Centroids and centers of mass

The formula

For a set of point masses, the center of mass is the mass-weighted average position:

x̄ = ( Σ mᵢ·xᵢ ) / ( Σ mᵢ )      (same form for ȳ, z̄)

Reading: add up each mass times its coordinate, divide by the total mass. That's it — a weighted average, no more mysterious than a grade-point average where heavier parts count more.

For a solid body with density ρ, the sum becomes an integral over the volume:

x̄ = ( ∫ x·ρ dV ) / ( ∫ ρ dV )

Reading: same average, taken over infinitely many tiny chunks of material.

The numerator has a name of its own — the first moment about the axis:

Q_yz = ∫ x·ρ dV        (first moment of mass about the y–z plane)

Reading: the first moment is "how much mass, times how far out it sits." Set it to zero and you've found the axis through the centroid — the point about which the first moments cancel.

If the density is uniform, ρ divides out and you're left with a pure geometry problem — the centroid of the area or volume:

x̄ = ( ∫ x dA ) / ( ∫ dA ) = ( ∫ x dA ) / A

Reading: for a constant-density part, the center of mass and the geometric centroid are the same point. That's why engineers use the words almost interchangeably — and why they get in trouble when the density isn't constant.

The workhorse in practice is the composite version — break a messy shape into simple pieces you already know:

x̄ = ( Σ Aᵢ·x̄ᵢ ) / ( Σ Aᵢ )       (areas)
x̄ = ( Σ mᵢ·x̄ᵢ ) / ( Σ mᵢ )       (masses)

Reading: find the centroid of each piece from a table, weight each by its area or mass, average. Cutouts (holes) count as negative area.

Where you meet it

Lift planning on the floor. Before a crane picks a fixture, an engine section, or a satellite panel, somebody has to know where the CG is so the sling legs bracket it. Hook the load off-center and it swings, rotates, and slaps the rigging until it hangs with its CG straight under the hook. On a bench that's a nuisance; on a 4,000-lb hoist it's a torn tag line and a stop-work.
Test stand and shaker fixtures. A vibration fixture that carries the unit-under-test off its own centroid feeds a rocking moment into the shaker armature instead of clean axial motion. The control accelerometer reads one thing, the corners of the part see another, and your notching gets ugly. Fixture designers push the combined CG down onto the armature centerline on purpose.
Vehicle and airframe stability review. CG location relative to the center of pressure (rockets) or the aerodynamic center (aircraft) is a go/no-go item at a design review. Too far aft and the thing is statically unstable; too far forward and it won't rotate. The CG budget gets tracked mass-item by mass-item through the whole build.
Section properties for a beam. Before you can compute bending stress you need the neutral axis, and the neutral axis of a symmetric elastic section passes through the area centroid. Get the centroid of a built-up I-section or a welded tee wrong and every stress number downstream is wrong with it.

How it works

The whole idea rests on one physical fact: for statics, you can replace a distributed weight with a single force acting at the CG and nothing about the external equilibrium changes. That's what makes the point useful — it collapses a cloud of material into one arrow you can put on a free-body diagram.

The moves that actually save time:

Symmetry is free. The centroid always lies on any axis of symmetry. A symmetric part hands you one, sometimes two, coordinates for nothing. Don't integrate what you can read off by inspection.
The centroid need not be inside the material. A ring, a boomerang, a C-channel — all have centroids sitting in empty space. Nothing wrong with that. It's a location, not a piece of metal.
Composite plus subtraction handles holes. Treat a bolt hole or a lightening pocket as a piece of negative area (or negative mass) at its own centroid. The algebra takes care of the rest. This is the single most common real-world method, and it's why centroid tables for standard shapes exist.

The mistakes that bite people:

Confusing centroid with center of mass. They coincide only when density is uniform. A part that's steel on one end and aluminum on the other has its CG shifted toward the steel, nowhere near the geometric centroid. Cast parts, potted electronics boxes, and fueled tanks all break the uniform-density assumption. Use the mass form, not the area form.
Center of mass vs. center of gravity. They're identical in a uniform gravity field, which is essentially everything you'll ever build on Earth. They diverge only when gravity varies noticeably across the object — think a long spacecraft in low orbit, where the gravity-gradient torque that arises from that tiny divergence is real enough that satellites use it for passive stabilization. For a bracket on a bench, the distinction is academic; for orbital mechanics it isn't.
Averaging centroids instead of weighting them. (x̄₁ + x̄₂)/2 is only right when the two pieces have equal area or mass. Weight by area or mass every time, or a small piece far out will pull your answer to a place it doesn't belong.

A quick sanity check you can trust: the centroid of a triangle sits one-third of the way up from each side — at (b/3, h/3) for a right triangle with legs b and h on the axes. A semicircle of radius R has its centroid 4R/(3π) ≈ 0.424·R off the flat edge, not at R/2. If your composite answer lands outside the bounding box of the part, you made a sign error on a cutout.

History

The subject is almost as old as rigorous geometry itself. Archimedes of Syracuse (287–212 BC) laid the foundation in his treatise On the Equilibrium of Planes, where he started from the law of the lever — equal weights balance at equal distances — and used it to locate the centers of gravity of flat figures ^[1]^[2]. In the first book he nailed down the parallelogram, the triangle, and the trapezoid; the second book he devoted entirely to one hard problem, the center of gravity of a parabolic segment ^[2]. He proved that a triangle balances at the intersection of its medians, the one-third point every engineer still uses.

What nobody knew until much later was how he found these results before proving them. A palimpsest surfaced in 1906 containing his lost work The Method, in which Archimedes admits he first discovered his theorems by a mechanical trick — imagining the figure sliced into strips and hung on a balance beam — and only afterward supplied the airtight geometric proof ^[2]. It's the earliest description we have of an engineer's instinct: get the answer by physical reasoning, then make it rigorous.

The idea sat mostly still for eighteen centuries. Then Pappus of Alexandria, writing in his Collection in the 4th century AD, stated the two theorems that tie centroids to solids of revolution — the surface and volume you sweep out are the path length of the centroid times the length or area you're rotating — provided the axis doesn't cut through the figure ^[3]^[4]. Pappus's statement was largely forgotten until the 1600s, when Johannes Kepler (1615) and, more fully, the Swiss mathematician Paul Guldin worked out the same relationships; Guldin's Centrobaryca of roughly 1640 ran to hundreds of pages on centers of gravity alone ^[3]^[4]. That's why the result carries the double name Pappus–Guldin to this day.

You can check Pappus yourself in one line: spin a small circle of radius r whose center rides a distance R from the axis (with r < R, so the circle clears the axis) and you get a torus of volume 2π·R · (π·r²) = 2π²·R·r². The centroid did the bookkeeping.

Related tools

/tools/moment-of-inertia-shapes
/tools/section-modulus-rect
/tools/beam-deflection
/tools/bolt-torque
/tools/convert-mass-force
/tools/convert-area-volume
/tools/shaft-speed-critical
/tools/vibration-natural-freq

Sources

Written by HE in our own words from the cited sources — engineering judgment included, your stamp still required. All entries →