Which column section series gives the best weight efficiency for a given axial load?

For pure axial compression, W14 sections are the most weight-efficient column series in the AISC manual — they have depth-to-width ratios near 1.0, producing nearly equal rx and ry values and maximising buckling resistance for a given mass. For European practice, HEA sections are the column-optimised series. UC (universal column) sections serve the same role in British and Australian practice.

What KL/r range should I target for efficient column design?

Target KL/r between 40 and 100 for efficient steel column design. Below KL/r = 40, the column is stocky and the capacity approaches the squash load — adding more steel is the only way to increase capacity. Above KL/r = 120, the column is slender and the capacity is dominated by elastic buckling. Between 40 and 100, the column operates in the inelastic buckling range where the AISC column curve provides reasonable material efficiency.

How do I select between W12, W14, and W10 column sections?

W14 sections offer the best column efficiency because their depth and flange width are nearly equal (rx/ry = 1.5-2.0), providing good buckling resistance about both axes. W12 sections are the second choice — slightly deeper than wide, rx/ry = 1.8-2.5. W10 sections are typically used for lighter columns. The selection rule: if the column is axially dominated, start with W14; if the column resists moment, W12 or W14 will both work.

Can I use a beam section (W-shape, UB) as a column?

Yes, beam sections can function as columns, but they are less efficient for pure compression because their depth-to-width ratios are higher (2.0-3.0 vs 1.0-1.5 for column sections). The weak-axis radius of gyration ry is proportionally smaller, so the weak-axis KL/r controls. Using a beam section as a column can make sense when the column also resists significant strong-axis moment or when the unbraced length differs between axes.

How do I handle columns with different unbraced lengths in the two axes?

When the strong-axis and weak-axis unbraced lengths differ, compute KL/r for each axis separately. The axis with the larger KL/r governs. In many cases, the weak axis has a shorter unbraced length but a smaller r, and the strong axis has a longer unbraced length but a larger r — the governing axis may surprise you. Always compute both axes — never assume which governs without checking.

Column Section Selection Guide — W vs UC vs HEA, KL/r Optimization

| Depth | 353 mm | 260 mm | 230 mm | | Flange width | 254 mm | 256 mm | 240 mm | | Flange thickness | 16.4 mm | 17.3 mm | 12.0 mm | | Web thickness | 10.2 mm | 10.3 mm | 7.5 mm | | Area Ag | 11,500 mm^2 | 11,300 mm^2 | 10,600 mm^2 | | rx | 153 mm | 112 mm | 101 mm | | ry | 97.3 mm | 65.4 mm | 59.9 mm | | rx/ry | 1.57 | 1.71 | 1.69 | | Iy (10^6 mm^4) | 108 | 48.0 | 38.6 |

At approximately 90 kg/m, the W14x61 provides 2.8 times the weak-axis moment of inertia of the HEA 240 and 2.3 times that of the UC section. This translates directly into higher column buckling capacity for a given unbraced length. The W14 section also has the largest rx and ry, producing the lowest KL/r values. For pure axial compression at this weight, the W14 is the most efficient choice.

KL/r — The Governing Parameter

The slenderness ratio KL/r is the single most important parameter in column design. It directly determines the column buckling stress Fcr through the column curve, and the available compressive strength is phi Pn = phi x Fcr x Ag.

KL/r ranges and column behaviour:

KL/r range	Behaviour	Design regime	Typical section type
0-40	Stocky column	Squash load controls, Fcr approx Fy	Heavy W14, UC, HEA for high-rise lower floors
40-80	Intermediate	Inelastic buckling, Fcr = 0.658^(Fy/Fe) x Fy	Typical building columns, 3-10 storeys
80-120	Slender	Transitional, approaching elastic buckling	Lightly loaded columns, bracing, mezzanine posts
120-200	Very slender	Elastic buckling, Fcr = 0.877 x Fe	Long unbraced columns, industrial pipe racks
>200	Not permitted	AISC 360 E2 prohibits KL/r > 200	Redesign required — add bracing or increase section

The elastic buckling stress Fe = pi^2 x E / (KL/r)^2. For A992 steel (Fy = 345 MPa), the transition between inelastic and elastic buckling occurs at KL/r = 4.71 x sqrt(E/Fy) = 113. Below this value, the AISC column curve uses the inelastic buckling equation; above it, the elastic equation with the 0.877 factor for residual stress and initial imperfections.

Column Weight Efficiency — KL/r vs Mass

For a given axial load and unbraced length, the most efficient section is the one with the lowest mass that satisfies phi Pn >= Pu. This is determined by KL/r:

For KL/r < 60: Section weight scales approximately linearly with Pu because the column operates near the squash load. Phi Pn approx 0.9 x Fy x Ag, so Ag_req approx Pu / (0.9 x 345). The lightest section with sufficient Ag is the optimum.

