Units and Dimensional Analysis in Mechanics

Physics connects mathematical symbols to measurable properties of the physical world. A numerical value by itself is therefore usually incomplete. The statement that a length is “5” becomes physically meaningful only after a unit is supplied, such as 5 m or 5 cm. At a deeper level, both measurements have the same physical dimension: length.

This distinction between unit and dimension is one of the most useful pieces of error checking in introductory mechanics. It also leads to powerful scaling arguments and, at a more advanced level, to the construction of dimensionless groups such as those used in similarity theory and the Buckingham Pi theorem.

The main ideas of this article are:

1. physical quantities have dimensions independent of the units chosen to measure them;
2. equations representing physical laws must be dimensionally homogeneous;
3. units may be converted by multiplying by ratios equal to one;
4. scaling laws often follow from powers of length, area, volume, time, or mass;
5. dimensional analysis can constrain the possible form of a physical relation, although it generally cannot determine dimensionless numerical constants;
6. when several dimensional variables appear, the physics can often be reorganized into a smaller number of dimensionless combinations.

1 Physical quantities, dimensions, and units

A physical quantity is a measurable property such as length, time, mass, speed, force, or energy. A dimension describes the physical kind of a quantity, while a unit is an agreed standard used to assign it a numerical value.

For example, a distance may be written as

(1)

or as

(2)

The numerical values and units differ, but the physical dimension is the same:

(3)

Here square brackets mean “the dimensions of.”

1.1 Base dimensions in mechanics

Most elementary mechanics can be expressed using three base dimensions:

(4)

(5)

(6)

Other branches of physics require additional base quantities, such as electric current and Thermodynamic temperature. The complete SI system contains seven base quantities, but the M-L-T set is sufficient for most of the mechanics developed in the early part of this learning path.

A general mechanical dimension can be written as

(7)

where the exponents may be positive, negative, zero, or fractional.

1.2 Dimensions of common mechanical quantities

Derived SI units are abbreviations for combinations of base units. For example,

(8)

and

(9)

2 Dimensions obey algebraic rules

Dimensions multiply, divide, and carry powers just like algebraic quantities. If

(10)

then

(11)

If

(12)

then

(13)

For velocity,

(14)

so

(15)

Differentiation by time contributes a factor of T⁻¹. Thus

(16)

Integration by time contributes a factor of T. For example, since impulse is

(17)

its dimensions are

(18)

which are the dimensions of momentum.

3 Dimensional homogeneity

An equation intended to represent a physical law must be dimensionally homogeneous: terms that are added or subtracted must have the same dimensions, and the two sides of an equality must have the same dimensions.

Consider

(19)

The four terms have dimensions

(20)

(21)

(22)

and

(23)

The equation passes the dimensional test.

Dimensional consistency is a necessary condition for correctness, but not a sufficient one. The incorrect equation

(24)

is dimensionally homogeneous even though its numerical coefficients are not those of constant-acceleration motion.

3.1 Functions require dimensionless arguments

The argument of a sine, cosine, exponential, or logarithm must be dimensionless. For example,

(25)

requires

(26)

so

(27)

Angles measured in radians are dimensionless ratios, although the symbol rad is often retained as a useful label.

Similarly, an exponential such as

(28)

requires t∕τ to be dimensionless, so τ must have the dimension of time.

4 Worked example 1: use dimensions to detect an error

A student proposes the equation

(29)

for constant-acceleration motion. Determine whether it can be correct.

The first two contributions have dimension length:

(30)

(31)

But the final term has dimensions

(32)

That is a velocity, not a length. Therefore the proposed equation cannot be correct.

Replacing at by at² restores dimensional homogeneity:

(33)

This example illustrates the fastest use of dimensional analysis: finding algebraic or transcription errors before inserting numbers.

5 Units and conversion factors

A unit conversion changes the numerical representation of a physical quantity without changing the quantity itself. The method is to multiply by one in a useful form.

Because

(34)

the ratio

(35)

is equal to one.

For example,

(36)

5.1 Powers of units matter

Area and volume conversions require powers of the conversion factor. Since

(37)

we have

(38)

and

(39)

A common error is to convert the numerical factor only once when the unit is squared or cubed.

