From Pressure to Bernoulli: Why Does Fluid Flow?

A balloon inflates, stretching its rubber membrane; the deeper the water, the more noticeable the pressure near your ears; a pipe narrows, and the flow speeds up. These seem like three separate things—gas, still water, and flowing water—but behind them lies a single chain: molecules hitting walls create pressure, pressure differences drive fluids, and flow conditions link pressure and velocity. Following this chain, the hydrostatic equation, Bernoulli's equation, and wing lift are no longer three isolated pieces of knowledge.

Pressure is a local quantity, not a push in one direction

Pressure is defined as the normal force per unit area:

$p=\frac{F_n}{S}$

At a point in space, take a sufficiently small surface and examine the ratio of the force perpendicular to it to its area; as the area shrinks, you obtain the pressure at that point. Pressure is a scalar; force is a vector. Pressure describes the local intensity of force, and its direction is determined by the normal of the surface it acts on. Pressure itself cannot be said to be "a force pointing in some direction."

In a stationary fluid, the pressure at a given point is the same in all directions. This does not mean the pressure is the same everywhere in the fluid: the deeper below the surface, the greater the pressure usually is. It only means that no direction is favored at a single point. If a stationary fluid had a sustained shear stress internally, it would deform and flow, so in static equilibrium the shear stress is zero; a viscous fluid in motion, however, can have shear stress.

Macroscopically, we see how much force a surface experiences; to find its origin, zoom in on the gas molecules: each collision with the wall exchanges only a little momentum, but the average effect of dense collisions steadily remains on the wall.

Where does gas pressure come from?

Treat a gas as a large number of molecules in random motion. When a molecule hits the container wall, its normal momentum changes; the rate of change of momentum is the force on the wall. The resultant force per unit area from all collisions manifests macroscopically as pressure.

For an ideal gas with molecular mass$m$and number density$n$, kinetic theory gives

$p=\frac{1}{3}nm\langle c^2\rangle$

where$c$is the thermal speed of molecules relative to the gas's overall average motion, and the angle brackets denote averaging over many molecules. For an ideal gas we also have

$p=nk_{\mathrm B}T$

These two equations link macroscopic pressure, microscopic motion, and temperature: at constant molecular number density, the higher the temperature, the more vigorous the thermal motion, the greater the average momentum transfer from wall collisions, and the greater the pressure.

Here we must distinguish "bulk motion" from "thermal motion." If the gas as a whole translates at velocity$\mathbf V$, the velocity of an individual molecule can be written as$\mathbf V+\mathbf c$. Bulk translation itself is not enhanced thermal motion, nor does it necessarily change the pressure. A sealed gas box on a uniformly moving train will not have its internal pressure change out of nowhere just because we switch to a ground reference frame, as long as the temperature and density inside remain unchanged. What truly determines thermal pressure is the random part relative to the gas's bulk motion, not the gas's bulk velocity relative to the ground.

This explains what determines the pressure at a location, but not yet why it varies in space. Now zoom back from molecules to a small parcel of fluid: at rest, how do the pressures on its top and bottom surfaces support its weight?

Deriving how pressure varies with height from static equilibrium

Take a thin layer of stationary fluid with cross-sectional area$A$and thickness$dz$, with the$z$axis pointing vertically upward. The bottom surface of the layer experiences an upward pressure force$p(z)A$, the top surface a downward pressure force$p(z+dz)A$, and gravity acts downward with magnitude$\rho gA dz$. Static equilibrium requires the net vertical force to be zero:

$p(z)A-p(z+dz)A-\rho gA dz=0$

Dividing both sides by$A dz$and taking the layer thickness to zero gives

$\frac{dp}{dz}=-\rho g$

This is the hydrostatic equation. It is not the empirical rule "deeper means higher pressure," but the result of Newton's second law applied to a stationary fluid element: the pressure below must be greater than above to support the weight of this layer of fluid.

If the fluid density is approximately constant and$g$can also be treated as constant, integrating from height$z_0$to$z$gives

$p(z)-p(z_0)=-\rho g(z-z_0)$

Therefore, at depth$h$below the liquid surface, the pressure is$p=p_0+\rho gh$. The pressure in a liquid increases linearly with depth, relying on the condition that density is approximately constant.

In a liquid, density hardly changes with depth, so each additional segment downward adds roughly the same weight of water, and the pressure increases linearly. Air, however, changes density with pressure; the hydrostatic equation still holds, but the result is no longer a straight line.

Why atmospheric pressure is not a simple linear function

Air also obeys the hydrostatic equation. The difference is that air is compressible: the higher up, the lower the pressure, and the lower the air density usually is, so the entire atmosphere cannot be treated as a liquid of constant density.

First take a simplified model: the atmosphere is in hydrostatic equilibrium, air can be treated as an ideal gas, and the temperature$T$and gravitational acceleration$g$are approximately constant over the range considered. The ideal gas equation of state is written as$p=\rho R_sT$, where$R_s$is the specific gas constant of air. Substituting into the hydrostatic equation:

$\frac{dp}{dz}=-\frac{g}{R_sT}p$

Separating variables and integrating gives

$p(z)=p_0\exp\left(-\frac{z-z_0}{H}\right),\qquad H=\frac{R_sT}{g}$

Thus, in the isothermal ideal gas model, pressure varies exponentially with height. The exponential form is not because "air experiences another kind of force," but because density changes along with pressure: the lower the pressure, the less air mass needs to be supported over the same small height increment, and the smaller the drop in pressure becomes.

