𝜕𝑣𝑠 𝜕𝑠 = 𝜕𝜃 𝐞 𝜕𝑡 𝑛 𝜕𝜃 − 𝐞𝑛 𝜕𝑠 1 = − 𝐞𝑛 ] 𝐞𝑠 + [−𝑣𝑠 𝑅 𝜕𝜃 𝜕𝑡 − 𝑣𝑠2 𝑅 ] 𝐞𝑛 : local 𝑎𝑠 in the direction of flow 𝜕𝑡 𝑣𝑠 𝜕𝑣𝑠 𝜕𝑡 𝜕𝐞𝑠 =− = −𝑣𝑠 𝜕𝑣𝑠 𝜕𝑠 𝑣𝑠2 𝑅 𝜕𝜃 𝜕𝑡 : local 𝑎𝑛 normal to the direction of flow : convective 𝑎𝑠 due to convergence/divergence of streamlines : normal 𝑎𝑛 due to streamline curvature Euler Equation: 𝜌𝐚 = −∇(𝑝 + 𝜌𝑔𝑧) 𝜕𝑣 Steady flow 𝑠-direction equation: 𝜌𝑣𝑠 𝑠 = − 𝜕𝑠 𝜕 𝜕𝑠 ( 𝑣𝑠2 2 𝜕 𝜕𝑠 (𝑝 + 𝜌𝑔𝑧) 𝑝 + + 𝑔𝑧) = 0, i.e., B=const. along streamline 𝜌 Steady flow 𝑛-direction equation: −𝜌 −∫ 𝑣𝑠2 𝑅 𝑝 𝑣𝑠2 𝑅 =− 𝜕 𝜕𝑛 (𝑝 + 𝜌𝑔𝑧) 𝑑𝑛 + + 𝑔𝑧 = const. across streamline 𝜌 058:0160 Jianming Yang Fall 2012 Chapter 2 31 11 Flow Patterns: Streamlines, Streaklines, Pathlines 1.) A Streamline is a curve everywhere tangent to the local velocity vector at a given instant. Instantaneous lines; convinent to compute mathematically. 2.) A Pathline is the actual path traveled by an individual fluid particle over some time period. Generated as the passage of time; convinent to generate experimentally. 058:0160 Jianming Yang Fall 2012 Chapter 2 32 3.) A Streakline is the locus of particles that have earlier passed through a prescribed point. Generated as the passage of time; convinent to generate experimentally. 4.) A Timeline is a set of fluid particles that form a line at a given instant. Instantaneous lines; convinent to generate experimentally. Note: 1. Streamlines, pathlines, and streaklines are identical in a steady flow. 2. For unsteady flow, streamline pattern changes with time, whereas pathlines and streaklines are generated as the passage of time. 058:0160 Jianming Yang Chapter 2 33 Fall 2012 11.1 Streamline By definition we must have 𝐕 × 𝑑𝐫 = 0 which upon expansion yields the equation of the streamlines for a given time 𝑡 = 𝑡1 𝑑𝑥 𝑢 = 𝑑𝑦 𝑣 = 𝑑𝑧 𝑤 = 𝑑𝑠 𝑠 = integration parameter So if (𝑢, 𝑣, 𝑤) is known, integrate with respect to 𝑠 for 𝑡 = 𝑡1 with initial condition (𝑥0 , 𝑦0 , 𝑧0 , 𝑡0 ) at 𝑠 = 0 and then eliminate 𝑠. 11.2 Pathline The pathline is defined by integration of the relationship between velocity and displacement. 