Dimensional analysis and its applications

This is the capstone topic of the CBSE Class 11 unit Units and Measurements, sitting directly after dimensions and dimensional formulae in the same chapter and closing out the chapter before the course moves into kinematics. It draws on nothing beyond the dimensional formulae you have already tabulated for velocity, acceleration, force, energy and so on, and on the SI base units introduced earlier in the same unit. The prescribed skill set is exactly three applications — checking an equation, converting units, deriving the form of a relation — plus a clear-eyed list of the method's limitations; nothing about error propagation or significant figures (covered separately in this unit) is needed here.

Every measurable mechanical quantity can be written as a product of powers of mass, length and time: $[\text{quantity}] = [\text{M}^a\text{L}^b\text{T}^c]$. Speed is $[\text{LT}^{-1}]$, acceleration is $[\text{LT}^{-2}]$, force is $[\text{MLT}^{-2}]$, and so on — these are the dimensional formulae from the previous topic. The principle of homogeneity simply insists that in a true physical equation, every term that is added, subtracted, or set equal to another must carry the same triple $(a,b,c)$.

where: each "term" is one additive piece of the equation (for instance, in $v^2=u^2+2as$, the terms are $v^2$, $u^2$, and $2as$), and $[\text{M}^a\text{L}^b\text{T}^c]$ is the shared dimensional formula every one of them must reduce to. The physical reading is blunt: a length can only ever be compared with another length, never with a time or a mass, no matter how the equation is dressed up algebraically. This is why "$5\ \text{metres} + 3\ \text{seconds}$" is not merely bad style — it is meaningless, and no amount of correct arithmetic can rescue it.

Application 2 — converting units. The numerical value $n$ and the unit $u$ of any physical quantity are tied together by $n_1u_1=n_2u_2$ (the quantity itself does not change when you switch from CGS to SI, only the number describing it does). Writing the unit itself in terms of the fundamental units raised to the dimensional powers $a$, $b$, $c$, this relation rearranges into a ready-made conversion formula:

where: $M_1,L_1,T_1$ are the mass, length and time units of the old system and $M_2,L_2,T_2$ those of the new system (for example grams versus kilograms, centimetres versus metres), $a,b,c$ are the powers from the quantity's dimensional formula $[\text{M}^a\text{L}^b\text{T}^c]$, and $n_1,n_2$ are the numerical values in the old and new systems. Physically, this says a conversion factor is never arbitrary — it is built entirely out of the ratios of the base units, each raised to exactly the power that quantity's dimensions demand. Get the power wrong and the conversion is wrong, even if you remember the right individual unit ratios.

Application 3 — deriving the form of a relation. If you can guess which physical quantities a result plausibly depends on, dimensional analysis lets you find how — the powers each one enters with — even before doing any experiment or full derivation:

where: $Z$ is the quantity being predicted, $A$, $B$, $C$ are the quantities it is assumed to depend on, $p,q,r$ are unknown exponents to be found by matching dimensions, and $k$ is a dimensionless constant that dimensional analysis can never determine — it must come from experiment or a full derivation using calculus or geometry. Substituting the dimensional formulae of $A$, $B$, $C$ into the right-hand side and comparing the resulting powers of M, L, T against those of $Z$ on the left gives (at most) three simultaneous equations, since M, L, T are the only three independent "slots" available in a mechanics problem.

Board level - Checking $v^2 = u^2 + 2as$ by dimensional homogeneity

Step 1. List the dimensional formula of each additive term in the equation of uniformly accelerated motion $v^2=u^2+2as$, using $[v]=[u]=[\text{LT}^{-1}]$, $[a]=[\text{LT}^{-2}]$, and $[s]=[\text{L}]$: $[v^2]=[u^2]=[\text{LT}^{-1}]^2=[\text{L}^2\text{T}^{-2}]$.

Step 2. Find the dimension of the remaining term, remembering that the pure number $2$ is dimensionless and simply drops out of the check: $[2as]=[a][s]=[\text{LT}^{-2}][\text{L}]=[\text{L}^2\text{T}^{-2}]$.

Step 3. Compare all three terms side by side: $v^2$, $u^2$, and $2as$ each reduce to $[\text{L}^2\text{T}^{-2}]$, so the principle of homogeneity is satisfied and the equation is dimensionally correct.

Step 4. State the limit of what this proves. Passing the check confirms the equation is consistent, not that the coefficient "2" in front of $as$ is numerically right — that guarantee only comes from the actual kinematic derivation (integrating $a\,dv = a\cdot a\,dt$ style working), which dimensional analysis cannot supply.

