Column Design Guide — Engineering Reference

AISC 360 Chapter E steel column design: Fcr formula, effective length K, KL/r slenderness check, combined axial-bending interaction, and section selection.

Overview

Steel column design per AISC 360 Chapter E determines axial compression capacity based on the slenderness ratio KL/r, where K is the effective length factor, L is the unbraced length, and r is the governing radius of gyration. The designer must identify the critical buckling axis (usually weak-axis for W-shapes), classify the section per Table B4.1a, and compute the critical stress Fcr.

For beam-columns subjected to combined axial compression and bending, the interaction equations of Chapter H (H1-1a and H1-1b) govern the design. Selecting an efficient column requires balancing axial demand against bending from frame action, eccentricity, or lateral loads.

Axial compression capacity

AISC 360 Chapter E uses two equations depending on the slenderness parameter KL/r relative to 4.71*sqrt(E/Fy):

• Inelastic buckling (KL/r <= 4.71*sqrt(E/Fy)): Fcr = [0.658^(Fy/Fe)] * Fy, where Fe = pi^2 * E / (KL/r)^2 is the Euler elastic buckling stress.
• Elastic buckling (KL/r > 4.71*sqrt(E/Fy)): Fcr = 0.877 * Fe. This reduction accounts for initial imperfections and residual stresses.

The available axial strength is phi*c * Pn = 0.90 _ Fcr * Ag.

Effective length K-factors

The K-factor depends on end restraint conditions. Common values from the AISC Commentary Table C-A-7.1:

• K = 1.0 for pinned-pinned (sidesway prevented)
• K = 0.65 for fixed-fixed (sidesway prevented, theoretical 0.5)
• K = 1.2 for pinned-fixed (sidesway permitted, practical)
• K = 2.1 for fixed-free cantilever (sidesway permitted, practical)

For frames with sidesway, the alignment chart (nomograph) or direct analysis method of Chapter C should be used. The direct analysis method applies notional loads and stiffness reductions, allowing K = 1.0 in many cases.

Combined axial and bending (H1 interaction)

For P_r/P_c >= 0.2: P_r/P_c + (8/9) x [M_rx/M_cx + M_ry/M_cy] <= 1.0 (Eq. H1-1a)

For P_r/P_c < 0.2: P_r/(2 x P_c) + [M_rx/M_cx + M_ry/M_cy] <= 1.0 (Eq. H1-1b)

Where P_r is the required axial strength, P_c is the available axial strength, M_rx/M_ry are the required flexural strengths (including second-order effects), and M_cx/M_cy are the available flexural strengths. Second-order effects (P-delta member curvature and P-Delta story sway) must be included in the required strengths via B1-B2 amplification or direct second-order analysis per AISC Chapter C.

The interaction equations produce a linear interaction surface in P-Mx-My space. Equation H1-1a (used when the axial ratio exceeds 0.2) weights the bending terms by 8/9, slightly reducing the penalty for bending. Equation H1-1b (used when axial is light) removes most of the axial penalty and lets the bending terms control.

Worked example — W14x82 beam-column

Given: W14x82, A992 (F_y = 50 ksi), braced frame, K_x = K_y = 1.0, L = 14 ft, P_u = 400 kip, M_ux = 180 kip-ft (strong axis), M_uy = 0. Properties: A = 24.0 in^2, r_x = 6.05 in., r_y = 2.48 in., Z_x = 139 in^3, L_p = 8.76 ft, L_r = 28.5 ft.

1. Axial capacity: KL/r_y = (1.0 x 14 x 12)/2.48 = 67.7. F_e = pi^2 x 29000/67.7^2 = 62.5 ksi. F_cr = 0.658^(50/62.5) x 50 = 0.658^0.80 x 50 = 0.706 x 50 = 35.3 ksi. P_n = 35.3 x 24.0 = 847 kip. phi x P_n = 0.90 x 847 = 762 kip.
2. Flexural capacity: L_b = 14 ft. Since L_p (8.76) < L_b (14.0) < L_r (28.5), inelastic LTB. C_b = 1.0 (conservative for beam-column). M_n = M_p - (M_p - 0.7 x F_y x S_x) x (L_b - L_p)/(L_r - L_p) = 6950 - (6950 - 4079) x (14 - 8.76)/(28.5 - 8.76) = 6950 - 2871 x 0.265 = 6189 kip-in. phi x M_n = 0.90 x 6189/12 = 464 kip-ft.
3. Interaction: P_r/P_c = 400/762 = 0.525 > 0.2, so use H1-1a: 0.525 + (8/9) x (180/464 + 0) = 0.525 + 0.889 x 0.388 = 0.525 + 0.345 = 0.870 <= 1.0. OK.