The temperature of the real atmosphere changes with height, and gravity is not strictly the same everywhere, so this isothermal exponential formula cannot be taken as an exact atmospheric model at all heights. Its value lies in showing that the linear relationship for liquids comes from approximately constant density; gas density changes with state, so the pressure profile must be solved together with the equation of state. If compared near the same surface density, the slope of the exponential curve at the start is still determined by the hydrostatic equation; it later flattens because the air above becomes thinner, not because "exponential decrease must be faster than a straight line."

The same pressure gradient exactly supports the weight of the fluid when at rest; once flow is established, the unbalanced part becomes the net force accelerating fluid parcels.

How a pressure difference accelerates fluid

For a fluid in motion, if the pressure is high at one place and low at another, the pressure forces on the two ends of a small volume no longer cancel. The net pressure force points toward the lower pressure, so the fluid accelerates. For an ideal fluid neglecting viscosity and subject only to body forces such as gravity, the equation of motion is

$\rho\frac{D\mathbf v}{Dt}=-\nabla p+\rho\mathbf g$

The left side is the mass density of the fluid element times acceleration; the right side includes the pressure gradient force and gravity. In a local flow without gravitational effects, the pressure gradient force points in the direction of decreasing pressure. When viscosity cannot be neglected, a viscous stress term must be added to the equation of motion.

Note the levels of causal statement: before flow is established, a pressure difference provides the net force that accelerates the fluid; in a steady flow, the pressure distribution, velocity distribution, and boundary conditions together form a solution. At that point, one cannot single out either "high velocity" or "low pressure" and treat it as a one-way causal law valid for all flows.

The pressure difference explains why fluid accelerates; once the flow is established, another question arises: how do pressure energy, kinetic energy, and gravitational potential energy convert into one another? Integrating the equation of motion along a streamline gives Bernoulli's equation. This energy account has clear conditions of applicability; let us write out the account first.

What the Bernoulli equation says, and what it does not say

For steady, incompressible flow with negligible viscosity, along the same streamline it can be obtained by integrating the equation of motion

$p+\frac{1}{2}\rho v^2+\rho gh=C$

The three terms are respectively the pressure energy, kinetic energy, and gravitational potential energy per unit volume of fluid. If a pump or turbine does work in the flow, if energy losses are significant, or if density changes cannot be neglected, a more general energy relation must be used instead. To extend the same constant to different streamlines, additional conditions are needed, such as irrotational flow.

At the same height and on the same streamline, the equation gives: where the speed is greater, the static pressure is usually lower. This is a result of the energy bookkeeping. The Bernoulli equation itself does not automatically tell us what the velocity field is; one must first know what boundary constraints the fluid is subject to and how it flows around objects in order to determine how velocity and pressure are distributed. Treating "higher speed, lower pressure" directly as a slogan that explains all flows omits both the range of validity of the equation and the order in which it is solved.

The difficulty with an airfoil lies precisely here: Bernoulli can convert a known velocity distribution into pressure, but it cannot by itself calculate how the airflow flows around the airfoil. If "the upper surface path is longer" is taken as the reason the speed increases, then the flow field that is actually to be determined has been skipped over.

Airfoil lift: it is not that the air must arrive at the same time

A common claim is: air traveling over the upper surface of an airfoil has a longer path, so it must accelerate in order to reach the trailing edge at the same time as the air on the lower surface. This "equal transit time" condition does not come from the laws of fluid mechanics, and there is no reason the actual flow must satisfy it, so it cannot be used to explain airfoil lift.

The shape of the airfoil, the angle of attack, the freestream velocity, viscosity, and the upstream and downstream boundaries together determine the surrounding flow field. For a typical subsonic airfoil, a suitable angle of attack produces a velocity and pressure distribution around the airfoil; the local speed on the upper surface is often greater and the static pressure lower, but lift comes from the resultant of the pressure distribution over the entire airfoil surface, not merely from the upper surface "sucking" the airfoil. The Bernoulli equation can be used to relate velocity and pressure once the flow field is known, but it cannot by itself deduce the flow field from "the upper surface path is longer."

Another way to look at it is to track the momentum of the air: the airfoil deflects the surrounding air downward as a whole, and the air's momentum changes downward; by conservation of momentum and Newton's third law, the air exerts an upward force on the airfoil. Integrating the pressure over the airfoil surface gives the resultant force on the airfoil; observing the momentum change of the air far away also allows the corresponding lift to be calculated. The two statements describe different aspects of the same flow solution; it is not Bernoulli and Newton competing with each other.

Bernoulli gives one way to calculate the pressure on the airfoil surface, while the momentum viewpoint tracks as a whole how the air is deflected. Both describe the same flow: one looks at the force on the airfoil surface, the other at the momentum change of the air far away.

From molecular collisions to airfoil lift

Push the scales up layer by layer: molecular collisions produce pressure, and spatial pressure differences determine the force on a fluid element; hydrostatic equilibrium gives the hydrostatic equation, and flow is governed by Newton's equations. If the conditions of steadiness, incompressibility, and negligible viscosity are further satisfied, the Bernoulli equation writes pressure, velocity, and height into the same energy account. Airfoil lift comes from the resultant pressure over the entire surface and also corresponds to the momentum change when the air is deflected downward.

From gas molecules striking a wall, to the atmosphere thinning with altitude, to an airfoil changing the airflow, the subject of discussion is always how a fluid is acted on by forces and how it moves. Grasping this order means the formulas are not merely conclusions to be memorized, but tools that can be derived step by step along the model and its conditions.

References: the fluid mechanics chapter of OpenStax "University Physics"; NASA Glenn Research Center explanations of the Bernoulli equation and airfoil forces; derivations of the isothermal atmosphere model from MIT OpenCourseWare and others.