Fluid Flow & Bernoulli · Guide · glidepathscience.org
GuideFluid Flow & Bernoulli
Contents
This Guide explains the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round, in the order a textbook would take them. Part 1 prepares the Intro track and Part 2 the Advanced track; each takes about a quarter of an hour to read. Each section ends with a worked example that is not one of the sim’s own problems.
Part 1 · The essentials
What moves a fluid
Pressure & Buoyancy studied fluids at rest. Water in a pipe, air over a wing and blood in an artery move because the pressure behind each small parcel of fluid differs from the pressure ahead of it: a pressure difference is a net force, and it accelerates the parcel. Gravity adds its own push wherever the fluid climbs or falls.
This lab looks at steady flow, in which the speed and the pressure at each point do not change with time. A steady flow keeps two books: one of mass (whatever enters a pipe each second must leave it), which gives the continuity equation, and one of energy (the work the pressure does on a parcel becomes kinetic energy and height), which gives Bernoulli’s equation. Almost everything that follows comes from reading the two together, and from noticing where real fluids break the second.
Flow rate and continuity
The first question about any flow is how much passes. The flow rate $Q$ is the volume that crosses a section of the pipe each second. If the fluid moves with an average speed $v$ through a cross-section of area $A$, a slab of length $v$ passes each second, so
$Q = V/t = Av$
Liters per second or per minute are easier to picture than cubic meters per second: a kitchen faucet runs at about 6 L/min. Flow rate is not speed. A wide, slow river and a narrow, fast hose can carry the same flow, and a common slip is to treat the two as the same thing or to use a diameter where the area needs a radius.
Now follow the flow through a pipe that changes size. A liquid’s density barely changes, and in steady flow no fluid can pile up between two sections, so the same $Q$ must pass both. This is the continuity equation:
$A_1v_1 = A_2v_2$
For any fluid, liquid or gas, it is the mass per second that is the same, $ρ_1A_1v_1 = ρ_2A_2v_2$; for a liquid, $ρ$ cancels. Because the area of a round pipe grows as its diameter squared, a narrowing speeds the flow up as the square of the diameter ratio. The intuition that a narrowing “holds the water back” has it backward: every liter that enters must leave, and in a narrow part it can do so only by moving faster. A narrowing can lower the flow a pump or a reservoir drives, by adding resistance (a narrowed artery does), but whatever that flow is, the narrow part carries it faster.
Something must push the water to make it go faster through the nozzle: a pressure difference, the subject of the next section.
Bernoulli’s equation
When water speeds up into a narrowing, each parcel gains kinetic energy, and only the pushes of the fluid behind and ahead of it (and gravity) can supply it. So the pressure behind a parcel that is speeding up must exceed the pressure ahead: where a level flow is fast, its pressure is low. This is the opposite of the idea that a squeezed fluid is at a higher pressure, and it is what a Venturi tube’s columns show.
To make this exact, follow a small volume $ΔV$ of fluid along a streamline from point 1 to point 2. The fluid behind it does work $p_1ΔV$ pushing it in, and the fluid ahead takes away $p_2ΔV$. The difference changes the parcel’s kinetic energy, $½ρΔV(v_2^2 - v_1^2)$, and its gravitational energy, $ρΔVg(h_2 - h_1)$. Divide by $ΔV$ and collect the terms for each point:
$p_1 + ½ρv_1^2 + ρgh_1 = p_2 + ½ρv_2^2 + ρgh_2$
This is Bernoulli’s equation: the work–energy theorem per unit volume. Each term is an energy per volume (a pascal is a joule per cubic meter): $p$ is the pressure’s capacity to do work, $½ρv^2$ the kinetic energy and $ρgh$ the gravitational energy. Along a streamline of a steady flow without friction their sum is constant. The bars in the sim show the three terms at each tap; their total stays level while pressure is traded for speed or height.