The interaction ratio of 0.870 indicates the W14x82 has 13% reserve capacity for this load combination.

Column selection strategy

Efficient column selection depends on the axial-to-bending demand ratio:

• High axial, low bending (P_r/P_c > 0.5): Select based on A_g (cross-sectional area). W14 columns are preferred because they have large A_g values in a compact footprint and favorable r_y values.
• Moderate axial, moderate bending (0.2 < P_r/P_c < 0.5): Use AISC Table 6-1 (W-shapes selected for combined loading) which plots the interaction equation graphically.
• Low axial, high bending (P_r/P_c < 0.2): Select based on Z_x. Deep sections (W18, W21, W24) are more efficient in bending but may have lower r_y, increasing KL/r.

The W14 family is the workhorse for columns because: (1) r_y is relatively large (2.48 in. for W14x82 vs. 1.57 in. for W21x83), reducing KL/r; (2) many weights are available from W14x22 to W14x730; (3) column splices are simplified when all columns share the same nominal depth.

Code comparison — column design

Common mistakes to avoid

• Using K = 1.0 for columns in unbraced frames — sway-permitted K values are significantly larger (1.2 to 2.5+), dramatically reducing capacity. The direct analysis method (AISC Chapter C) permits K = 1.0 only when notional loads and stiffness reductions are applied to the analysis model.
• Neglecting weak-axis buckling — W-shapes have r_y much less than r_x (often 40-60%), so weak-axis KL/r usually governs unless the column is braced about the weak axis at shorter intervals.
• Ignoring P-delta amplification — second-order effects increase column moments by 10-30% in typical braced frames and much more in moment frames. The B1 amplifier (member P-delta) and B2 amplifier (story P-Delta) must both be applied.
• Using axial-only tables for beam-columns — AISC Table 4-1 gives pure axial capacity. Any bending reduces the available axial capacity via the H1 interaction equations. Even small moments from connection eccentricity should be checked.
• Not checking both strong and weak axis bending — for corner columns and columns at re-entrant corners, bending occurs about both axes simultaneously. Both M_rx and M_ry terms appear in the interaction equation.

AISC Chapter E design procedure — step by step

The following numbered procedure covers the full AISC 360-22 Chapter E (Section E3) design check for a doubly symmetric section under concentric axial compression. Every column design in a steel building follows these six steps.

Step 1 — Determine the factored axial load P_u. Apply LRFD load combinations from ASCE 7-22 / IBC 2024 to obtain the maximum required compressive strength. For gravity-only columns, the controlling combination is typically 1.2D + 1.6L. For columns in lateral-force-resisting systems, include wind and seismic combinations.

Step 2 — Select a trial section. Start with a W-shape from AISC Table 4-1 (available axial strength) based on the required capacity and KL. For beam-columns with significant bending, also check Table 6-1 (shapes for combined loading). The W14 family is the default starting point for building columns because of its favorable weak-axis radius of gyration and wide range of weights (W14x22 through W14x730).

Step 3 — Compute the slenderness ratio KL/r. Determine the effective length factor K (see the K-factor section below) and the unbraced length L for each axis. Calculate KL/r for both the strong axis (KL_x/r_x) and the weak axis (KL_y/r_y). The larger value governs. AISC recommends KL/r less than or equal to 200 for compression members.

(KL/r)_governing = max[ (K_x * L_x) / r_x ,  (K_y * L_y) / r_y ]

Step 4 — Determine the critical stress F_cr per AISC E3. First compute the elastic buckling stress:

F_e = pi^2 * E / (KL/r)^2

Then determine the limit slenderness:

lambda_r = 4.71 * sqrt(E / F_y)

For A992 steel (F_y = 50 ksi, E = 29,000 ksi), lambda_r = 4.71 * sqrt(29000/50) = 113.4.

• If KL/r <= lambda_r (inelastic buckling):

F_cr = [0.658^(F_y / F_e)] * F_y

• If KL/r > lambda_r (elastic / Euler buckling):

F_cr = 0.877 * F_e

Step 5 — Calculate the design compressive strength.

phi * P_n = phi_c * F_cr * A_g    (phi_c = 0.90)

Where A_g is the gross cross-sectional area from the AISC Steel Construction Manual Part 1.

Step 6 — Check the demand-to-capacity ratio.