𝑑𝑥 𝑑𝑡 =𝑢 𝑑𝑦 𝑑𝑡 =𝑣 𝑑𝑧 𝑑𝑡 =𝑤 Integrate 𝑢, 𝑣, 𝑤 with respect to 𝑡 using initial condition (𝑥0 , 𝑦0 , 𝑧0 , 𝑡0 ), then eliminate 𝑡. 11.3 Streakline To find the streakline, use the integrated result for the pathline retaining time as a parameter. Now, find the integration constant which causes the pathline to pass through (𝑥0 , 𝑦0 , 𝑧0 ) for a sequence of times 𝜏 < 𝑡. Then eliminate 𝜏. 058:0160 Jianming Yang Chapter 2 34 Fall 2012 Example: an idealized velocity distribution is given by: 𝑢= 𝑥 1+𝑡 𝑣= 𝑦 𝑤=0 1+2𝑡 calculate and plot: 1) the streamlines 2) the pathlines 3) the streaklines which pass through (𝑥0 , 𝑦0 , 𝑧0 ) at 𝑡 = 0. 1. First, note that since 𝑤 = 0 there is no motion in the 𝑧 direction and the flow is 2-D 𝑑𝑥 𝑥 𝑑𝑦 𝑦 =𝑢= =𝑣= 𝑑𝑠 1+𝑡 𝑠 𝑑𝑠 𝑥 = 𝐶1 exp( ) 1+𝑡 𝑠 = 0 at (𝑥0 , 𝑦0 ): 1+2𝑡 𝑠 𝑦 = 𝐶2 exp( ) 1+2𝑡 𝐶1 = 𝑥0 𝐶2 = 𝑦0 and eliminating 𝑠: 𝑠 = (1 + 𝑡)ln 𝑥 𝑥0 𝑛 𝑥 = (1 + 2𝑡)ln 𝑦 = 𝑦0 ( ) where 𝑛 = 𝑥0 𝑦 𝑦0 1+𝑡 1+2𝑡 This is the equation of the streamlines which pass through (𝑥0 , 𝑦0 ) for all times 𝑡. 𝑦 𝑡 = 0, 𝑡 = ∞, 𝑦0 𝑦 𝑦0 = 𝑥 𝑥0 𝑥 1⁄2 =( ) 𝑥0 058:0160 Jianming Yang Chapter 2 35 Fall 2012 2. To find the pathlines we integrate 𝑑𝑥 𝑑𝑡 =𝑢= 𝑥 𝑑𝑦 1+𝑡 𝑑𝑡 =𝑣= 𝑦 1+2𝑡 𝑐 𝑐 𝑥 = 𝐶1 (1 + 𝑡) 𝑦 = 𝐶2 (1 + 2𝑡)1⁄2 (∫ 𝑑𝑥 = 𝑙𝑛|𝑎𝑥 + 𝑏| + 𝐶) 𝑎𝑥+𝑏 𝑎 𝑡 = 0 (𝑥, 𝑦) = (𝑥0 , 𝑦0 ): 𝐶1 = 𝑥0 𝐶2 = 𝑦0 now eliminate 𝑡 between the equations for (𝑥, 𝑦) 𝑦 = 𝑦0 [1 + 2 ( 𝑥 𝑥0 − 1)] 1⁄2 This is the pathline through (𝑥0 , 𝑦0 ) at 𝑡 = 0 and does not coincide with the streamline at 𝑡 = 0. 