If the speed does not change, as in a pipe of one size, only height trades with pressure: about 9.8 kPa for every meter the water climbs. If nothing moves at all, the equation becomes $p = p_0 + ρgh$, the pressure with depth of Pressure & Buoyancy. The usual slips are to square the difference of the speeds, $(v_2 - v_1)^2$, instead of taking the difference of their squares, and to drop the ½.
Read in reverse, Bernoulli’s equation turns a measured pressure difference into a speed. That is how flow meters and airspeed indicators work.
Venturi meters and pitot tubes
A Venturi meter is a short narrowing put into a pipe on purpose, with a pressure tap on the wide part and one in the throat. The columns over the two taps differ by $Δh$, so the pressure difference is $Δp = ρgΔh$. Bernoulli gives $Δp = ½ρ(v_2^2 - v_1^2)$, continuity gives $v_2 = v_1(A_1/A_2)$, and together
$v_1 = \sqrt{2gΔh/((A_1/A_2)^2 - 1)}$, $Q = A_1v_1$
The density cancels when the columns hold the same liquid as the pipe, so a water meter needs no knowledge of the water. Two slips are common: dropping the −1, which treats the water in the wide part as still, and using the ratio of the diameters where the ratio of the areas belongs. A good Venturi gets back almost all of its pressure in the long cone after the throat; a cheaper orifice plate loses most of its reading for good.
A pitot tube measures speed at a point. It is a small tube whose open mouth faces into the flow. The fluid arriving at the mouth is brought to rest there, at the stagnation point, and its $½ρv^2$ becomes extra pressure: the mouth feels the total pressure, $p + ½ρv^2$. Holes in the side of the tube, where the flow passes undisturbed, feel only the static pressure $p$. The difference between them gives the speed:
$v = \sqrt{2Δp/ρ}$
Aircraft carry a pitot-static probe for this. Here $ρ$ is the density of the moving fluid, air for an airplane, even when $Δp$ is read on a water manometer.
Jets, and Torricelli’s law
A free jet, a stream of fluid that has left its pipe or tank and moves through the open air, is at the pressure of the air around it, however fast it moves. This is a free-surface approximation: the jet’s surface is open to the air, and once the jet has narrowed just past the hole its streamlines run side by side, so nothing holds its inside at another pressure (surface tension adds about 15 Pa in a jet 1 cm across). Gravity then bends the jet, but every parcel falls freely, so the bend needs no pressure difference either.
Consider a wide tank open to the air, with a small hole a depth $h$ below the water’s surface. Follow a streamline from the surface, where the water is almost still and at the air’s pressure, to the jet just outside the hole, also at the air’s pressure. The pressures cancel in Bernoulli’s equation, and what is left is $ρgh = ½ρv^2$:
$v = \sqrt{2gh}$
This is Torricelli’s law: the water leaves as fast as if it had fallen freely through $h$, and its speed grows as the square root of the depth. Once out, the jet is a projectile. Leaving horizontally from a height $y$ above the floor, it takes $\sqrt{2y/g}$ to fall and lands a distance
$x = v\sqrt{2y/g}$
from the tank. With $v = \sqrt{2gh}$ this is $x = 2\sqrt{hy}$, and $g$ cancels. In a tank of depth $H$, a hole at height $y$ has $h = H - y$, so holes at $y$ and at $H - y$ land at the same spot, and a hole halfway up throws farthest.
Bernoulli and the wing, without the myth
The most famous application of Bernoulli’s equation is also the most often misexplained. A popular account says that the air over the curved top of a wing has farther to go, must meet the air going under it at the back edge, and so goes faster. The conclusion, faster air and lower pressure above, is true; the reasoning is not, since nothing requires the two parcels to meet again. Smoke released in pulses in a wind tunnel shows the air over a lifting wing arriving well before the air underneath. And a symmetric wing at zero angle, where the two paths really are equal, gives no lift at all.
