Home / Courses / Fluid Mechanics / Dimensional Analysis

Fluid Mechanics · Chapter 7 of 10 · Advanced

Dimensional Analysis and Modeling

Why does a tiny model in a wind tunnel predict a full-size aircraft? Because the physics depends only on dimensionless groups. Match those groups and the model and the real thing behave alike.

01

Readiness check

This chapter reasons with dimensions and ratios. Tick only what you can do closed-notes.

  • Write the primary dimensions M, L, T of a quantity.
  • Check dimensional homogeneity of an equation.
  • Recall the Reynolds number.
  • Solve a small set of linear equations for exponents.
  • Work with ratios and scale factors.
0 or 1 weak itemsContinue with this chapter.
2 weak itemsReview dimensions and the Reynolds number in Chapter 1.
3 or more weak itemsRevisit units and dimensional homogeneity first.
02

The core idea

Any physical relation can be rewritten among dimensionless groups; the Buckingham Pi theorem says there are n − m of them, and matching them makes a model dynamically similar to the real thing.

number of Π groups = n − mCD = f(Re)similarity: Remodel = Reprototype

Because nature does not care about our units, every valid equation can be expressed purely in dimensionless groups. The Buckingham Pi theorem counts them: n variables built from m primary dimensions give n − m independent groups. This collapses a forest of variables into a few, so that scattered drag data, for instance, fall onto a single curve CD = f(Re). The same principle lets a scale model predict a prototype: match the governing groups (Reynolds, Froude, Mach) and the flows are similar.

The skill works when: you count the groups correctly and match the ones that govern the flow between model and prototype.
The skill breaks down when: a relevant group is left unmatched, or dimensional homogeneity is violated in forming the groups.
The concept. Drag measured across many fluids, speeds, and sizes collapses onto a single curve when plotted as the dimensionless CD versus Re. Dimensionless groups are the natural variables of fluid mechanics.
03

The skills, taught in order

Dimensional analysis is bookkeeping with dimensions, made powerful by similitude. Five skills cover homogeneity, the Pi theorem, the named groups, similarity, and model testing.

7.1 Dimensions and homogeneity

Every quantity reduces to primary dimensions, mass M, length L, time T (and temperature Θ when needed). A valid equation is dimensionally homogeneous: every term carries the same dimensions. This both checks work and is the engine of the Pi theorem.

7.2 The Buckingham Pi theorem

If a problem has n variables built from m primary dimensions, it can be written with k = n − m independent dimensionless groups (the Π groups). Choose m repeating variables that together span the dimensions, then combine each remaining variable with them to form a Π. The physics becomes a relation among the Πs.

7.3 Common dimensionless groups

A handful of named groups recur throughout fluid mechanics.

GroupDefinitionRatio of
Reynolds, ReρVL/μinertia to viscous
Froude, FrV/√(gL)inertia to gravity
Mach, MaV/cspeed to sound speed
Drag coefficient, CDF/(½ρV²A)drag to dynamic force

7.4 Similitude

A model is similar to a prototype on three levels: geometric (same shape, scaled), kinematic (same flow pattern), and dynamic (same force ratios). Dynamic similarity, the goal, is achieved by matching the dimensionless groups that govern the flow.

7.5 Model testing and scaling

To predict a prototype from a model, match the governing group (Reynolds for most internal and submerged flows, Froude for free-surface and ship flows). Matching it fixes the model test speed; the dimensionless result (such as CD) then scales the force back to full size.

Engineering connection: wind-tunnel and tow-tank testing, pump and turbine scaling, and the very form of every empirical fluids correlation rest on dimensional analysis.

04

Worked example 1: drag by the Pi theorem

The drag force F on a sphere depends on the density ρ, speed V, diameter D, and viscosity μ. Use the Buckingham Pi theorem to find how many dimensionless groups there are and what they are.

Figure 1. Five variables built from three primary dimensions reduce to two dimensionless groups: a drag coefficient and the Reynolds number. The whole drag law is CD = f(Re).
  1. ProblemFind the number and form of the dimensionless groups for the drag in Figure 1.
  2. Given / findVariables F, ρ, V, D, μ (n = 5); primary dimensions M, L, T (m = 3). Find the Π groups.
  3. AssumptionsThese five variables fully describe the drag; incompressible (no Mach), no free surface (no Froude).
  4. ModelApply the Pi theorem: k = n − m groups, with ρ, V, D as repeating variables.
  5. Equationsk = n − m = 5 − 3 = 2 Π₁ = F/(ρV²D²), Π₂ = ρVD/μ
  6. SolveThere are 2 groups. Checking dimensions, both Π₁ = F/(ρV²D²) and Π₂ = ρVD/μ come out dimensionless (M⁰L⁰T⁰). Π₁ is a drag coefficient (within a constant of CD) and Π₂ is the Reynolds number, so CD = f(Re).
  7. CheckThe result matches the universal drag curve: drag data for spheres across all sizes and fluids collapse onto one CD-versus-Re line, exactly what two groups predict.
  8. ConclusionFive variables became one curve. Dimensional analysis tells you the form of the relationship even before any experiment, and shrinks the testing to a single dimensionless plot.
Result. Two groups: CD = F/(½ρV²A) and Re = ρVD/μ, so CD = f(Re).
05

Worked example 2: a wind-tunnel model

A car travels at 30 m/s. A 1:5 scale model is tested in a wind tunnel using the same air. Find the tunnel speed for dynamic similarity, and how the measured model drag relates to the full-size drag.

