Moment of Inertia: Why Beam Shape Beats Weight

The moment of inertia (also called the second moment of area, symbol I) is a geometric property that measures how a cross-section's area is distributed relative to an axis. The farther the material is from the axis, the more it contributes to I — and the contribution grows with the square of the distance.

For a rectangle of width b and height h:

I = bh³ / 12

The cube in h³ is the key insight: doubling the depth increases the moment of inertia eight-fold (2³ = 8), while doubling the width only doubles it. This is why structural steel sections are deep and narrow — not short and wide.

The moment of inertia appears in two of the most important equations in structural engineering:

• Beam deflection: δ = 5wL⁴ / (384EI). Deflection is inversely proportional to I — double I, halve the deflection.
• Column buckling: P_cr = π²EI / (KL)². The critical buckling load is directly proportional to I.

In steel design tables, the moment of inertia is listed as I_x (about the strong axis, the major axis of bending) and I_y (about the weak axis). For an I-beam, I_x is typically 5–12× larger than I_y because the flanges are far from the x-axis but close to the y-axis.

For standard rolled sections (W, S, HP, C, L, HSS), you never need to calculate I from scratch — the values are tabulated in the AISC Steel Construction Manual, the European section tables (IPE, HEB, HEA), or the Indian Standard section tables (ISMB, ISMC). Software like CalcSteel stores these values in its profile database.

But understanding the calculation builds intuition. For an I-beam, I_x is computed using the parallel axis theorem:

I_x = I_web + 2 × (I_flange + A_flange × d²)

where d is the distance from the flange centroid to the section's neutral axis. The A_flange × d² term (the "transfer term") dominates — for a typical W section, 85–95% of I_x comes from the flanges, and most of that from the transfer term, not the flanges' own inertia.

For a W410×60: each flange is 178 mm × 12.8 mm at approximately ±197 mm from the centroid. The transfer term alone gives 2 × (178 × 12.8) × 197² = 2 × 2 278 × 38 809 = 177 × 10⁶ mm⁴ — that is 82% of the total I_x = 216 × 10⁶ mm⁴. The remaining 18% comes from the web and the flanges' own bh³/12.

This arithmetic explains why deeper is stiffer: increasing the depth moves the flanges farther from the neutral axis, and the I grows with d². It also explains why removing flange area (as in a coped beam) devastates the moment of inertia far more than removing web area.

In a beam, the moment of inertia controls two things: deflection and stress distribution.

Deflection. For a simply supported beam with uniform load: δ_max = 5wL⁴ / (384EI). The E (elastic modulus) is 200 000 MPa for all structural steel — you cannot change it. The only section property you can change is I. A beam with twice the I deflects half as much, everything else being equal.

For a W410×60 (I_x = 216 × 10⁶ mm⁴) spanning 8 m with 20 kN/m service load: δ = 5 × 20 × 8 000⁴ / (384 × 200 000 × 216 × 10⁶) = 12.2 mm. The L/360 limit is 22.2 mm — comfortable.

But swap to a W310×33 (I_x = 65 × 10⁶ mm⁴, much lighter but shallower): δ = 5 × 20 × 8 000⁴ / (384 × 200 000 × 65 × 10⁶) = 40.6 mm. That is L/197 — far beyond the L/360 limit. The lighter beam cannot work, not because it lacks strength, but because it deflects too much.

Stress distribution. The extreme-fibre bending stress is σ = M × c / I, where c is the distance to the extreme fibre. The ratio I/c is the elastic section modulus S_x. A higher I (for the same depth) means a higher S_x and lower stress. This is why heavier sections of the same depth family (W410×60 vs W410×46) have higher moment capacity — they have thicker flanges, which increases I more than it increases c.

Every cross-section has two principal moments of inertia: I_x (about the horizontal x-axis, the strong or major axis) and I_y (about the vertical y-axis, the weak or minor axis).

For an I-beam, the flanges are far from the x-axis but close to the y-axis. This geometric asymmetry creates a huge ratio between I_x and I_y:

• W410×60: I_x = 216 × 10⁶ mm⁴, I_y = 20.5 × 10⁶ mm⁴. Ratio: 10.5×.
• W200×27: I_x = 25.8 × 10⁶ mm⁴, I_y = 3.32 × 10⁶ mm⁴. Ratio: 7.8×.

This ratio has profound design implications:

• Beam bending: I-beams are meant to bend about the strong axis. A beam loaded about the weak axis wastes 80–90% of its material efficiency.
• Column buckling: A column buckles about its weakest axis (smallest I) unless braced. The weak-axis moment of inertia I_y and the corresponding radius of gyration r_y = √(I_y/A) determine the effective length factor K and the slenderness ratio KL/r_y.
• Lateral torsional buckling: The LTB capacity depends on I_y, J (torsional constant), and C_w (warping constant). Sections with low I_y relative to I_x are more susceptible to LTB.

Hollow sections (HSS, RHS, CHS) have I_x ≈ I_y, which makes them ideal for columns (equal buckling resistance in both directions) and for members under bi-axial bending or torsion.

The answer depends on whether you want absolute I_x (the stiffest section available) or I_x per unit weight (the most efficient section).