What lifts a wing is that it turns the passing air downward; the air pushes back on the wing, up. A curving flow needs lower pressure on the inside of its curve, so the air bending down over the top of the wing is at lower pressure than the air far away, and the air underneath at somewhat higher pressure. Bernoulli’s equation, along each streamline from the same undisturbed air upstream, then says that the air above moves faster and the air below more slowly. Speed and pressure go together; the turning of the flow sets both.
In the engine’s ideal flow over the 5° section the lower parcels trail by about a third of the time the free stream takes to cross one chord. Part 2 explains what sets this lead, the circulation, and how it fixes the lift.
Viscosity, Poiseuille’s law and the Reynolds number
Bernoulli’s equation leaves out friction. Real fluids have viscosity, $η$, a resistance to sliding past themselves: water’s is about 1.0 mPa·s, blood’s 3 to 4, honey’s thousands of times more. Fluid touching a wall does not slip, so in slow, smooth, laminar flow through a tube the speed across it is a parabola, zero at the wall.
Pushing fluid through a level tube of one size then takes a pressure difference, and the flow is proportional to it. For a tube of radius $r$ and length $l$, the resistance is
$R = 8ηl/(πr^4)$, $Q = Δp/R$
This is Poiseuille’s law, and the fourth power is its surprise: narrowing a tube cuts its area as $r^2$, and in a thinner tube more of the fluid is close to a wall and held back, which costs another $r^2$. It is tempting to argue that continuity keeps $Q$ the same and the fluid simply speeds up; but continuity only says that one $Q$ passes every section of a tube, while the pressure difference and the resistance decide how large it is.
Faster flow eventually breaks into swirls: it becomes turbulent, and the resistance rises steeply. Which regime a pipe is in depends on one dimensionless number, the Reynolds number,
$N_R = 2ρvr/η$
which compares the fluid’s momentum with its viscous stresses. In a pipe the flow is laminar below about 2,000, turbulent above about 3,000, and unstable in between, where it may switch back and forth. At a fixed flow rate, $vr$ goes as $1/r$, so narrowing a pipe raises $N_R$.
A narrowed artery
Plaque that narrows a short stretch of an artery is a constriction like a Venturi throat: by continuity the blood speeds up through it, and by Bernoulli its pressure there dips. In mmHg, the unit of blood pressure (1 mmHg = 133.3 Pa), the dip is $½ρ(v_2^2 - v_1^2)/133.3$ with blood’s density, about 1,060 kg/m³. Be careful with the units: a dip computed in pascals and typed into a box in mmHg is 133 times too large.
Ideal flow would give the dip back after the narrowing. A real artery recovers only part of it: the jet leaving the narrowing mixes into the slower blood and turns some of its energy into heat, and a long narrowing adds Poiseuille’s resistance. That lost pressure is the narrowing’s real cost (Part 2).
Part 2 · Going further
Newton along and across streamlines
Part 1 derived Bernoulli’s equation from work and energy. Newton’s second law gives it more directly, and it also shows what the equation does not say. Take a small parcel moving along a streamline, with $s$ the distance along it. Its acceleration along the streamline is $v\,dv/ds$, and the forces on it per unit volume are the pressure gradient and gravity:
$ρv\,dv/ds = -dp/ds - ρg\,dz/ds$
Integrating along the streamline gives $p + ½ρv^2 + ρgz$ = constant. Read in this form, the cause is clear: the pressure falling along the flow is what accelerates the parcel, and a pump or a reservoir upstream sets that pressure. Saying that “fast flow lowers the pressure” reads the equation backward.
The derivation also shows the fine print: the flow must be steady, inviscid and incompressible, and the two points must lie on the same streamline. A hair dryer’s jet is fast, yet its pressure is the room’s, because its streamlines began inside the dryer at a higher pressure. Comparing it with the still room air is comparing two different streamlines.