P_u <= phi * P_n
D/C = P_u / (phi * P_n)  <=  1.0

If D/C exceeds 1.0, select a larger section and repeat from Step 3. If D/C is well below 1.0, consider a lighter section. Target a D/C between 0.80 and 0.95 for efficient designs with reasonable reserve.

Euler buckling vs inelastic buckling

Column buckling behavior divides into two distinct regimes based on the slenderness parameter KL/r. Understanding which regime governs is essential because the failure mechanism and the F_cr equation are fundamentally different.

Inelastic buckling (KL/r <= 4.71 * sqrt(E/F_y)). Short and intermediate columns buckle inelastically. At these slenderness levels, the column yield stress is reached on at least part of the cross section before elastic buckling can occur. Residual stresses from rolling and welding play a major role — they cause premature yielding at the tips of flanges, reducing the effective stiffness and triggering buckling at loads below the squash load F_y * A_g.

The AISC E3 equation for this regime is:

F_cr = [0.658^(F_y / F_e)] * F_y

This is a parabolic-type curve that transitions smoothly from the squash load (F_cr = F_y when KL/r = 0) toward the Euler curve at the transition point. The exponent (F_y/F_e) controls how rapidly the capacity drops with increasing slenderness. For stocky columns, the exponent approaches zero and F_cr approaches F_y. For columns near the transition slenderness, the exponent is close to 1.0 and F_cr drops significantly.

Elastic (Euler) buckling (KL/r > 4.71 * sqrt(E/F_y)). Slender columns buckle elastically — the material remains entirely elastic at the point of bifurcation, and failure is governed by the classic Euler hyperbola. The AISC equation applies a 0.877 reduction factor to account for initial out-of-straightness (taken as L/1000) and residual stresses:

F_cr = 0.877 * F_e = 0.877 * pi^2 * E / (KL/r)^2

The 0.877 factor comes from the Structural Stability Research Council (SSRC) Curve 2P. Without this factor, the Euler equation would be unconservative for real columns with geometric imperfections.

Quick threshold values for common steels:

For A992 columns, any member with KL/r below 113 is in the inelastic regime, which covers the vast majority of practical building columns. Elastic buckling only governs for unusually long, slender columns or unbraced members.

Verification at the boundary. At the transition slenderness (KL/r = 4.71 * sqrt(E/F_y)), both equations give the same result. For A992:

At KL/r = 113.4:
F_e = pi^2 * 29000 / 113.4^2 = 22.3 ksi
F_cr (inelastic) = 0.658^(50/22.3) * 50 = 0.658^2.242 * 50 = 0.392 * 50 = 19.6 ksi
F_cr (elastic)    = 0.877 * 22.3 = 19.6 ksi   (matches)

Effective length (K factor) quick reference

The K factor translates the actual end-restraint conditions of a column into an equivalent pin-ended length. The AISC Commentary (Table C-A-7.1) provides theoretical and recommended design K values for six idealized end conditions:

Why recommended values exceed theoretical. True fixed supports do not exist in practice — even stiff foundations rotate slightly, and beam-to-column connections have some flexibility. The recommended K values account for this partial fixity.

Alignment chart method. For columns in frames, the K factor is determined from the alignment chart (nomograph) using the G parameters at each end:

G = sum(I_col / L_col) / sum(I_beam / L_beam)

Where the summation includes all columns meeting at the joint (numerator) and all beams framing into the joint (denominator). Enter the alignment chart with G_A (bottom) and G_B (top) to read K. For sidesway-inhibited (braced) frames, use the braced frame chart. For sidesway-uninhibited (moment) frames, use the sway frame chart.

Direct analysis method simplification. AISC Chapter C permits the use of K = 1.0 for all columns when the direct analysis method is used with: (a) notional loads equal to 0.002 times the gravity load applied laterally at each level, and (b) reduced flexural stiffness (0.80 * EI) for all members contributing to the stability of the structure. Most modern designs use this approach because it eliminates the alignment chart and avoids the difficulty of estimating true K values in complex frames.

Worked example — W14x61 column (axial only)

Given: W14x61, A992 steel (F_y = 50 ksi, E = 29,000 ksi), braced frame, K = 1.0, L = 14 ft, P_u = 400 kips. Pure axial compression — no applied bending.

Section properties (AISC Manual Table 1-1):

Step 1 — Slenderness ratio. Weak axis governs for W-shapes unless braced differently:

KL/r_y = (1.0 * 14 * 12) / 2.45 = 168 / 2.45 = 68.6

Check against the recommended limit: 68.6 < 200. OK.