For KL/r = 60-100: Section weight scales non-linearly — both Ag and r matter. A section with a larger ry (hence lower KL/r_y) can have a smaller Ag and lighter weight. This is where column series selection matters most: a W14 shape may be lighter than a W12 of equivalent Ag because its larger ry produces a higher Fcr.

For KL/r > 100: The column is buckling-dominated. Fcr is approximately proportional to E / (KL/r)^2, so doubling the effective length reduces the capacity by a factor of 4. At these slenderness ratios, adding mass is ineffective — the most efficient strategy is to reduce KL by adding intermediate bracing.

Selection Workflow — Step by Step

Step 1: Determine the effective length KL for each axis. Compute KLx = Kx x Lx and KLy = Ky x Ly separately. For a typical braced-frame building column, Kx = Ky = 1.0 (pinned-pinned assumption) unless a frame stability analysis justifies lower values. For sway frames, K typically exceeds 1.0 and should be determined from the alignment chart (AISC 360 Commentary Chapter C) or a direct analysis method.

Step 2: Estimate the required cross-sectional area. Ag_est = Pu / (0.85 x Fy) where the 0.85 factor accounts for approximately 15% reduction from the squash load due to slenderness effects. This gives a starting Ag for filtering sections.

Step 3: Filter by Ag and ry. From the section table, select sections with Ag >= Ag_est. Sort by mass ascending. For each candidate, compute KL/r_y = Ky x Ly / ry. The weak axis typically governs because ry < rx — even if Ly < Lx, the ratio KL/r_y is often larger.

Step 4: Compute phi Pn for each candidate. Using the governing code column curve:

AISC 360 E3: Fe = pi^2 x E / (KL/r)^2. If KL/r <= 4.71 sqrt(E/Fy), Fcr = 0.658^(Fy/Fe) x Fy. Otherwise, Fcr = 0.877 x Fe. phi Pn = 0.90 x Fcr x Ag.
AS 4100 Section 6: alpha_b = member slenderness reduction factor from Table 6.3.3 based on the modified slenderness lambda_n = (KL/r) x sqrt(kf) x sqrt(Fy/250). phi Ns = 0.9 x kf x An x Fy x alpha_c where alpha_c is the column curve factor.
EN 1993-1-1 Clause 6.3.1: Nb,Rd = chi x A x Fy / gamma_M1 where chi is the buckling reduction factor from the appropriate buckling curve (a, b, c, or d) based on the section type and axis.

Step 5: Select the lightest section with phi Pn >= Pu. Beginning with the lightest candidate that satisfies Ag_est, compute phi Pn. If it fails, move to the next heavier section. The first section that passes is the mass-optimum.

Step 6: Verify local buckling and compactness. Check that the flange and web slenderness satisfy the compact or non-compact limits per AISC 360 Table B4.1b. A slender element reduces the available compressive strength. W14 sections with heavy flanges are almost always compact; lighter W14 and W12 sections may have slender webs at high h/tw ratios.

Column Preliminary Selection Table — Axial Load vs Height

Axial load Pu (kN)	Height 3.5 m	Height 5.0 m	Height 7.0 m	Height 10.0 m
500	W10x33 (49 kg/m)	W10x39 (58 kg/m)	W12x45 (67 kg/m)	W12x58 (86 kg/m)
1,000	W12x50 (75 kg/m)	W12x58 (86 kg/m)	W14x68 (101 kg/m)	W14x90 (134 kg/m)
2,000	W14x82 (122 kg/m)	W14x90 (134 kg/m)	W14x120 (179 kg/m)	W14x176 (262 kg/m)
3,000	W14x132 (196 kg/m)	W14x145 (216 kg/m)	W14x193 (287 kg/m)	W14x283 (421 kg/m)
5,000	W14x211 (314 kg/m)	W14x233 (347 kg/m)	W14x311 (463 kg/m)	W14x398 (592 kg/m)

Assumptions: A992 steel (Fy = 345 MPa), K = 1.0 both axes, weak-axis governs (KL/r_y). Sections shown are the lightest W14 or W12 shape satisfying phi Pn >= Pu per AISC 360-22 Chapter E. For preliminary use only — verify with the actual K factor, unbraced length, and governing code.

