Intuition The big pictureWhen fluid flows past a flat plate, a thin boundary layer grows where viscosity glues the fluid to the wall (no-slip). Far away the flow zips along at U ∞ U_\infty U ∞ exact (well, exactly-reduced) answer to the question: what is the velocity profile and how thick is this layer?
The magic trick: the profile looks the same shape at every x x x y y y similarity solution — one universal curve describes the whole plate.
Definition Boundary layerA thin region near a solid wall where viscous forces are comparable to inertial forces , so the velocity rises from 0 0 0 ≈ U ∞ \approx U_\infty ≈ U ∞
WHY can't we just solve the full Navier–Stokes equations? Because they're brutal nonlinear PDEs. Prandtl's insight (1904): inside a thin layer, the equations simplify enormously. Blasius (1908, his PhD) then solved that simplified set for the flat plate.
Start with 2D steady incompressible flow, no pressure gradient (flat plate, ∂ p / ∂ x = 0 \partial p/\partial x = 0 ∂ p / ∂ x = 0
Continuity:
∂ u ∂ x + ∂ v ∂ y = 0 \frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} = 0 ∂ x ∂ u + ∂ y ∂ v = 0
x-momentum (Navier–Stokes):
u ∂ u ∂ x + v ∂ u ∂ y = ν ( ∂ 2 u ∂ x 2 + ∂ 2 u ∂ y 2 ) u\frac{\partial u}{\partial x} + v\frac{\partial u}{\partial y} = \nu\left(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2}\right) u ∂ x ∂ u + v ∂ y ∂ u = ν ( ∂ x 2 ∂ 2 u + ∂ y 2 ∂ 2 u )
∂ 2 u / ∂ x 2 \partial^2 u/\partial x^2 ∂ 2 u / ∂ x 2 thin : δ ≪ x \delta \ll x δ ≪ x y y y fast ; variation along it (in x x x slow . So ∂ 2 u / ∂ y 2 ≫ ∂ 2 u / ∂ x 2 \partial^2 u/\partial y^2 \gg \partial^2 u/\partial x^2 ∂ 2 u / ∂ y 2 ≫ ∂ 2 u / ∂ x 2
Prandtl boundary-layer equation:
u ∂ u ∂ x + v ∂ u ∂ y = ν ∂ 2 u ∂ y 2 u\frac{\partial u}{\partial x} + v\frac{\partial u}{\partial y} = \nu\frac{\partial^2 u}{\partial y^2} u ∂ x ∂ u + v ∂ y ∂ u = ν ∂ y 2 ∂ 2 u
with boundary conditions:
u ( x , 0 ) = 0 , v ( x , 0 ) = 0 , u ( x , ∞ ) = U ∞ u(x,0)=0,\quad v(x,0)=0,\quad u(x,\infty)=U_\infty u ( x , 0 ) = 0 , v ( x , 0 ) = 0 , u ( x , ∞ ) = U ∞
Intuition The scaling guessBalance inertia ∼ U ∞ 2 / x \sim U_\infty^2/x ∼ U ∞ 2 / x ∼ ν U ∞ / δ 2 \sim \nu U_\infty/\delta^2 ∼ ν U ∞ / δ 2 δ ∼ ν x / U ∞ \delta \sim \sqrt{\nu x / U_\infty} δ ∼ ν x / U ∞ η = y U ∞ ν x \eta = y\sqrt{\frac{U_\infty}{\nu x}} η = y ν x U ∞ η \eta η one curve .