Newton’s second law across the streamlines says something new. A parcel following a curved streamline of radius $R$ has a centripetal acceleration $v^2/R$ toward the center of curvature, and the pressure gradient and gravity together supply it. With $n$ measured outward, away from that center, and $g_n$ gravity’s component along $n$:
$∂p/∂n = ρv^2/R + ρg_n$
In a straight stream ($R$ infinite) with $n$ pointing up, $g_n = -g$: the pressure falls by $ρg$ per meter of height, as in still water. In a horizontal bend gravity has no component along $n$, and the pressure gradient alone turns the flow:
$∂p/∂n = ρv^2/R$ (horizontal bend)
with the pressure rising away from the center of curvature. Measured from the pressure the still fluid would have at each height, the curve’s own part is $ρv^2/R$ in any plane; for air, the still air’s change is only 12 Pa per meter. The pressure is lowest on the inside of a curved flow. This is why air hugging a curved surface is at low pressure next to it, why the paper strip at your lip rises when a free jet curves over it, and why the pressure on the outside of a bend in a pipe is higher than on the inside.
Circulation, the Kutta condition and lift
In ideal, frictionless flow the equations round a body have many solutions, which differ in their circulation $Γ$, the net swirl of the flow round the body (the integral of the velocity along any loop enclosing it). A spinning cylinder shows the family plainly: adding circulation speeds the flow on one side and slows it on the other, and the two stagnation points slide toward each other until they meet, at $Γ = 4πUa$ for a cylinder of radius $a$ in a stream $U$. The asymmetry gives a sideways force, the Magnus effect of a spinning ball.
For every body in steady, inviscid, two-dimensional flow the lift per unit span is the Kutta–Joukowski theorem, $L′ = ρVΓ$, and the drag is exactly zero, whatever the shape: d’Alembert’s paradox. Real bodies drag because viscosity makes their boundary layers separate, which ideal flow cannot describe.
Which circulation does a real airfoil get? With too little, the ideal flow would have to turn round its sharp trailing edge from below, at an infinite speed, which real air will not do. As a wing starts moving, viscosity sheds a starting vortex from the edge and leaves the wing with just the circulation that makes the flow leave the trailing edge smoothly: the Kutta condition. It is not equal transit; with it, the air over the top arrives first.
Ideal flow never stalls; real wings stall past about 12 to 15°, when the flow separates (Flight Lab).
Airspeed, density and compressibility
An airspeed indicator is a pitot-static probe connected to a dial. The dial cannot know the air’s density, so it is marked as if the air were at sea level, $ρ_0$ = 1.225 kg/m³: it reads $\sqrt{2q/ρ_0}$, where $q = ½ρv^2$ is the dynamic pressure. In thinner air at altitude the same true airspeed gives a smaller $q$, so the dial reads low. The reading is the indicated airspeed, and
$V_I = V_T\sqrt{ρ/ρ_0}$
The square root is easy to lose, and so is the direction: the indicated airspeed is below the true one wherever the air is thinner than at sea level. Yet it is the right instrument to fly by: a wing’s lift also depends on $q$, so at a given weight, configuration (flaps and gear) and load factor an airplane stalls at the same indicated airspeed at every altitude, as long as the speed is low enough that compressibility and the instrument’s calibration errors are small.
At high speed air compresses as it is brought to rest at the probe’s mouth, and the probe reads more than $½ρv^2$. For subsonic flow of air (γ = 1.4) the isentropic relation gives
$(p_0 - p)/(½ρv^2) = [(1 + 0.2M^2)^{3.5} - 1]/(0.7M^2) ≈ 1 + M^2/4$
where $M$ is the Mach number. The excess is 2.3 % at Mach 0.3, the usual limit below which air is treated as incompressible, and grows quickly after that; the error in the speed is about half the error in the pressure. Fast aircraft therefore use the compressible formula.
The laminar profile, turbulent friction and head loss
Poiseuille’s law follows from a force balance on a cylinder of fluid of radius $s$ inside the tube: the pressure difference pushes it with $Δp\,πs^2$ and the viscous stress on its surface, $η\,du/ds$ over the area $2πsl$, holds it back. Setting them equal and integrating from the still wall inward gives the parabola
$u(s) = u_{max}(1 - s^2/R^2)$, $u_{max} = 2\bar v$
and integrating $u$ over the cross-section gives $Q = πR^4Δp/(8ηl)$. The center moves at twice the mean speed $\bar v = Q/(πR^2)$. The law holds only while the flow is laminar, so check $N_R$ before using it.
