Boundary layer separation — adverse pressure gradient: Foundations

Before you can read the parent note, you must own each letter it throws at you. We'll meet them one at a time, each with a picture and a reason it exists.

The stage: what are $x$ and $y$ here?

Everything in this topic lives near a wall. We first need a coordinate frame glued to that wall, not to the room.

Figure 1 (below). A curved wall in white with the flow-direction axis $x$ (blue arrow) bending to follow the surface, and the height axis $y$ (yellow arrow) shooting straight out perpendicular to it. The shaded region below is the solid; the wall is the line $y = 0$ . Notice how $x$ never leaves the surface even though the wall curves — that is the whole point of this frame.

Figure — Boundary layer separation — adverse pressure gradient

The velocity components $u$ and $v$

A fluid particle moves in a direction. We split that motion into two arrows, one along each axis.

What is a streamline?

The word streamline is about to appear in the Bernoulli section, so let's build it now.

The boundary layer, its thickness $δ$ , and the edge

Before we can talk about "just outside the layer," we need to say precisely where the layer ends. That edge has a name and a symbol.

Figure 2 (below). The white wall at $y = 0$ with the red dashed edge curve $y = δ (x)$ that thickens as $x$ increases. At three stations, a blue velocity profile shows $u$ starting at $0$ on the wall and climbing to the free-stream value at the red dot on the edge; green arrows mark the along-wall direction and yellow arrows the uniform free stream $U (x)$ above the layer. Watch the red edge climb — that is $δ$ growing with $x$ .

Why pressure does not change across the layer ( $p = p (x)$ )

The parent note quietly assumes $p$ depends on $x$ only, not on height $y$ . That is not obvious — let's actually derive it.

Intuition The thin-layer scaling argument — piece by piece

Give the layer a length $L$ along $x$ and a tiny height $δ$ , with $δ ≪ L$ . Call the free-stream speed $U$ . Two facts drive everything:

The along-wall speed $u$ ranges over $\sim U$ across a distance $\sim L$ , so $\partial u / \partial x \sim U / L$ .
Mass conservation, $\partial u / \partial x + \partial v / \partial y = 0$ , then forces $\partial v / \partial y \sim U / L$ too. Since $v$ changes over the tiny height $δ$ , that means $v$ itself is only $\sim U δ / L$ — small, because $δ / L$ is small.

Now weigh each term in the $y$ -equation:

Inertia $ρ u \partial v / \partial x \sim ρ U \cdot (U δ / L) / L = ρ U^{2} δ / L^{2}$ .
Compare with the along-wall pressure term (from the $x$ -equation), which is order $ρ U^{2} / L$ .
Their ratio is $(ρ U^{2} δ / L^{2}) / (ρ U^{2} / L) = δ / L ≪ 1$ .

So every term that could bend the pressure in the $y$ -direction is smaller than the main flow terms by the factor $δ / L$ . Shrink the layer ( $δ / L \to 0$ ) and they vanish, leaving only $\frac{\partial p}{\partial y} \approx 0.$

The no-slip condition — why $u = 0$ at the wall

Slope and curvature: reading a velocity profile

Draw $u$ sideways against height $y$ : that curve is the velocity profile. Two things about its shape decide everything.

Figure 3 (below). Two velocity profiles ( $u$ horizontal, height $y$ vertical). Left (green): a favourable profile — steep slope at the wall and bending only one way, no inflection, healthy and attached. Right (red): an adverse profile — the wall slope has nearly flattened to zero and a yellow dot marks the inflection point where the bend switches direction. Compare the wall slopes: the green one is steep (strong forward pull); the red one is nearly flat (about to reverse).

The Greek letters: $ρ$ , $μ$ , $ν$

Newton's law of friction: why $τ_{w} = μ \partial u / \partial y$

The wall shear formula is not a definition dropped from the sky — it rests on one physical assumption about the fluid.

Pressure and its gradient

The bridge you already have: Bernoulli

We keep saying " $U (x) ↓$ forces $p (x) ↑$ ." That link comes from Bernoulli's equation — but only under specific conditions, and only in a specific form. Let's state both. (Recall from above that a streamline is a curve the fluid flows along but never across.)

How the pieces feed the topic

The diagram below is a dependency map: each box is one idea from this page, and an arrow " $A \to B$ " means you must own $A$ before $B$ makes sense. Read it top to bottom.

Equipment checklist

Cover the right side and recite each before reading the parent note.

What do

x

and

y

measure in this topic?

x

runs along the surface (flow direction);

y

runs perpendicular, away from the wall, with the wall at

y = 0

What is the difference between

u

and

U

u

is the actual along-wall speed inside the layer (0 at the wall up to

U

);

U (x)

is the fast free-stream speed at and beyond the edge

y = δ

What is a streamline?

A curve everywhere tangent to the velocity; fluid flows along it but never crosses it, so energy bookkeeping (Bernoulli) is done along one streamline.

What is the boundary-layer thickness

δ (x)

The height where

u

reaches

99%

U

; below it friction matters, above it is free stream, and

δ

grows with

x

Why can we write

p = p (x)

with no

y

-dependence?

In the

y

-momentum equation every term is smaller than the main flow by the factor

δ / L ≪ 1

, so

\partial p / \partial y \approx 0

; the outer pressure stamps straight down onto the wall.

What does the no-slip condition state?

The fluid touching a stationary wall is frozen:

u = 0

and

v = 0

y = 0

Why write

\partial u / \partial y

with a curly

\partial

instead of

d

u

depends on both

x

and

y

; the partial means "change with height

y

only, holding

x

fixed."

What does

\frac{\partial ^{2} u}{\partial y ^{2}}

physically represent?

The curvature (bend) of the velocity profile — how its slope changes as you rise; a sign change gives an inflection point.

Relate

μ

ρ

and

ν

ν = μ / ρ

, i.e.

μ = ρ ν

;

μ

is stickiness,

ρ

is density,

ν

is momentum diffusivity.

State Newton's law of friction and the assumption behind it.

τ = μ \partial u / \partial y

, valid for a Newtonian fluid where shear stress is linearly proportional to the velocity slope (air, water).

Write the wall shear stress and say when it signals separation.

τ_{w} = μ \partial u / \partial y ∣_{y = 0}

; separation begins where

τ_{w} = 0

, with backflow just downstream.

What makes a pressure gradient "adverse"?

d p / d x > 0

— pressure rising downstream, decelerating the flow (

d U / d x < 0

State Bernoulli's validity conditions and its differential form.

Steady, incompressible, inviscid along a streamline; differentiating gives

d p / d x = - ρ U d U / d x

, so

U ↓\Rightarrow p ↑

Related builds: Boundary layer theory · Reynolds number & transition · Drag — form vs skin friction · Flow over a cylinder and sphere · Diffusers and nozzles · Stall on an aerofoil.