Define the similarity variable and a dimensionless stream function f f f η = y U ∞ ν x , ψ = ν x U ∞ f ( η ) \boxed{\eta = y\sqrt{\frac{U_\infty}{\nu x}}}, \qquad \psi = \sqrt{\nu x\, U_\infty}\; f(\eta) η = y ν x U ∞ , ψ = ν x U ∞ f ( η )
Since u = ∂ ψ / ∂ y u = \partial\psi/\partial y u = ∂ ψ / ∂ y v = − ∂ ψ / ∂ x v=-\partial\psi/\partial x v = − ∂ ψ / ∂ x WHY we use a stream function), we compute:
u = U ∞ f ′ ( η ) u = U_\infty f'(\eta) u = U ∞ f ′ ( η )
Why this step? ∂ ψ / ∂ y = ν x U ∞ f ′ ( η ) ⋅ ∂ η / ∂ y = ν x U ∞ f ′ ⋅ U ∞ / ( ν x ) = U ∞ f ′ \partial\psi/\partial y = \sqrt{\nu x U_\infty}\, f'(\eta)\cdot \partial\eta/\partial y = \sqrt{\nu x U_\infty}\, f' \cdot \sqrt{U_\infty/(\nu x)} = U_\infty f' ∂ ψ / ∂ y = ν x U ∞ f ′ ( η ) ⋅ ∂ η / ∂ y = ν x U ∞ f ′ ⋅ U ∞ / ( ν x ) = U ∞ f ′
v = 1 2 ν U ∞ x ( η f ′ − f ) v = \frac{1}{2}\sqrt{\frac{\nu U_\infty}{x}}\,(\eta f' - f) v = 2 1 x ν U ∞ ( η f ′ − f )
Substituting u , v u,v u , v every x x x U ∞ U_\infty U ∞
This is a 3rd-order nonlinear ODE with no closed form — solved numerically. The key numerical result:
f ′ ′ ( 0 ) = 0.332 f''(0) = 0.332 f ′′ ( 0 ) = 0.332
Why this step (the 0.664 0.664 0.664 C f = 0.332 μ U ∞ U ∞ / ν x 1 2 ρ U ∞ 2 = 2 ( 0.332 ) U ∞ x / ν = 0.664 R e x C_f = \dfrac{0.332\,\mu U_\infty \sqrt{U_\infty/\nu x}}{\tfrac12\rho U_\infty^2} = \dfrac{2(0.332)}{\sqrt{U_\infty x/\nu}} = \dfrac{0.664}{\sqrt{Re_x}} C f = 2 1 ρ U ∞ 2 0.332 μ U ∞ U ∞ / ν x = U ∞ x / ν 2 ( 0.332 ) = R e x 0.664
Worked example Example 1 — thickness of the layerAir (ν = 1.5 × 10 − 5 m 2 / s \nu=1.5\times10^{-5}\,\text{m}^2/\text{s} ν = 1.5 × 1 0 − 5 m 2 / s U ∞ = 10 m/s U_\infty=10\,\text{m/s} U ∞ = 10 m/s δ \delta δ x = 0.5 m x=0.5\,\text{m} x = 0.5 m
Step 1 — Reynolds number: R e x = 10 × 0.5 1.5 × 10 − 5 = 3.33 × 10 5 Re_x = \frac{10\times0.5}{1.5\times10^{-5}} = 3.33\times10^{5} R e x = 1.5 × 1 0 − 5 10 × 0.5 = 3.33 × 1 0 5 Why? Confirms laminar (< 5 × 10 5 <5\times10^5 < 5 × 1 0 5 Step 2 — δ = 5.0 x R e x = 5.0 × 0.5 3.33 × 10 5 = 2.5 577 = 4.3 mm \delta = \frac{5.0 x}{\sqrt{Re_x}} = \frac{5.0\times 0.5}{\sqrt{3.33\times10^5}} = \frac{2.5}{577} = 4.3\,\text{mm} δ = R e x 5.0 x = 3.33 × 1 0 5 5.0 × 0.5 = 577 2.5 = 4.3 mm Why? Plug into the boxed formula. A 4 mm layer over half a metre — genuinely thin ✓.
Worked example Example 2 — wall shear at a pointSame flow, find τ w \tau_w τ w x = 0.5 x=0.5\, x = 0.5 ρ = 1.2 kg/m 3 \rho=1.2\,\text{kg/m}^3 ρ = 1.2 kg/m 3
Step 1 — C f = 0.664 / R e x = 0.664 / 577 = 1.15 × 10 − 3 C_f = 0.664/\sqrt{Re_x} = 0.664/577 = 1.15\times10^{-3} C f = 0.664/ R e x = 0.664/577 = 1.15 × 1 0 − 3 Why? Local friction coefficient first — dimensionless, easy.