3. To find the streakline, we use the pathline equations to find the family of particles that have passed through the point (𝑥0 , 𝑦0 ) for all times 𝜏 < 𝑡. 𝑥 = 𝐶1 (1 + 𝑡) 𝑦 = 𝐶2 (1 + 2𝑡)1⁄2 𝑥0 = 𝐶1 (1 + 𝜏) 𝑦0 = 𝐶2 (1 + 2𝜏)1⁄2 𝐶1 = 𝑥= 𝑥0 1+𝜏 𝑥0 1+𝜏 (1 + 𝑡) 𝜏 = (1 + 𝑡) 𝑥0 𝑥 𝑦0 1+2𝜏)1⁄2 𝑦 = ( 0)1⁄2 (1 + 2𝑡)1⁄2 1+2𝜏 1 𝑦0 2 𝐶2 = ( 𝑦 − 1 = [(1 + 2𝑡) ( ) − 1] 2 𝑦 058:0160 Jianming Yang Chapter 2 36 Fall 2012 𝑦0 2 ( ) = 𝑦 𝑡 = 0: 1+2𝑡 𝑥 1+2[(1+𝑡)( 0 )−1] 𝑥 𝑦 𝑦0 = [1 + 2 ( 𝑥0 𝑥 − 1)] −1⁄2 The streakline does not coincide with either the equivalent streamline or pathline. Physically, the streakline reflects the streamline behavior before the specified time 𝑡 = 0, while the pathline reflects the streamline behavior after 𝑡 = 0. 058:0160 Jianming Yang Chapter 2 37 Fall 2012 11.4 Stream Function The stream function is a powerful tool for 2D flows in which 𝐕 is obtained by differentiation of a scalar 𝛹 which automatically satisfies the continuity equation. Continuity equation: say: 𝑢 = 𝜕 then: 𝐕= 𝜕𝑥 𝜕𝛹 𝜕𝑦 𝜕𝛹 ( 𝜕𝛹 )+ 𝜕𝑦 𝜕𝛹 𝜕𝑥 𝜕 𝜕𝑦 + 𝜕𝑣 =0 𝜕𝑥 𝜕𝑦 𝜕𝛹 𝑣=− 𝜕𝑦 𝐢− 𝜕𝑢 (− 𝜕𝑥 𝜕𝛹 𝜕𝑥 )= 𝜕2 𝛹 𝜕𝑥𝜕𝑦 − 𝜕2 𝛹 𝜕𝑥𝜕𝑦 =0 𝐣 ⇒ ∇ × 𝐕 = 𝛚 ⇒ 𝜔𝑧 = 𝜔 = −∇2 𝛹 2D vorticity transport equation: 𝜕𝜔 𝜕𝑡 +𝑢 𝜕𝜔 𝜕𝑥 +𝑣 𝜕𝜔 𝜕𝑦 = 𝜈∇2 𝜔 Replace 𝑢, 𝑣, 𝜔: 𝜕 𝜕𝑡 2 (∇ 𝛹) + 𝜕𝛹 𝜕 (∇ 𝛹) − 𝜕𝑦 𝜕𝑥 𝜕𝛹 𝜕 Steady flow: 2 𝜕𝑦 𝜕𝑥 𝜕𝛹 𝜕 4 (∇ 𝛹) = 𝜈∇ 𝛹 𝜕𝑥 𝜕𝑦 𝜕𝛹 𝜕 (∇2 𝛹) − 2 𝜕𝑥 𝜕𝑦 4 (∇ 𝛹 = 𝜕4 𝛹 𝜕𝑥 4 +2 𝜕4 𝛹 𝜕𝑥 2 𝜕𝑦 2 (∇2 𝛹) = 𝜈∇4 𝛹 It is a single fourth-order scalar equation, which requires 4 boundary conditions At infinity: 𝑢 = 𝜕𝛹/𝜕𝑦 = 𝑈∞ 𝑣 = −𝜕𝛹/𝜕𝑥 = 0 On body: 𝑢 = 𝑣 = 0 = 𝜕𝛹/𝜕𝑦 = −𝜕𝛹/𝜕𝑥 + 𝜕4 𝛹 𝜕𝑦 4 ) 058:0160 Jianming Yang Chapter 2 38 Fall 2012 11.4.1 Irrotational Flow ∇×𝐕=0 ⇒ ∇2 𝛹 = 0 Second-order linear Laplace equation At infinity 𝑆∞ : 𝛹 = 𝑈∞ 𝑦 + const. On body 𝑆𝐵 : 𝛹 = const. 11.4.2 Geometric Interpretation of 𝜳 Besides its importance mathematically 𝛹 also has important geometric significance. 