For any pipe flow, laminar or turbulent, engineers write the pressure lost to friction with the Darcy–Weisbach equation,
$Δp = f(L/D)½ρv^2$
where $f$ is the friction factor. In laminar flow Poiseuille’s law gives $f = 64/N_R$ exactly. In turbulent flow $f$ comes from experiment (Colebrook’s equation, the Moody chart): about 0.02 to 0.04 for ordinary pipes, more for rough ones. Because it changes so little, turbulent losses grow nearly as $v^2$; laminar ones grow as $v$.
The losses that never come back, and critical stenosis
A fast stream leaving a narrow passage abruptly into a wider one does not spread smoothly: the jet mixes with slower fluid in eddies. A momentum balance on the mixing region gives the pressure the expansion recovers, less than Bernoulli’s by the Borda–Carnot loss,
$Δp_{loss} = ½ρ(v_1 - v_2)^2$
where $v_1$ is the jet’s speed and $v_2$ the speed after the expansion. Note the square of the difference, not the difference of the squares: the latter is the recovery ideal flow would give. In a narrowed artery this loss, together with Poiseuille’s drop through the throat, is the pressure lost for good.
Cardiologists judge a narrowed heart valve the same way: Doppler ultrasound measures the jet’s speed, and the peak pressure gradient is
$ΔP = ½ρ(v_2^2 - v_1^2)$
For blood, ½ρ is 3.98 mmHg per (m/s)², so with the speed upstream left out this becomes the clinic’s rule, $ΔP ≈ 4v^2$ in mmHg with $v$ in m/s, fine when the speed upstream is under about 1 m/s. Part of it is recovered past the valve, so the net loss is smaller.
How much flow does a narrowing cost? It sits in series with the vascular bed beyond it, so with $a$ its Poiseuille resistance and $b$ its Borda–Carnot coefficient the driving pressure is $ΔP = (R_{bed} + a)Q + bQ^2$. At rest the bed’s resistance dominates, and the flow hardly changes until the narrowing is severe. Exercise dilates the bed and lowers its resistance, so the same narrowing costs more, which is why an exercise test can unmask it (Gould, 1978). The Artery Lab’s values are illustrative: it is an educational model, not a diagnostic tool.
Draining tanks, and the push of a jet
A tank of area $A$ draining through a hole of area $a$ loses water at the rate the jet carries it away. If the level falls slowly enough, Torricelli’s law holds at each moment (the flow is quasi-steady), and $A\,dh/dt = -C_da\sqrt{2gh}$. The square root of the depth then falls linearly in time, and a tank that starts at depth $H$ empties in
$t = (A/(C_da))\sqrt{2H/g}$
The level falls fastest at first, so it takes twice as long as the first second’s rate would suggest. $C_d$, the discharge coefficient, is the fraction of the ideal flow a real hole passes: past a sharp edge the jet contracts to a narrower neck, the vena contracta, and $C_d$ is about 0.61; a rounded inlet gives about 0.98.
A jet also pushes back on whatever launches it. Each second it carries away a momentum $ρQv$, which the nozzle must supply: $F = ρQv = ρAv^2$. With the nozzle pressure turned into the jet’s speed, $½ρv^2 = p$, this is $F = 2pA$, twice the pressure alone on the opening. The fire service’s rule for a smooth-bore tip, NR = 1.57 $d^2$ NP (pounds, inches, psi), is the same $2pA$.
The Reference collects the usual mistakes and every equation.
Reference
Common mistakes
- A diameter used as an area. Area goes as the diameter squared: halving a pipe’s diameter quarters its area. And $A = πr^2$ needs the radius.
- “Squeezed means higher pressure.” Fluid speeds up into a narrowing, so its pressure there is lower.