When to Use Each Column Section Series

Use W14 sections when:

The project is in North America and the column is predominantly axially loaded
The column height exceeds 5 m (W14 sections maintain good KL/r at longer lengths)
High weak-axis stiffness is needed (unbraced length differs between axes)

Use UC sections when:

The project is in the UK or Australia and procurement favours British/AS sections
The column is in a braced frame with equal unbraced lengths in both axes (square UC profile optimises both axes simultaneously)
The connection is a simple shear connection with a fin plate — UC sections accommodate standard bolt gauges well

Use HEA sections when:

The project specifies EN standards (Europe, Middle East, parts of Asia)
The column depth must be minimised for architectural or clearance reasons (HEA sections pack more area into less depth than W sections of equivalent mass)
Combined axial load and weak-axis bending — HEA sections have wider flanges than IPE sections and better weak-axis modulus

Use UC/HEA over W sections when:

The project is in a metric-design region even if W-shapes are available — mixing series complicates procurement, connection detailing, and erection
Fire engineering uses the section factor Am/V — UC and HEA sections have different section factors than W-shapes, and this may favour one series over another depending on the fire resistance period required

Worked Example: Office Building Column Selection

A 7-storey office building with 4.0 m storey heights. The ground-floor interior column carries a factored axial load Pu = 2,800 kN from the gravity load combination. The column is part of a braced frame with Kx = Ky = 1.0. Unbraced length L = 4.0 m about both axes. Steel grade: ASTM A992 (Fy = 345 MPa, E = 200,000 MPa).

Step 1: Estimate required area. Ag_est = Pu / (0.85 x Fy) = 2,800,000 / (0.85 x 345) = 9,550 mm^2

Step 2: Filter W14 sections with Ag >= 9,550 mm^2. W14 sections sorted by mass:

W14x68: Ag = 8,710 mm^2 — below threshold, skip
W14x74: Ag = 9,420 mm^2 — below threshold, skip
W14x82: Ag = 10,500 mm^2, ry = 63.2 mm, mass = 122 kg/m — candidate
W14x90: Ag = 11,400 mm^2, ry = 97.5 mm, mass = 134 kg/m — candidate
W14x99: Ag = 12,600 mm^2, ry = 97.3 mm, mass = 147 kg/m — candidate

Step 3: Compute KL/r. KL/r_y = 1.0 x 4,000 / 63.2 = 63.3 for W14x82. KL/r_y = 1.0 x 4,000 / 97.5 = 41.0 for W14x90.

Step 4: Compute phi Pn. For W14x82 (KL/r = 63.3): Fe = pi^2 x 200,000 / (63.3)^2 = 493 MPa Fy/Fe = 345 / 493 = 0.70, Fcr = 0.658^0.70 x 345 = 0.746 x 345 = 257 MPa phi Pn = 0.90 x 257 x 10,500 / 1,000 = 2,430 kN — FAILS (2,430 < 2,800)

For W14x90 (KL/r = 41.0): Fe = pi^2 x 200,000 / (41.0)^2 = 1,174 MPa Fy/Fe = 345 / 1,174 = 0.294, Fcr = 0.658^0.294 x 345 = 0.878 x 345 = 303 MPa phi Pn = 0.90 x 303 x 11,400 / 1,000 = 3,110 kN — OK, utilisation = 2,800/3,110 = 0.90

Selected: W14x90 (134 kg/m).

Note that the W14x90 is heavier than the W14x82 but its larger ry (97.5 mm vs 63.2 mm) produces a much lower KL/r, and the combination of higher Fcr and larger Ag gives adequate capacity. The W14x90 is not a heavier version of the same section — it's a fundamentally different section geometry (deeper web, wider flanges) optimised for a different axial load range.

FAQ

How do I select the K-factor for column design?

The effective length factor K depends on the frame classification (braced or sway) and the rotational restraint at the column ends. For braced frames with typical beam-to-column shear connections, K = 1.0 is the standard assumption. For moment frames (continuous columns), use the alignment chart (AISC 360 Commentary Figure C-A-7.1 for braced frames, C-A-7.2 for sway frames) or the direct analysis method which permits K = 1.0 with notional loads capturing the P-delta effects. Never assume K < 1.0 unless a rigorous frame stability analysis justifies it — the pinned-end assumption is conservative and defensible when end restraint is uncertain.

Which buckling curve applies to a given section?

AISC 360 uses a single column curve (Chapter E3) for all hot-rolled sections — the 0.877 factor on Fe accounts for residual stresses and initial out-of-straightness without distinguishing between section types. AS 4100 uses multiple column curves (Table 6.3.3) where alpha_b varies for hot-rolled UB/UC sections (alpha_b = 0.0), welded sections (alpha_b = 0.5), and cold-formed sections (alpha_b varies). EN 1993-1-1 uses five buckling curves (a0, a, b, c, d) per Table 6.2, with hot-rolled H-sections typically using curve a or b depending on the axis and h/b ratio. The governing code determines which curve applies — check before importing a column curve assumption across standards.

When should I use a built-up column instead of a rolled section?

Built-up (fabricated) columns — two or more sections laced, battened, or connected with cover plates — are specified when: (1) the required axial capacity exceeds the largest available rolled section (W14x730, phi Pn approximately 18,000 kN), (2) the column must pass through services or architectural features that a solid section would obstruct, or (3) the unbraced length is very long (KL/r > 100) and a box or cruciform section provides better buckling resistance than any single rolled section of equivalent mass. Built-up columns are governed by AISC 360 Section E6 (built-up members) with modified slenderness limits accounting for shear deformation between connectors.