6 Scaling and similarity

Dimensional reasoning becomes especially useful when a system is enlarged or reduced while keeping the same geometric shape. Let every linear dimension be multiplied by a scale factor s.

Then

(40)

area scales as

(41)

and volume scales as

(42)

For objects of the same material and density, mass also scales as

(43)

The surface-area-to-volume ratio therefore scales as

(44)

A larger similar object has less surface area per unit volume than a smaller one.

Scaling arguments are common in mechanics. For example, if a period has the form

(45)

then making the characteristic length four times larger makes the period twice as large.

7 Worked example 2: infer the pendulum-period dependence

Suppose the small-amplitude period T of a pendulum depends only on pendulum length L and gravitational acceleration g. Assume a power-law form

(46)

where C is dimensionless.

The dimensions are

(47)

(48)

and

(49)

Therefore

(50)

Matching the exponent of time gives

(51)

so

(52)

Matching the exponent of length gives

(53)

so

(54)

Hence dimensional analysis determines

(55)

The exact small-angle dynamics gives C = 2π, but dimensional analysis alone cannot determine that dimensionless constant. This limitation is fundamental.

8 Dimensional analysis as a way to infer physical form

Suppose a quantity Q depends on several variables and no detailed equation is known. If the relation is expected to be a product of powers, write

(56)

Replace every variable by its dimensions and equate the exponents of the independent base dimensions. The result is a linear algebra problem for the unknown exponents.

This procedure cannot generally determine:

• dimensionless numerical coefficients such as 2, 1∕2, or 2π;
• additive structure, such as whether two terms should be added;
• dependence on a dimensionless function;
• signs or vector directions.

Nevertheless, the method can strongly constrain the possible form of a law.

9 Worked example 3: infer the speed scale of a circular orbit

Suppose an orbital speed v depends only on the gravitational constant G, central mass M, and orbital radius r. Assume

(57)

The dimensions are

(58)

(59)

(60)

and

(61)

Thus

(62)

Matching powers gives

(63)

so

(64)

For mass,

(65)

so

(66)

For length,

(67)

so

(68)

Therefore

(69)

Newtonian circular-orbit dynamics gives C = 1. Again, dimensions determine the dependence on G, M, and r, while dynamics supplies the dimensionless coefficient.

10 Dimensionless quantities

A dimensionless quantity has dimension

(70)

Dimensionless quantities are important because they can be compared across systems with different scales and units.

Examples include

(71)

(72)

and

(73)

The numerical value of a properly defined dimensionless quantity is independent of the coherent unit system used to evaluate it.

11 Buckingham-Pi intuition

The Buckingham Pi theorem provides a systematic way to reorganize a dimensional problem. In introductory form, the idea is:

If a physical relation contains n dimensional variables built from k independent base dimensions, the relation can often be rewritten using n − k independent dimensionless combinations.

A full proof belongs in a more advanced treatment, but the counting idea is already useful in freshman mechanics.

Consider a drag force F that depends on fluid density ρ, speed v, characteristic size L, and dynamic viscosity μ. There are five variables:

(74)

They use the three base dimensions M, L, and T. Buckingham counting therefore suggests

(75)

independent dimensionless groups.

Two possible choices are

(76)

and

(77)

The second group is the Reynolds number. The original five-variable relation can then be written schematically as

(78)

where Φ is a dimensionless function that dimensional analysis alone cannot determine.

This viewpoint is the foundation of similarity testing, wind-tunnel modeling, fluid scaling, and many engineering correlations.

12 Common mistakes

1. Confusing a unit with a dimension. Meter and centimeter are different units of the same dimension, length.
2. Treating dimensional consistency as proof. A dimensionally correct equation may still contain the wrong coefficient or wrong physical dependence.
3. Adding unlike dimensions. A length cannot be added directly to a speed or energy.
4. Forgetting powers in unit conversions. Converting m² or m³ requires squaring or cubing the conversion factor.
5. Putting dimensional quantities inside sine, exponential, or logarithm functions. Their arguments must be dimensionless.
6. Dropping units too early. Carrying units through a numerical calculation is a powerful error check.
7. Expecting dimensional analysis to determine 2π. Pure numerical constants are invisible to dimensional reasoning.
8. Assuming every dimensionless quantity is unimportant. Dimensionless groups often control the qualitative behavior of a system.

