Find the question you need
Choose a question or read the twelve chapters in order. Teaching chapters remain open; the complete grouped index is expandable.
What the meter really tells us · Derive the two profiles · Calculate momentum and energy · Interpret real engineering evidence · Practise and read further
A litre per second tells us how much liquid moves through a pipe—but it does not describe where the liquid is moving fastest. Two velocity profiles can share the same discharge and still carry different momentum across the same section. This guide derives that difference and explains when it matters.
Chapter index
What the meter really tells us · Chapters 1–2
1. The surprising limit of a flow-meter reading · 2. Define the same flow rate without hiding assumptions
Derive the two profiles · Chapters 3–5
3. Build an ideal uniform reference profile · 4. Build a fully developed laminar parabolic profile · 5. Verify that the parabolic profile has exactly the same discharge
Calculate momentum and energy · Chapters 6–8
6. Calculate the second moment and the 4/3 momentum factor · 7. Work a unit-consistent numerical example · 8. Do not confuse momentum correction with kinetic-energy correction
Interpret real engineering evidence · Chapters 9–10
9. Connect the flux to a complete force balance · 10. Understand measurement and profile uncertainty
Practise and read further · Chapters 11–12
11. Frequently asked questions and three practice checks · 12. Use the result, then follow the next learning route
A pipe-flow meter may report a single volumetric discharge, such as one litre per second. That is a valuable statement about how much volume crosses a pipe section each second. It does not, by itself, tell us how rapidly every portion of the liquid is moving or how much momentum is crossing the section.
Imagine two idealised velocity fields in an identical round pipe. Both move the same amount of the same liquid through the same area per second. In the first, every part of the cross-section moves at the same speed. In the second, flow is fastest near the centre and approaches zero at the wall. The first is a useful mathematical comparison; the second resembles fully developed laminar flow of a Newtonian fluid.
The important distinction is between a first moment of the velocity profile—the average needed for discharge—and a second moment involving velocity squared, needed for momentum flux. Knowing the first does not generally determine the second.
This article derives the difference, works a numerical case with coherent units, explains what the calculation does and does not imply about pipe forces, and then tests how a changed profile changes the answer. All numbers are stipulated classroom quantities, not field measurements.
Take a straight circular cross-section of radius R and area A = πR². Let the positive axial velocity at radial coordinate r be u(r). For a steady, one-directional, axisymmetric flow, volumetric flow rate is Q = ∫Au dA, and cross-sectional mean speed is U = Q/A.
An annular ring of radius r and thickness dr has area dA = 2πr dr. This matters: equal increments of r are not equal areas. Rings close to the outer wall occupy more area than equally thick rings near the centre. Simply averaging a few centre-to-wall velocity values without area weights can produce the wrong mean.
We stipulate constant density ρ, a fixed circular geometry, no reverse axial flow and a steady snapshot at the section. These conditions let us compare profiles using clean one-dimensional expressions. A real flow can also have turbulent fluctuations, swirl, compressibility, temperature gradients or three-dimensional effects requiring more careful treatment.
On both profiles, choose the same U and hence the same Q = AU. When we later compare their momentum fluxes, do not quietly change the liquid density, the cross-sectional area or the observation boundary. Those changes would answer a different question.
The uniform reference takes u(r) = U everywhere across the circular area. Its flow rate is ∫U dA = UA, so it satisfies the specified mean by construction.
The ideal convective transport of axial momentum through that section is ρ∫Au²dA. In the uniform case this becomes ρAU². The result has dimensions kg·m/s², the same as a newton; it represents the rate at which axial momentum passes through the plane, not automatically a measured reaction force on the pipe.
For a viscous liquid obeying no-slip at a stationary wall, a perfectly flat velocity profile right up to the boundary cannot represent fully developed pipe flow. We use it as a reference or idealised nearly flat core profile, not as a second physically attainable steady laminar solution under identical boundary conditions.
This honesty matters. An instructive comparison can be mathematically valid without claiming that the two profiles coexist as unrestricted solutions to the same apparatus. When applying the momentum equation to an actual fitting, the inlet and outlet profiles must match the physical flow and boundary conditions.
For a steady, fully developed, laminar flow of an incompressible Newtonian fluid through a straight circular pipe, the classical Poiseuille profile is u(r) = 2U[1 − (r/R)²]. The axial speed is 2U at the centre and zero at the stationary wall.
