Verified Engineering ReferenceUpdated 2026-09-13

Steel Section Properties: Verified IPE and HEB Tables

Area, second moment of area, section modulus and radius of gyration for European rolled I-sections, derived from nominal geometry including root fillets and cross-checked against EN 10365.

Direct answer. The textbook three-rectangle formula I = [bh³ − (b−tw)hw³]/12 under-predicts the real properties of a rolled I-section, because it ignores the four root fillets at the web-to-flange junctions. Across the 27 sections below those fillets add 2.6–4.9% to area and 2.5–5.7% to Iy. Include them and the derivation matches the published table to within 0.29%.

Deflection check

Elastic mid-span deflection of a simply supported beam under uniform load. Pure mechanics, no code factors — this is a serviceability sanity check, not a design verification.

IPE sections (EN 10365)

Narrow-flange beams. Scroll the table sideways for the full property set.

Sectionh (mm)b (mm)twtfrMass (kg/m)A (cm²)Iy (cm⁴)Wy (cm³)iy (cm)Iz (cm⁴)Wz (cm³)iz (cm)Max dev.
IPE 100100554.15.778.110.3217134.24.07165.81.240.05%
IPE 120120644.46.3710.413.2131853.04.90288.61.450.29%
IPE 140140734.76.9712.916.4354177.35.744512.31.650.04%
IPE 160160825.07.4915.820.09869108.76.586816.71.840.18%
IPE 180180915.38.0918.823.951,317146.37.4210122.22.050.02%
IPE 2002001005.68.51222.428.481,943194.38.2614228.52.240.18%
IPE 2202201105.99.21226.233.372,772252.09.1120537.32.480.02%
IPE 2402401206.29.81530.739.123,892324.39.9728447.32.690.02%
IPE 2702701356.610.21536.145.955,790428.911.2342062.23.020.09%
IPE 3003001507.110.71542.253.818,356557.112.4660480.53.350.10%
IPE 3303301607.511.51849.162.6111,767713.113.7178898.53.550.09%
IPE 3603601708.012.71857.172.7316,266903.614.951,043122.83.790.03%
IPE 4004001808.613.52166.384.4623,1281156.416.551,318146.43.950.04%
IPE 4504501909.414.62177.698.8233,7431499.718.481,676176.44.120.03%
IPE 50050020010.216.02190.7115.5248,1991927.920.432,142214.24.310.02%
IPE 55055021011.117.224105.5134.4267,1172440.622.352,668254.14.450.02%
IPE 60060022012.019.024122.4155.9892,0833069.424.303,387307.94.660.04%

HEB sections (EN 10365)

Wide-flange columns, depth roughly equal to width.

Sectionh (mm)b (mm)twtfrMass (kg/m)A (cm²)Iy (cm⁴)Wy (cm³)iy (cm)Iz (cm⁴)Wz (cm³)iz (cm)Max dev.
HEB 1001001006.010.01220.426.0445089.94.1616733.52.530.19%
HEB 1201201206.511.01226.734.01864144.15.0431852.93.060.03%
HEB 1401401407.012.01233.742.961,509215.65.9355078.53.580.06%
HEB 1601601608.013.01542.654.252,492311.56.78889111.24.050.03%
HEB 1801801808.514.01551.265.253,831425.77.661,363151.44.570.04%
HEB 2002002009.015.01861.378.085,696569.68.542,003200.35.070.01%
HEB 2202202209.516.01871.591.048,091735.59.432,843258.55.590.05%
HEB 24024024010.017.02183.2105.9911,259938.310.313,923326.96.080.01%
HEB 26026026010.017.52493.0118.4414,9191147.611.225,135395.06.580.04%
HEB 30030030011.019.027117.0149.0825,1661677.712.998,563570.97.580.02%

Why the simple formula is wrong, and by how much

A rolled I-section is not three rectangles. Where the web meets each flange there is a root fillet of radius r, and there are four of them. Each fillet is the region left when a quarter disc is removed from a square of side r:

Af = r² (1 − π/4) ≈ 0.2146 r²
c   = (5/6 − π/4) / (1 − π/4) ≈ 0.2234 r   (centroid, measured from the corner)
If = r⁴ (1 − 5π/16) − Af c² ≈ 0.007545 r⁴  (about its own centroid)

Each fillet is then shifted onto the section's neutral axis by the parallel-axis theorem:

Iy = Iy,ideal + 4 [ If + Af (hw/2 − c)² ]
Iz = Iz,ideal + 4 [ If + Af (tw/2 + c)² ]

These constants are exact, integrated from the geometry — not fitted to make the numbers agree. That distinction matters: a fudge factor tuned to one catalogue tells you nothing about the next section you meet.

Verification result: all 27 sections agree with EN 10365. The largest deviation across area, Iy, Wy and mass for every section is 0.29%, comfortably inside the 1.0% flag threshold applied site-wide. Deviations at this level reflect rounding in the published tables.

Frequently asked

Why is the strong axis called y-y here and x-x elsewhere?

European practice (EN 1993, EN 10365) labels the major axis y-y and the minor axis z-z. US practice (AISC) labels them x-x and y-y. Same physical quantity, different name — so an American Ix is the European Iy. Getting this backwards is a common and expensive mistake when working across the two systems.

Do I use these values or the published table for design?

Use the published table for your governing standard. These derived values exist to verify that table and to show where the numbers come from, not to replace it. They agree to within 0.29%, so if your source disagrees with both, the source is worth checking.

Does the deflection check above include self-weight?

No. The load w is whatever you enter. Add the section's own mass (the Mass column, in kg/m × 9.81 ÷ 1000 = kN/m) if you want it included. The formula is elastic deflection only — it applies no partial factors, no buckling check and no serviceability limit.

Why is the radius of gyration useful?

i = √(I/A) expresses how far the area sits from the axis. It is the property that drives slenderness, λ = Lcr/i, and therefore buckling capacity — which is why a wide-flange HEB makes a far better column than an IPE of the same mass.