- Speed taken as the cause of low pressure. The pressure difference accelerates the fluid; Bernoulli relates the two along a streamline.
- Bernoulli used across streamlines. A free jet is at the surrounding pressure, however fast it is; compare points on one streamline.
- Equal transit. The air over a lifting wing arrives at the back edge first; nothing makes the parcels meet.
- Continuity taken to fix the flow. Continuity says the same $Q$ passes every section; the pressure and the resistance decide how large $Q$ is.
- The square root lost from indicated airspeed, $V_I = V_T\sqrt{ρ/ρ_0}$, or from Torricelli’s $\sqrt{2gh}$.
- The −1 dropped in the Venturi formula, which treats the wide part’s water as still.
- Pascals typed for mmHg. 1 mmHg = 133.3 Pa.
- Poiseuille in turbulent flow. Check $N_R$: above about 2,000 the flow is not laminar and the law fails.
The equation sheet
| Part 1 | |
|---|---|
| flow rate | $Q = V/t = Av$ |
| continuity | $A_1v_1 = A_2v_2$ |
| Bernoulli | $p_1 + ½ρv_1^2 + ρgh_1 = p_2 + ½ρv_2^2 + ρgh_2$ |
| columns | $Δp = ρgΔh$ |
| pitot | $v = \sqrt{2Δp/ρ}$ |
| Torricelli | $v = \sqrt{2gh}$ |
| the jet’s range | $x = v\sqrt{2y/g}$ |
| Poiseuille | $R = 8ηl/(πr^4)$, $Q = Δp/R$ |
| Reynolds | $N_R = 2ρvr/η$ |
| Part 2 | |
|---|---|
| curved flow | $∂p/∂n = ρv^2/R + ρg_n$ |
| a horizontal bend | $∂p/∂n = ρv^2/R$ |
| airspeed | $V_I = V_T\sqrt{ρ/ρ_0}$ |
| laminar profile | $u_{max} = 2\bar v$ |
| friction | $Δp = f(L/D)½ρv^2$, $f = 64/N_R$ (laminar) |
| Borda–Carnot | $Δp_{loss} = ½ρ(v_1 - v_2)^2$ |
| valve gradient | $ΔP = ½ρ(v_2^2 - v_1^2)$ |
| draining | $t = (A/(C_da))\sqrt{2H/g}$ |
| a jet’s push | $F = ρQv = 2pA$ |
| lift | $L′ = ρVΓ$ |
Symbols: $Q$ flow rate; $A$ area; $v$ speed; $p$ pressure (gauge unless “absolute”); $ρ$ density (water 998.2 kg/m³ at 20 °C, blood 1,060 kg/m³, sea-level air $ρ_0$ = 1.225 kg/m³); $h$ height or depth; $g$ = 9.81 m/s²; $η$ viscosity; $r$, $R$ radius (in $∂p/∂n$, the streamline’s radius of curvature, $n$ pointing away from its center and $g_n$ gravity’s component along $n$); $N_R$ Reynolds number; $f$ friction factor; $L$, $D$ a pipe’s length and diameter; $C_d$ discharge coefficient; $Γ$ circulation; 1 mmHg = 133.3 Pa.
Sources and further reading
- OpenStax, University Physics Volume 1 (Ling, Sanny and Moebs), chapter 14, “Fluid Mechanics,” §14.5–14.7, CC BY-NC-SA 4.0: the Intro track and Part 1 follow its order and notation, in this lab’s own words. Fluids at rest (§14.1–14.4) belong to Pressure & Buoyancy.
- H. Babinsky, “How do wings work?”, Physics Education 38, 497 (2003); NASA Glenn Research Center’s pages on incorrect theories of lift; D. McLean, Understanding Aerodynamics (Wiley, 2012).
- F. M. White, Fluid Mechanics (McGraw-Hill), for friction, minor losses and meters; K. L. Gould (1978) on critical stenosis. The Field Manual (H) describes the model and what it leaves out.
Fluid Flow & Bernoulli v1.2 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org