The maximum being twice the mean is not a general rule for every pipe. It follows from this particular parabolic model. OpenStax University Physics explains Poiseuille flow, the role of viscosity and the no-slip boundary. The MIT fluid mechanics lecture sequence provides broader conservation and velocity-profile context.
Set s = r/R to make the arithmetic clearer. At s=0, u=2U. At s=0.5, u=1.5U. At s=0.8, u=0.72U. At s=1, u=0. The centre region is faster than the section mean; flow near the wall is slower.
These pointwise values are not four equally weighted portions of the area. The cross-sectional mean is obtained by integrating rings. That distinction is central to understanding why an apparently simple average can go astray in fluid mechanics.
CHAPTER 5 OF 12
5. Verify that the parabolic profile has exactly the same discharge
Use dA = 2πr dr and substitute s = r/R. Dividing by pipe area, the area average is (1/A)∫u dA = 2∫₀¹u(s)s ds. The factor 2s represents the relative ring area.
For u(s)=2U(1−s²), the mean becomes 4U∫₀¹(s−s³)ds = 4U(1/2−1/4) = U. Consequently Q=AU, exactly the same as the uniform reference.
The equality has not been achieved by inventing extra water in the centre. A higher central speed is offset by reduced speed closer to the wall. Conservation of volume transport is satisfied because every ring contributes its proper area and speed.
This verification is worth practising before introducing a correction coefficient. If a velocity profile claimed to have mean U does not integrate to U, then either its scale factor is wrong or the averaging measure has been mishandled. All the later comparisons would inherit that error.
Convective axial momentum flux is M = ρ∫Au²dA. Unlike discharge, it squares the local velocity before averaging. For the parabolic profile, (1/A)∫u²dA = 2∫₀¹[2U(1−s²)]²s ds.
Expand the squared profile: the expression equals 8U²∫₀¹(s − 2s³ + s⁵)ds = 8U²(1/2 − 1/2 + 1/6) = (4/3)U². Thus M = (4/3)ρAU².
Define the momentum correction coefficient β = M/(ρAU²). It equals 1 for the uniform reference and 4/3 for the ideal fully developed laminar circular-pipe profile. The two fields share mean speed and discharge, yet the laminar profile transports one-third more axial momentum through the same section than the uniform reference.
The fluid-flow textbook discussion indexed by ScienceDirect gives the same 4/3 result for Newtonian laminar pipe flow. Our equations explain why: velocities above the average contribute disproportionately when squared.
| Quantity | Uniform reference | Parabolic laminar |
|---|---|---|
| Area and mean speed | A and U | A and U |
| Discharge Q | AU | AU |
| Centreline speed | U | 2U |
| Momentum correction β | 1 | 4/3 |
| Convective momentum flux | ρAU² | (4/3)ρAU² |
| Kinetic-energy correction α | 1 | 2 |
Choose a hypothetical high-viscosity Newtonian liquid with density ρ = 1,000 kg/m³, cross-sectional area A = 0.010 m² and mean speed U = 0.10 m/s. The discharge is Q = AU = 0.0010 m³/s, equivalent to 1 litre per second. Because A=πR², the radius is about 0.0564 m.
The uniform reference’s convective momentum flux is ρAU² = 1,000×0.010×0.10² = 0.100 N. Under the parabolic laminar model it is (4/3)×0.100 = 0.1333 N. The difference in the fluxes is about 0.0333 N across the stipulated section.
To keep the laminar label physically coherent, also stipulate dynamic viscosity μ = 10 Pa·s. The nominal Reynolds number based on pipe diameter is Re = ρU(2R)/μ ≈ 1.13, comfortably within a very low-Reynolds-number viscous regime. These values are an invented consistency check, not a proposed pipeline design.
Note the units and interpretation. A value in newtons for a momentum flux term does not imply that a pressure gauge reads 0.1333 N or that the tube support feels exactly that force. To infer a net force, account for all momentum fluxes and external forces over a defined control volume.
CHAPTER 8 OF 12
8. Do not confuse momentum correction with kinetic-energy correction
Another frequently quoted coefficient is the kinetic-energy correction factor α. The energy flux associated with axial kinetic motion contains a velocity-cubed integral, proportional to ∫u³dA. It therefore weights faster portions of the cross-section even more strongly than the momentum integral does.
For an ideal uniform profile, α = (1/AU³)∫u³dA = 1. For the same parabolic pipe profile, α = 2. The momentum factor β remains 4/3. Memorising these two numbers without remembering which moment each represents invites an error in a Bernoulli-type or momentum-balance calculation.