𝛹 = const. = streamline Recall definition of a streamline: 𝐕 × 𝑑𝐫 = 0 𝑑𝑥 𝑢 = 𝑑𝑦 𝑣 ⇒ 𝑢𝑑𝑦 − 𝑣𝑑𝑥 = 0 Compare with 𝑑𝛹 = 𝜕𝛹 𝜕𝑥 𝑑𝑥 + 𝜕𝛹 𝜕𝑦 𝑑𝑦 = −𝑣𝑑𝑥 + 𝑢𝑑𝑦 i.e., 𝑑𝛹 = 0 along a streamline Or 𝛹 = const. along a streamline and curves of constant 𝛹 are the flow streamlines. If we know 𝛹(𝑥, 𝑦) then we can plot 𝛹 = const. curves to show streamlines. 058:0160 Jianming Yang Chapter 2 39 Fall 2012 11.4.3 Physical Interpretation 𝑛𝑥 = 𝑑𝑦 𝑑𝑠 𝑛𝑦 = − 𝑑𝑥 𝑑𝑠 𝜕𝛹 𝜕𝛹 𝑑𝑦 𝑑𝑥 𝜕𝛹 𝜕𝛹 𝑑𝑄 = 𝐕 ∙ 𝐧𝑑𝐴 = ( 𝐢− 𝐣) ∙ ( 𝐢 − 𝐣) 𝑑(𝑠 ∙ 1) = 𝑑𝑦 + 𝑑𝑥 = 𝑑𝛹 𝜕𝑦 𝜕𝑥 𝑑𝑠 𝑑𝑠 𝜕𝑦 𝜕𝑥 i.e. change in 𝑑𝛹 is the volume flux and along a streamline 𝑑𝑄 = 0. Consider flow between two streamlines 𝐵 𝐵 𝑄𝐴𝐵 = ∫𝐴 𝐕 ∙ 𝐧𝑑𝐴 = ∫𝐴 𝑑𝛹 = 𝛹𝐵 − 𝛹𝐴 058:0160 Jianming Yang Chapter 2 40 Fall 2012 11.4.4 Incompressible Plane Flow in Polar Coordinates Continuity equation: or 𝜕 𝜕𝑟 (𝑟𝑣𝑟 ) + say: 𝑣𝑟 = then: 𝜕 𝜕𝑟 𝜕𝑣𝜃 𝜕𝜃 1 𝜕𝛹 𝑟 𝜕𝜃 1 𝜕𝛹 (𝑟 𝑟 𝜕𝜃 1 𝜕 𝑟 𝜕𝑟 (𝑟𝑣𝑟 ) + 𝑟 𝜕𝜃 =0 =0 𝑣𝜃 = − )+ 1 𝜕𝑣𝜃 𝜕 𝜕𝜃 (− 𝜕𝛹 𝜕𝑟 𝜕𝛹 𝜕𝑟 )=0 As before 𝑑𝛹 = 0 along a streamline and 𝑑𝑄 = 𝑑𝛹 Volume flux = change in stream function 11.4.5 Incompressible Axisymmetric Flow Continuity equation: say: 𝑣𝑟 = − then: 1 𝜕 𝑟 𝜕𝑟 (𝑟 1 𝜕𝛹 𝑟 𝜕𝑧 −1 𝜕𝛹 𝑟 𝜕𝑧 1 𝜕 𝑟 𝜕𝑟 𝑣𝑧 = )+ (𝑟𝑣𝑟 ) + 𝜕𝑣𝑧 𝜕𝑧 =0 1 𝜕𝛹 𝑟 𝜕𝑟 𝜕 1 𝜕𝛹 ( 𝜕𝑧 𝑟 𝜕𝑟 )=0 As before 𝑑𝛹 = 0 along a streamline and 𝑑𝑄 = 𝑑𝛹 Volume flux = change in stream function 058:0160 Jianming Yang Chapter 2 41 Fall 2012 11.4.6 Generalization to Steady Plane Compressible Flow In steady compressible flow, the continuity equation is 𝜕𝜌𝑢 𝜕𝑥 + 𝜕𝜌𝑣 𝜕𝑦 Define: 𝜌𝑢 = =0 𝜕𝛹 𝜕𝑦 𝜌𝑣 = − 𝜕𝛹 𝜕𝑥 Streamline: 𝑢𝑑𝑦 − 𝑣𝑑𝑥 = 0 Compare with 𝑑𝛹 = 𝜕𝛹 𝜕𝑥 1 𝜕𝛹 𝜌 𝜕𝑦 𝑑𝑥 + 𝑑𝑦 + 𝜕𝛹 𝜕𝑦 1 𝜕𝛹 𝜌 𝜕𝑥 𝑑𝑦 ⇒ 𝑑𝑥 = 0 1 𝜌 (𝑑𝛹) = 0 , i.e., 𝑑𝛹 = 0 and 𝛹 = const. is a streamline. Now: 𝑑𝑚̇ = 𝜌(𝐕 ∙ 𝐧)𝑑𝐴 = 𝑑𝛹 𝐵 𝑑𝑚̇𝐴𝐵 = ∫𝐴 𝜌(𝐕 ∙ 𝐧)𝑑𝐴 = 𝛹𝐵 − 𝛹𝐴 Change in 𝛹 is equivalent to the mass flux.