Highest absolute I_x in the AISC manual: the W44×335 (W1100×335 in metric) with I_x ≈ 8 710 × 10⁶ mm⁴. This is a monster — 1 118 mm deep, 335 kg/m. Used for long-span transfer girders, crane runways, and bridge girders.

Highest I_x per weight: deep, lightweight W sections like the W610×82 (I_x/weight = 762/82 = 9.3 × 10⁶ mm⁴ per kg/m) outperform shallow, heavy sections like the W310×67 (145/67 = 2.2). The rule of thumb: go deeper before going heavier. A deeper section puts more material far from the neutral axis, maximising the transfer term in the parallel-axis theorem.

In European sections, the IPE family (narrow flange) gives excellent I_x/weight for beams, while the HEB family (wide flange) gives better I_y/weight for columns. CalcSteel lets you sort profiles by I_x, weight, or the ratio I_x/weight to find the optimal section for your application.

I-beams and hollow sections represent two different strategies for distributing material:

I-beams concentrate material in the flanges (far from the x-axis) to maximise I_x. The web is thin — it carries shear but contributes little to I_x. Result: very high I_x per kg, but I_y is much smaller.

Hollow sections (RHS, SHS, CHS) distribute material evenly around all four sides. Result: I_x ≈ I_y, plus high torsional stiffness (J is 100–1000× higher than I-beams). But for the same weight, I_x is lower than an I-beam's because material is "wasted" near the neutral axis (the side walls).

Numerical comparison at ~60 kg/m:

• W410×60: I_x = 216 × 10⁶ mm⁴, I_y = 20.5 × 10⁶ mm⁴. Ratio I_x/I_y = 10.5.
• HSS 254×254×9.5 (≈59 kg/m): I_x = I_y ≈ 89 × 10⁶ mm⁴. Ratio = 1.0.

The I-beam wins on I_x by 2.4×. The hollow section wins on I_y by 4.3×. This trade-off drives design decisions:

• Use I-beams for beams bending about one axis (floor beams, rafters).
• Use hollow sections for columns (equal buckling resistance), torsion members (high J), and exposed architectural elements (clean look, no open flanges).

CalcSteel stores the full set of section properties for every profile in its database — AISC (W, S, HP, HSS, C, L), European (IPE, HEB, HEA, UPN), Indian (ISMB, ISMC, ISLB), and Brazilian (NBR 6355 cold-formed). Here is how to access them:

In the profile browser: Click the profile name in the Properties panel to open the section database. You can filter by family (W, IPE, HSS, etc.) and sort by any property — including I_x, I_y, weight, or Z_x. The table shows all properties at a glance: A, d, b_f, t_f, t_w, I_x, I_y, S_x, S_y, Z_x, Z_y, r_x, r_y, J, and C_w.

In the design verification panel: After running the analysis, open the design check for any member. The section properties used in the verification are displayed alongside the calculations, so you can trace exactly how I_x feeds into the deflection check or how r_y feeds into the buckling slenderness.

Sorting by I_x/weight: For preliminary design, sort the profile table by I_x in descending order and look for the lightest section that meets your deflection requirement. This is faster than trial-and-error and ensures you are not over-designing.

CalcSteel also supports custom profiles: if you have a built-up section (plate girder, double angle, etc.), you can enter the dimensions and CalcSteel computes I_x, I_y, and all derived properties automatically.

Let us compute I_x for a W410×60 from its dimensional properties and verify it against the tabulated value.

Dimensions: d = 407 mm (total depth), b_f = 178 mm (flange width), t_f = 12.8 mm (flange thickness), t_w = 7.7 mm (web thickness), h = d − 2t_f = 407 − 25.6 = 381.4 mm (web clear height).

Step 1 — Flange inertia about its own centroid. Each flange is a rectangle b_f × t_f: I_flange,own = b_f × t_f³ / 12 = 178 × 12.8³ / 12 = 31 127 mm⁴ ≈ 0.031 × 10⁶ mm⁴. Tiny — this is negligible.

Step 2 — Flange transfer term. Distance from the section centroid to the flange centroid: d_f = (d − t_f)/2 = (407 − 12.8)/2 = 197.1 mm. A_flange = b_f × t_f = 178 × 12.8 = 2 278 mm². Transfer term: A × d² = 2 278 × 197.1² = 88.5 × 10⁶ mm⁴. For two flanges: 177 × 10⁶ mm⁴.

Step 3 — Web inertia. I_web = t_w × h³ / 12 = 7.7 × 381.4³ / 12 = 35.6 × 10⁶ mm⁴.

Step 4 — Total. I_x = 2 × (I_flange,own + A_flange × d_f²) + I_web = 2 × (0.031 + 88.5) + 35.6 = 177.1 + 35.6 = 212.7 × 10⁶ mm⁴.

The tabulated value is 216 × 10⁶ mm⁴. The small difference (~1.5%) comes from the fillets (root radii) at the web-flange junction, which add area far from the centroid. The manual value accounts for the fillets; our simplified rectangle model does not.

The takeaway: 82% of I_x comes from the flange transfer terms. The web contributes only 17%, and the flanges' own bh³/12 is negligible. This is why adding even 1–2 mm of flange thickness increases I_x dramatically — and why cutting or coping a flange is so detrimental to stiffness.