Writing the weights out helps: discharge uses u, momentum flux uses u², and kinetic-energy transport uses u³. All three involve integration over the cross-section, but they are not interchangeable. Engineering LibreTexts on fluid energy conservation explores why profile assumptions affect averaged fluid quantities.
When a textbook uses different symbols for coefficients, read the definitions rather than assuming a Greek letter has a universal meaning. A proper report states both the integral and the reference mean speed. The physical quantity, not the memorised fraction, should drive the choice.
Newton’s second law in control-volume form links the net external force on fluid inside a defined region to the difference between outgoing and incoming momentum fluxes, plus any storage of momentum when conditions are unsteady. At a real bend or nozzle, pressure forces, wall stresses, gravity and momentum directions can all matter.
Suppose the inlet and outlet each have the same measured discharge. That does not make their vector momentum fluxes identical if area, direction or velocity profiles differ. Conversely, observing a 4/3 correction at one section is not enough to calculate the pipe-anchor force without information about the other boundaries.
Our 0.0333 N numerical difference is a comparison of two alternative section flux models, not a directly measured net thrust. In a real pipe, a uniform-looking entrance profile develops toward a viscous profile as the liquid travels downstream, and the pressure field and shear stresses participate in the momentum balance.
This distinction is especially useful in engineering education. The impulse–momentum method from MIT Engineering Mechanics II requires careful accounting of what crosses the control surface and which forces act. Begin by drawing the control volume, assigning outward normals and stating whether the flow is steady before inserting a profile coefficient.
A volumetric flow-meter reading normally tells us total discharge under its calibration and installation conditions. It generally does not provide a complete radial velocity distribution. The discharge can therefore be well measured while a derived momentum flux remains model dependent.
An actual velocity profile may be obtained through suitable instruments or flow simulations, but methods differ in spatial resolution, access, uncertainty and assumptions. Near-wall measurements can be especially demanding. If the profile is unknown, the correction factor should not be treated as perfectly known merely because Q is reported to many digits.
Furthermore, fully developed laminar flow is a particular regime. Entrance effects, turbulence, non-Newtonian rheology, wall motion, slip or pulsation can invalidate the parabolic formula. In the Rheology discussion of yield-stress pipe flow, a moving plug is a reminder that not all viscous profiles share Poiseuille geometry.
To audit a report, ask which fluid model was used, how pipe area and discharge were measured, whether density variation matters, how close the flow is to the assumed regime, and which correction coefficient was applied to which integral. An uncertainty statement should cover all of those dependencies relevant to the decision.
Does equal flow rate mean equal force on a pipe? No. It establishes the same volume crossing the section per time, not the total momentum and external-force balance. One needs the profile, pressure, wall effects and the rest of the control volume.
Is the parabolic velocity twice as fast everywhere? No. It is twice the mean only at the centreline, 1.5 times at half radius, and zero at a no-slip wall. The cross-sectional average remains U because ring areas have different weights.
Is 4/3 a universal fluid momentum coefficient? No. It belongs to the stipulated fully developed laminar circular-pipe profile. For a uniform profile β=1, and other geometries or flow regimes produce other integrals.
Practice A: with the same area, density and mean speed, double U. The uniform momentum flux grows by a factor of four, and the parabolic flux does likewise if its dimensionless profile is unchanged. Practice B: with U fixed and pipe area halved, both ideal section momentum fluxes halve, but Q also halves. Practice C: if density doubles while velocity and area remain unchanged, momentum flux doubles.
Each check changes exactly one condition. This habit is more useful than carrying a formula into a case whose assumptions have changed without noticing it.
The central result is precise: for steady incompressible axial flow, identical A and Q fix U, but they do not by themselves fix ∫u²dA. In the two illustrative profiles, β changes from 1 to 4/3. That difference is caused by velocity distribution, not an error in the flow meter.
Return to The Gold Standard of Fluid Mechanics to connect this calculation to conservation of mass, pressure, viscosity and control volumes. For non-Newtonian complications, follow Yield-Stress Pipe Flow and the Plug That Still Moves. Both show why profile shape matters whenever the model assumes more than a single mean.
Primary teaching references include OpenStax on viscosity and Poiseuille flow, MIT’s Engineering Mechanics II notes, Engineering LibreTexts on energy conservation, and a fluid-flow textbook summary of momentum flux. They explain established physics; our model dimensions, liquid properties and flow readings are invented for this worked example.
Continue through the Gold Standard subject index or How X Works. Before trusting a summary, ask which integral the instrument supplies—and whether that integral actually answers the engineering question.