Competitive level - Converting Young's modulus from CGS to SI and predicting orbital speed's dependence

Step 1. Young's modulus $Y$ has dimensional formula $[\text{ML}^{-1}\text{T}^{-2}]$ (force per area, same as pressure), so $a=1,\ b=-1,\ c=-2$. Suppose $Y = 2\times10^{12}$ dyne/cm$^2$ in the CGS system and you need the SI value. Using $n_2=n_1\left[\tfrac{M_1}{M_2}\right]^{1}\left[\tfrac{L_1}{L_2}\right]^{-1}\left[\tfrac{T_1}{T_2}\right]^{-2}$ with $M_1=1\ \text{g}=10^{-3}\ \text{kg}=M_2\times10^{-3}$, $L_1=1\ \text{cm}=10^{-2}\ \text{m}=L_2\times10^{-2}$, and $T_1=T_2=1\ \text{s}$: $n_2 = 2\times10^{12}\times(10^{-3})^{1}\times(10^{-2})^{-1}\times(1)^{-2} = 2\times10^{12}\times10^{-3}\times10^{2} = 2\times10^{11}$.

Step 2. So $Y = 2\times10^{11}\ \text{N/m}^2$ (pascal) — a figure that matches the textbook's known order of magnitude for steel-like Young's moduli, giving confidence the powers $a,b,c$ were applied correctly.

Step 3. Now use Application 3 on a second quantity to cross-check the method's reach: predict how orbital speed $v_o$ of a satellite depends on the gravitational constant $G$, the central mass $M$, and orbital radius $r$, assuming $v_o = k\,G^{p}M^{q}r^{s}$. With $[G]=[\text{M}^{-1}\text{L}^3\text{T}^{-2}]$, $[M]=[\text{M}]$, $[r]=[\text{L}]$, and $[v_o]=[\text{LT}^{-1}]$: matching M gives $0=-p+q$; matching T gives $-1=-2p$, so $p=\tfrac12$, hence $q=\tfrac12$; matching L gives $1=3p+s=\tfrac32+s$, so $s=-\tfrac12$.

Step 4. Substituting back: $v_o = k\sqrt{\dfrac{GM}{r}}$, which correctly reproduces the shape of the known orbital-velocity formula $v_o=\sqrt{GM/r}$ (with $k=1$ here, though dimensional analysis alone could not have told us that $k$ is exactly $1$ rather than, say, $2$ — that a full circular-orbit force balance confirms separately, outside this chapter's scope). Timing note: in a competitive exam, budget under 90 seconds for a straightforward unit conversion like Step 1 (three power multiplications, no thinking required beyond correctly reading off $a,b,c$) and under two minutes for a three-unknown power match like Steps 3–4 — writing the M, L, T equations directly from the dimensional formula table, without re-deriving any dimensional formula from scratch, is what keeps both well inside time.

• Treating a dimensionally correct equation as proven — homogeneity is a necessary but never a sufficient condition; a formula with the right dimensions but the wrong numerical constant (say $as$ instead of $2as$) sails through the check undetected.
• Trying to dimensionally check equations containing $\sin\theta$, $e^{-t/\tau}$, or $\log x$ — the arguments of these functions must themselves be dimensionless, and dimensional analysis offers no machinery to derive or verify that requirement; it simply does not apply to such terms.
• Forgetting that the method can only ever fix the powers of the assumed variables, never the leading dimensionless constant — writing $T=2\pi\sqrt{L/g}$ as if the $2\pi$ came from the dimensional argument itself is a very common scoring slip.
• Assuming a dependency list that is either incomplete or padded — if a real variable is left out of the assumed power-law form, the derived relation will still look perfectly dimensionally consistent while being physically wrong; conversely, more than three unknown exponents in a mechanics (M, L, T) problem cannot be solved uniquely, since only three independent equations are ever available.
• Mixing up which subscript is "old" and which is "new" in the unit-conversion formula — swapping $n_1u_1=n_2u_2$ around, or getting a sign wrong on an exponent, silently inverts the answer by a large power of ten.

• State the principle of homogeneity of dimensions and explain why it is necessary but not sufficient for an equation to be physically correct.
• Check whether a given kinematic or dynamic equation is dimensionally consistent by reducing every additive term to its $[\text{M}^a\text{L}^b\text{T}^c]$ formula.
• Convert a physical quantity's numerical value between two unit systems using the formula $n_2=n_1[M_1/M_2]^a[L_1/L_2]^b[T_1/T_2]^c$.
• Derive the power-law form of an unknown relation from an assumed set of dependent variables, by matching powers of M, L, T on both sides.
• List all four key limitations of dimensional analysis: undetermined dimensionless constants, trigonometric/exponential/logarithmic relations, more than three unknown exponents in a mechanics problem, and indistinguishability of same-dimension quantities.

Ask the tutor: Want to try deriving the formula for the speed of a wave on a stretched string, or the escape speed of a satellite, from scratch using dimensional analysis — ask the AI tutor to set one up as a guided practice problem.