Intuition The one core idea
A physical law cannot know what units you chose — the swing takes the same time whether you measure its rope in metres or miles. So every honest equation can be rewritten using only unit-free numbers , and the whole topic is just a way of counting how many such numbers a problem allows.
This page builds — from absolute zero — every symbol, word, and picture the parent note Buckingham π theorem leans on. Read it top to bottom; nothing below uses a thing not yet built above it.
Definition Dimension vs. unit
A dimension is the kind of quantity you are measuring — a length , a duration , a mass . A unit is one particular yardstick for that kind — a metre, a second, a kilogram. "Length" is a dimension; "metre" is a unit of that dimension.
Picture a ruler. The fact that it measures distance is its dimension. The little tick marks (cm, inch) are the unit. Change the ticks and the number changes, but the kind of thing you measured does not.
Why the topic needs this: the theorem's whole power is that laws survive when you swap units. To talk about "swapping units without changing the kind of thing," we must separate dimension (kind) from unit (yardstick). See Fundamental and derived units .
Definition Fundamental dimensions
M , L , T
Mechanics needs only three fundamental dimensions — building blocks that cannot be made from each other:
M = mass (how much stuff),
L = length (how far),
T = time (how long).
The square-bracket notation [ X ] is read "the dimensions of X ." It is a question-answering symbol : "what kinds of base quantity, and in what powers, make up X ?"
Intuition Why brackets, and why powers?
Speed is distance-per-time. Written in kinds: a length divided by a time. We record this as [ v ] = L T − 1 . The exponents (+ 1 on L , − 1 on T ) say "one length on top, one time on the bottom." Every mechanical quantity is some product of powers of M , L , T — that is all a dimension ever is.
Worked example Reading dimensions off a formula
Area = length × length, so [ A ] = L 2 .
Density = mass ÷ volume, so [ ρ ] = M L − 3 (the symbol ρ , Greek "rho", means density).
Force = mass × acceleration, and acceleration is speed-per-time = L T − 2 , so [ F ] = M L T − 2 .
Viscosity (the symbol μ , Greek "mu", = fluid "stickiness") works out to [ μ ] = M L − 1 T − 1 .
Why the topic needs this: every variable in the theorem is first turned into its [ X ] form. Without the bracket-and-power language, there is nothing to count.
Definition Dimensionless quantity
A dimensionless quantity is one whose dimensions are M 0 L 0 T 0 = 1 — a pure number with no units at all. It reads the same to everyone in the universe regardless of their rulers or clocks.
ratio kills the units
Divide a length by a length and the L 's cancel: L + 1 ⋅ L − 1 = L 0 = 1 . That is why ratios are unit-free. If a bug is 2 mm long on a 10 mm leaf, the ratio 0.2 is 0.2 whether you measured in mm, inches, or lightyears — the yardstick divides itself out.
Why the topic needs this: the theorem's output is a set of dimensionless groups (π 1 , π 2 , … ). "Dimensionless" is precisely the property that makes them unit-independent, hence physically meaningful. Compare Reynolds number , the most famous such number.
Definition Dimensional homogeneity
An equation is dimensionally homogeneous when every term added or equated has exactly the same dimensions . You may only add apples to apples: a length can never equal a time.
Intuition Picture a balance scale
Think of the = sign as a scale. On the left pan sits the dimensions of the left side; on the right, the right side's dimensions. A true law keeps the scale level — both pans show the same combination of M , L , T . If they differ, the "law" is nonsense no matter what numbers you plug in.
F = 6 π μ r v
Right side: [ μ ] [ r ] [ v ] = ( M L − 1 T − 1 ) ( L ) ( L T − 1 ) = M L − 1 + 1 + 1 T − 1 − 1 = M L T − 2 .
Left side: [ F ] = M L T − 2 . ✓ The pans balance. (The 6 π is a pure number, dimensionless, so it never affects the balance.)
Why the topic needs this: homogeneity is the single demand from which the whole theorem flows. See Dimensional homogeneity .
Definition The three counts
n = the number of physical variables in your problem (force, speed, density, …).
k = the number of independent fundamental dimensions actually present (often, but not always, 3 for M , L , T ).
p = the number of independent dimensionless groups you can build, given by p = n − k .
Intuition "Knobs" of the physics
Imagine each dimensionless group as a knob on a machine. n is how many raw ingredients you have; k is how many independent "unit rulers" tie them down; p is what's left free to turn. Fewer knobs = simpler the hidden recipe. That is the entire promise of the theorem.
Why the topic needs this: p = n − k is the formula of the theorem. Everything else is machinery for computing k honestly and building the p groups.
Definition Dimension matrix
D
Write each variable's exponents of M , L , T as a column. Stacking all n columns gives a grid of numbers called the dimension matrix D : k rows (one per dimension), n columns (one per variable).
Intuition Rank = number of truly independent rows
The rank of D is how many of its dimension-rows carry genuinely new information — rows that are not just copies or combinations of others. If speed, length and time are your only variables, "time" is hidden inside "speed ÷ length," so the rows collapse and the rank drops. The true k is this rank, not a blind "3."
Why the topic needs this: this is why p = n − k is true, and it warns you that k must be measured (as a rank), not assumed. Non-dimensionalising the full Navier–Stokes equations is where these ideas turn rigorous.
Definition Repeating variables and
π groups
Repeating variables are the k chosen "core" variables you reuse to build every group. They must (a) together contain all k dimensions and (b) be dimensionally independent (they can form no dimensionless group among themselves).
Each ==π group== π i (the symbol π here means "a dimensionless bundle," not 3.14159 ) is one repeater-combination that comes out unit-free.
Mnemonic "n minus k = π keys"
n variables, k dimensions, and the leftover p = n − k are the dimensionless π keys that unlock the law. Repeaters = the k "locks" you build everything from.
Why the topic needs this: the recipe in the parent note ("pick repeaters, combine with each leftover") produces the π i . These groups feed straight into Drag force and drag coefficient and Model testing and similarity .
Dimensionless = pure number
State the dimensions of speed v [ v ] = L T − 1 — a length over a time.
Difference between a dimension and a unit Dimension = the kind of quantity (length); unit = the yardstick for it (metre).
What does the bracket [ X ] ask? "Which powers of M , L , T make up X ?"
When is a quantity dimensionless? When its dimensions reduce to M 0 L 0 T 0 = 1 — a pure number, equal for all observers.
Why does a ratio of two lengths carry no units? L + 1 ⋅ L − 1 = L 0 = 1 ; the yardstick cancels itself.
State the rule of dimensional homogeneity Every term equated or added must have identical dimensions.
Check [ F ] against [ μ r v ] Both equal M L T − 2 , so F = 6 π μ r v is homogeneous.
Meaning of n , k , p n = number of variables; k = independent dimensions present; p = n − k = number of independent π groups.
Why is k a rank, not just "3"? If variables share hidden dependence, dimension-rows collapse; the honest count is rank ( D ) .
Two conditions on repeating variables (1) Together span all k dimensions; (2) be dimensionally independent (form no π among themselves).
Dimensional homogeneity — the balance-scale rule this page builds toward
Fundamental and derived units — where M , L , T and their yardsticks come from
Reynolds number — the flagship dimensionless number once you can build π groups
Drag force and drag coefficient — first payoff of the counting
Model testing and similarity — matching π groups across scales
Navier–Stokes equations — where rank/homogeneity turn rigorous