Figure 2. Matching the Reynolds number in the same fluid forces the model to run five times faster. Because the dimensionless drag is then identical, the model drag equals the prototype drag.
  1. ProblemFind the model test speed and the drag relation for the setup in Figure 2.
  2. Given / findVp = 30 m/s, scale Lm/Lp = 1/5, same air (ρ, μ equal). Find Vm and Fm/Fp.
  3. AssumptionsReynolds number governs (incompressible, no free surface); geometric similarity holds.
  4. ModelMatch Re to set Vm; equal CD then scales the force by ρV²L².
  5. EquationsρVmLm/μ = ρVpLp/μ → Vm = Vp(Lp/Lm) Fm/Fp = (Vm/Vp)²(Lm/Lp)²
  6. SolveVm = 30 × 5 = 150 m/s. Fm/Fp = (5)²(1/5)² = 25 × (1/25) = 1, so Fm = Fp.
  7. CheckMatching Re in the same fluid always gives equal model and prototype force, an exact consequence of the scaling. The catch: 150 m/s is near the speed of sound, where compressibility (Mach) would intrude, which is why such tests use water tunnels or pressurised air instead.
  8. ConclusionSimilitude fixes the test conditions and the scaling law together. Choosing the right fluid keeps the model speed practical while preserving the governing group.
Result. Test speed 150 m/s; with matched Re, the model drag equals the prototype drag.
06

Misconceptions and diagnostics

MistakeSymptomDiagnostic questionCorrection
Wrong group countToo many or too few Πs"Is it n − m groups?"Count variables and primary dimensions carefully.
Unmatched governing groupModel does not predict prototype"Which group governs this flow?"Match Re (submerged) or Fr (free surface), as appropriate.
Ignoring compressibilityHigh model speed gives wrong drag"Is the model Mach number too high?"Change fluid or scale so Ma stays low while matching Re.
Dimensional group not dimensionlessΠ still carries units"Does the group reduce to M⁰L⁰T⁰?"Re-derive the exponents so all dimensions cancel.
07

Practice ladder

Level 1 · Direct skill

A problem has 6 variables built from 3 primary dimensions. How many dimensionless groups describe it?

Show answer

k = n − m = 6 − 3 = 3 independent Π groups.

Level 2 · Mixed concept

A ship model is tested at 1:25 scale. Which dimensionless group must be matched, and what model speed results if the ship moves at 10 m/s?

Show answer

Free-surface (wave) flows are governed by the Froude number, Fr = V/√(gL). Matching it: Vm = Vp√(Lm/Lp) = 10 × √(1/25) = 10/5 = 2 m/s. Froude scaling lowers the model speed, unlike Reynolds scaling.

Level 3 · Independent problem

Confirm that the drag coefficient CD = F/(½ρV²A) is dimensionless.

Show answer

[F] = MLT⁻², [ρV²A] = (ML⁻³)(L²T⁻²)(L²) = MLT⁻². The ratio is M⁰L⁰T⁰, dimensionless. The ½ is a pure number, so CD carries no units.

Level 4 · Transfer to real engineering

Find a real scaled test (a wind-tunnel aircraft, a ship tow-tank, a spillway model). Identify the governing dimensionless group and how the result scales to full size.

What good work looks like

The governing group named (Re or Fr), the model condition derived from matching it, and the dimensionless result used to scale a force or rate to the prototype.

08

Working with AI, and proving it yourself

Use AI as an examiner, not a solver

"Check that I have n − m groups and each is dimensionless."
"Give me five flows; I will name the dimensionless group that governs each."
"Form the Pi groups." Choosing repeating variables and reducing dimensions is the skill.
"What model speed?" Matching the governing group yourself is the point.

Portfolio task

Apply dimensional analysis to one problem: form the Π groups, identify the governing one, and set up a model test with the correct scaling.

Must include: a group count (n − m), dimensionless groups verified, and a model-to-prototype scaling.
09

Retrieval and spaced review

Closed notes. Answer out loud, then reveal.

1. How many dimensionless groups does a problem have?

k = n − m, variables minus primary dimensions (Buckingham Pi).

2. What does the Reynolds number compare? The Froude number?

Re: inertia to viscous forces. Fr: inertia to gravity (free-surface flows).

3. Name the three levels of similarity.

Geometric, kinematic, and dynamic; dynamic is the goal.

4. How do you achieve dynamic similarity?

Match the dimensionless groups that govern the flow between model and prototype.

5. Which group governs ship (free-surface) testing?

The Froude number.

TodayFinish this quiz and Levels 1 and 2 of the ladder.
+1 dayRe-derive the drag groups and model scaling from a blank page.
+3 daysOne Pi-theorem and one similitude problem.
+7 daysSee Re govern pipe friction in Chapter 8.
+30 daysRecall CD = f(Re) when computing drag in Chapter 9.
10

Textbook mapping

ItemMapping
Primary sourceÇengel and Cimbala, Fluid Mechanics: Fundamentals and Applications, Chapter 7 (Dimensional Analysis and Modeling)
Cross-referenceWhite, Ch. 5 · Munson, Ch. 7
Core topics7.1 Homogeneity · 7.2 Buckingham Pi · 7.3 Dimensionless groups · 7.4 Similitude · 7.5 Model testing
Engineering connectionWind tunnels, tow tanks, and the form of every empirical correlation.
Read nextChapter 8: Internal (Pipe) Flow.