Step 2 — τ w = C f ⋅ 1 2 ρ U ∞ 2 = 1.15 × 10 − 3 × 1 2 × 1.2 × 100 = 0.069 Pa \tau_w = C_f\cdot\tfrac12\rho U_\infty^2 = 1.15\times10^{-3}\times\tfrac12\times1.2\times100 = 0.069\,\text{Pa} τ w = C f ⋅ 2 1 ρ U ∞ 2 = 1.15 × 1 0 − 3 × 2 1 × 1.2 × 100 = 0.069 Pa Why? Un-normalize using the dynamic pressure 1 2 ρ U ∞ 2 \tfrac12\rho U_\infty^2 2 1 ρ U ∞ 2
Worked example Example 3 — total dragPlate length L = 0.5 L=0.5 L = 0.5 b = 0.2 b=0.2 b = 0.2
Step 1 — R e L = 3.33 × 10 5 Re_L=3.33\times10^5 R e L = 3.33 × 1 0 5 C D = 1.328 / R e L = 2.30 × 10 − 3 C_D=1.328/\sqrt{Re_L}=2.30\times10^{-3} C D = 1.328/ R e L = 2.30 × 1 0 − 3 Why? C D C_D C D averaged friction (twice C f C_f C f Step 2 — F D = C D ⋅ 1 2 ρ U ∞ 2 ⋅ ( b L ) = 2.30 × 10 − 3 × 60 × 0.1 = 0.0138 N F_D = C_D\cdot\tfrac12\rho U_\infty^2\cdot(bL) = 2.30\times10^{-3}\times60\times0.1 = 0.0138\,\text{N} F D = C D ⋅ 2 1 ρ U ∞ 2 ⋅ ( b L ) = 2.30 × 1 0 − 3 × 60 × 0.1 = 0.0138 N Why? Multiply by plate area b L = 0.1 m 2 bL=0.1\,\text{m}^2 b L = 0.1 m 2
δ \delta δ x x x Why it feels right: the layer clearly thickens downstream, so a straight-line growth seems natural.
The fix: it grows as x \sqrt{x} x δ ∝ ν x / U ∞ \delta\propto\sqrt{\nu x/U_\infty} δ ∝ ν x / U ∞ decelerates . Linear growth would over-predict thickness far downstream.
δ = 5.0 ν x / U ∞ \delta=5.0\sqrt{\nu x/U_\infty} δ = 5.0 ν x / U ∞ Why it feels right: it's the formula you memorized.
The fix: Blasius is laminar only (R e x ≲ 5 × 10 5 Re_x \lesssim 5\times10^5 R e x ≲ 5 × 1 0 5 x 4 / 5 x^{4/5} x 4/5 δ / x ≈ 0.37 R e x − 1 / 5 \delta/x \approx 0.37 Re_x^{-1/5} δ / x ≈ 0.37 R e x − 1/5
Common mistake Forgettingv ≠ 0 v\neq 0 v = 0 Why it feels right: "the flow is parallel to the plate."
The fix: continuity forces a small upward v v v v = 1 2 ν U ∞ / x ( η f ′ − f ) ≠ 0 v=\frac12\sqrt{\nu U_\infty/x}(\eta f'-f)\neq 0 v = 2 1 ν U ∞ / x ( η f ′ − f ) = 0
Common mistake Treating Blasius equation as having a clean analytic solution.Why it feels right: physics formulas usually do.
The fix: f ′ ′ ′ + 1 2 f f ′ ′ = 0 f'''+\frac12 ff''=0 f ′′′ + 2 1 f f ′′ = 0 no closed form ; f ′ ′ ( 0 ) = 0.332 f''(0)=0.332 f ′′ ( 0 ) = 0.332
Recall Feynman: explain to a 12-year-oldImagine running your hand flat just above a carpet while air blows past. Right at the carpet the air can't move — it's stuck (sticky rule). Far up, the air zooms freely. In between there's a thin "slow zone" that gets a little thicker as you go further along the carpet — but it thickens slower and slower , like a shadow stretching. Blasius figured out the exact shape of how fast the air speeds up as you climb away from the carpet, and found that this shape is always the same if you measure height the clever way. He also found how hard the air drags on the carpet — and that drag gets weaker the faster or longer the flow goes.
Mnemonic Remember the numbers"Five thick, three-thirty drag, six-six-four friction, one-three-two-eight total."
δ \delta δ 5.0 · f ′ ′ ( 0 ) f''(0) f ′′ ( 0 ) 0.332 · C f C_f C f 0.664 · C D C_D C D 1.328 .
Notice: 0.664 = 2 × 0.332 0.664 = 2\times0.332 0.664 = 2 × 0.332 1.328 = 2 × 0.664 1.328 = 2\times0.664 1.328 = 2 × 0.664 doubles .
What approximation lets Prandtl simplify Navier–Stokes near a wall? The layer is thin (
δ ≪ x \delta\ll x δ ≪ x ), so
∂ 2 u / ∂ x 2 \partial^2 u/\partial x^2 ∂ 2 u / ∂ x 2 is negligible vs
∂ 2 u / ∂ y 2 \partial^2 u/\partial y^2 ∂ 2 u / ∂ y 2 .