13 Problem-solving workflow

For a mechanics calculation, the following routine is effective:

1. identify the physical quantity being sought;
2. write the dimensions expected for the answer;
3. carry units through the calculation;
4. check that every additive term has the same dimensions;
5. check that both sides of every physical equation have the same dimensions;
6. convert to a consistent unit system before numerical substitution when necessary;
7. after obtaining the result, verify units and inspect scaling or limiting behavior.

14 Practice problems

The following problems use stable mechanics identifiers. Short answers are supplied after the practice set.

MECH-FOUND-001 — Dimensions of basic mechanical quantities

Starting from [x] = L, [m] = M, and [t] = T, derive the dimensions of velocity, acceleration, force, kinetic energy, and power.

MECH-FOUND-002 — Check a kinematic equation

Determine whether

(79)

is dimensionally homogeneous.

MECH-FOUND-003 — Unit conversion

Convert 90 km∕h to m/s.

MECH-FOUND-004 — Area and volume scaling

Two geometrically similar solid objects are made of the same material. Every length of the second object is three times the corresponding length of the first. By what factors do surface area, volume, mass, and surface-area-to-volume ratio change?

MECH-FOUND-005 — Mass-spring period

Assume the period T of an ideal mass-spring oscillator depends only on mass m and spring constant k. Given

(80)

use dimensional analysis to find the form of T up to a dimensionless constant.

MECH-FOUND-006 — Wave-speed scaling

The speed v of transverse waves on a stretched string depends on Tension F and linear mass density μ, where [μ] = ML⁻¹. Determine the dependence of v on F and μ up to a dimensionless constant.

MECH-FOUND-007 — Pressure dimensions

Pressure is force per area. Derive its dimensions in M-L-T form and verify that the SI unit Pa is equivalent to kg/(m s²).

MECH-FOUND-008 — Dimensionless exponential

A damped signal contains the factor

(81)

If m is a mass, determine the dimensions required for b.

MECH-FOUND-009 — Pendulum scaling

Two small-angle pendulums are in the same gravitational field. The second has nine times the length of the first. What is the ratio of their periods?

MECH-FOUND-010 — Verify Buckingham groups

Using

(82)

(83)

(84)

and

(85)

verify that F∕(ρv²L²) and ρvL∕μ are dimensionless.

15 Short answer key

Practice problems:

1. [v] = LT⁻¹; [a] = LT⁻²; [F] = MLT⁻²; [K] = ML²T⁻²; [P] = ML²T⁻³.
2. Yes. Both sides have dimensions L²T⁻².
3. 25 m∕s.
4. Area ×9; volume ×27; mass ×27; A∕V becomes one third as large.
5. T = C.
6. v = C.
7. [p] = ML⁻¹T⁻²; Pa = kg/(m s²).
8. [b] = MT⁻¹.
9. T₂∕T₁ = 3.
10. Both groups reduce to M⁰L⁰T⁰.

16 Connections to later mechanics

Units and dimensional analysis are foundational topics. They reappear throughout mechanics:

• vector components must carry compatible units before they are added;
• derivatives introduce inverse powers of time and integrals introduce powers of time or length;
• force laws and potential energies can be checked dimensionally;
• oscillation periods and wave speeds often admit useful scaling arguments;
• non-dimensionalization reveals control parameters in drag, oscillation, fluid, orbital, and nonlinear problems;
• advanced perturbation and similarity methods frequently begin by identifying appropriate dimensionless variables.

GRE-specific test strategy and speed questions are intentionally separated from this core article. They are provided in the attachable companion entry M00-03G: GRE Physics Companion – Units and Dimensional Analysis.

The natural next foundation article is Scalars and Vectors in Mechanics.

17 Source and license note

This PhysicsLibrary entry is an original synthesis written for the M00-03 mechanics curriculum target. Its scope and terminology were benchmarked against the archived 2016 University Physics Volume 1 clone distributed through BCcampus under CC BY 4.0, UCD Physics 9A Classical Mechanics by Tom Weideman under CC BY-SA 4.0, and Jeffrey W. Schnick’s Calculus-Based Physics I under CC BY-SA 3.0. No source figures are reproduced; all figures in this package are original.

References