Define the Blasius similarity variable η \eta η η = y U ∞ / ( ν x ) \eta = y\sqrt{U_\infty/(\nu x)} η = y U ∞ / ( ν x ) .
What is the Blasius ODE and its BCs? f ′ ′ ′ + 1 2 f f ′ ′ = 0 f'''+\tfrac12 f f''=0 f ′′′ + 2 1 f f ′′ = 0 ;
f ( 0 ) = 0 , f ′ ( 0 ) = 0 , f ′ ( ∞ ) = 1 f(0)=0,\ f'(0)=0,\ f'(\infty)=1 f ( 0 ) = 0 , f ′ ( 0 ) = 0 , f ′ ( ∞ ) = 1 .
What does f ′ ( η ) f'(\eta) f ′ ( η ) The velocity ratio
u / U ∞ u/U_\infty u / U ∞ .
Why use a stream function ψ \psi ψ It satisfies continuity automatically (
u = ψ y , v = − ψ x u=\psi_y,\ v=-\psi_x u = ψ y , v = − ψ x ).
Value of f ′ ′ ( 0 ) f''(0) f ′′ ( 0 ) 0.332 0.332 0.332 ; sets the wall shear stress (velocity gradient at wall).
Boundary-layer thickness formula (99%)? δ = 5.0 x / R e x \delta = 5.0\,x/\sqrt{Re_x} δ = 5.0 x / R e x , i.e.
δ ∝ x \delta\propto\sqrt{x} δ ∝ x .
Local skin-friction coefficient? C f = 0.664 / R e x C_f = 0.664/\sqrt{Re_x} C f = 0.664/ R e x .
Total drag coefficient for one side of plate length L L L C D = 1.328 / R e L C_D = 1.328/\sqrt{Re_L} C D = 1.328/ R e L .
Validity limit of the Blasius solution? Laminar flow,
R e x ≲ 5 × 10 5 Re_x \lesssim 5\times10^5 R e x ≲ 5 × 1 0 5 .
How does δ \delta δ U ∞ U_\infty U ∞ δ ∝ 1 / U ∞ \delta\propto 1/\sqrt{U_\infty} δ ∝ 1/ U ∞ — faster flow ⇒ thinner layer.
Wall shear stress expression? τ w = 0.332 μ U ∞ U ∞ / ( ν x ) \tau_w = 0.332\,\mu U_\infty\sqrt{U_\infty/(\nu x)} τ w = 0.332 μ U ∞ U ∞ / ( ν x ) .
Prandtl boundary-layer eqn
No-slip and free-stream BCs
Scaling balance inertia vs viscosity
Layer thickness delta ~ sqrt of nu x over U
Stream function psi with f of eta
Blasius ODE 2f triple prime plus f f double prime equals 0
Intuition Hinglish mein samjhoSocho ek flat plate ke upar se hawa ya paani bah raha hai. Plate ki surface par "no-slip" rule lagta hai — matlab bilkul surface par fluid ki speed zero ho jaati hai, aur thodi door upar jaake free-stream speed U ∞ U_\infty U ∞ boundary layer kehte hain. Blasius solution isi layer ka exact (laminar) jawab deta hai: velocity ka profile kya hai, aur layer kitni moti hai.
Asli kamaal ye hai ki Prandtl ne pehle full Navier–Stokes ko simplify kiya (kyunki layer thin hai, x x x similarity variable η = y U ∞ / ν x \eta = y\sqrt{U_\infty/\nu x} η = y U ∞ / ν x y y y x x x same shape dikhta hai. Saari PDE ek single ODE ban jaati hai: f ′ ′ ′ + 1 2 f f ′ ′ = 0 f''' + \tfrac12 f f'' = 0 f ′′′ + 2 1 f f ′′ = 0 f ′ ′ ( 0 ) = 0.332 f''(0)=0.332 f ′′ ( 0 ) = 0.332
Yahin se saari important formulas nikalti hain. Layer ki thickness δ = 5.0 x / R e x \delta = 5.0\,x/\sqrt{Re_x} δ = 5.0 x / R e x δ \delta δ x x x C f = 0.664 / R e x C_f = 0.664/\sqrt{Re_x} C f = 0.664/ R e x C D = 1.328 / R e L C_D = 1.328/\sqrt{Re_L} C D = 1.328/ R e L
Kyun important hai? Aircraft wings, pipes, ships, turbine blades — har jagah drag aur heat transfer is boundary layer pe depend karta hai. Bas yaad rakho: Blasius sirf laminar flow (R e x < 5 × 10 5 Re_x < 5\times10^5 R e x < 5 × 1 0 5