Step 2 — Check the transition slenderness.

lambda_r = 4.71 * sqrt(29000 / 50) = 4.71 * 24.08 = 113.4

Since KL/r = 68.6 < 113.4, the column is in the inelastic buckling regime.

Step 3 — Elastic buckling stress.

F_e = pi^2 * 29000 / 68.6^2 = 286,754 / 4706 = 60.9 ksi

Step 4 — Critical stress F_cr.

F_cr = 0.658^(50 / 60.9) * 50
     = 0.658^0.821 * 50
     = 0.712 * 50
     = 35.6 ksi

Step 5 — Design compressive strength.

phi * P_n = 0.90 * 35.6 * 17.9 = 573.8 kips

Step 6 — Demand-to-capacity ratio.

D/C = P_u / (phi * P_n) = 400 / 573.8 = 0.697

The W14x61 carries the 400-kip demand with approximately 30% reserve capacity. This is an efficient but not overly stressed column for this load. A W14x53 (phi*P_n ~ 495 kips at this KL) would be marginal at D/C = 0.81, and a W14x48 would be overstressed.

Worked example — HSS6x6x3/8 column (axial only)

Given: HSS6x6x3/8, A500 Gr. B (F_y = 46 ksi, E = 29,000 ksi), K = 1.0, L = 14 ft, P_u = 250 kips. This is a lighter column typical of mezzanine or low-rise construction.

Section properties (AISC Manual Table 1-12):

Step 1 — Slenderness ratio. Square HSS has equal radii of gyration about both axes:

KL/r = (1.0 * 14 * 12) / 2.27 = 168 / 2.27 = 74.0

Check: 74.0 < 200. OK.

Step 2 — Transition slenderness for A500 Gr. B.

lambda_r = 4.71 * sqrt(29000 / 46) = 4.71 * 25.12 = 118.2

Since 74.0 < 118.2, inelastic buckling governs.

Step 3 — Elastic buckling stress.

F_e = pi^2 * 29000 / 74.0^2 = 286,754 / 5476 = 52.4 ksi

Step 4 — Critical stress F_cr.

F_cr = 0.658^(46 / 52.4) * 46
     = 0.658^0.878 * 46
     = 0.697 * 46
     = 32.1 ksi

Step 5 — Design compressive strength.

phi * P_n = 0.90 * 32.1 * 7.58 = 218.9 kips

Step 6 — Demand-to-capacity ratio.

D/C = P_u / (phi * P_n) = 250 / 218.9 = 1.14

The HSS6x6x3/8 is overstressed at P_u = 250 kips with KL = 14 ft. The demand exceeds capacity by 14%. The designer must select a larger section such as HSS6x6x1/2 (phiP_n ~ 280 kips) or HSS8x8x3/8 (phiP_n ~ 310 kips), or reduce the unbraced length by adding bracing.

Comparison with W14x61 above: The W14x61 (phiP_n = 573.8 kips) has 2.6 times the axial capacity of the HSS6x6x3/8 (phiP_n = 218.9 kips), but it also weighs 61 lb/ft vs. 25.84 lb/ft — roughly 2.4 times heavier. The capacity-to-weight ratio is similar, but the W-shape provides much more absolute capacity. HSS sections are preferred for architectural exposure (clean appearance, equal properties about both axes) and torsional resistance, while W-shapes are preferred for maximum axial efficiency.

Column selection table — phi*P_n (kips) for common W-shapes

The following table provides the design compressive strength phi*P_n (kips, LRFD) for five common column sections at various effective lengths KL. All sections are A992 steel (F_y = 50 ksi). Use this table for quick trial selection before running a detailed calculation.

How to use this table:

1. Determine your P_u and KL.
2. Find the KL column that matches your condition (interpolate for intermediate values).
3. Scan down to find the lightest section with phi*P_n >= P_u.
4. Verify with a full calculation (local buckling checks, actual KL/r for your geometry).

Key observations:

• The W14x82 provides 30-35% more capacity than the W14x61 at every KL length, but at only 34% more weight (82 vs. 61 plf). This is because the extra material goes primarily into the flanges, increasing A_g without proportionally increasing r_y.
• The W12x65 outperforms the W14x61 at all KL lengths despite similar weight, because r_y = 3.02 in. vs. 2.45 in. — a 23% advantage in the governing slenderness parameter. For pure axial capacity, W12 shapes often beat W14 shapes at the same weight per foot.
• The W8x31 is only viable for shorter columns (KL < 15 ft) with moderate loads. Its small r_y = 2.02 in. causes rapid capacity loss with increasing KL.

Combined axial and bending — AISC Chapter H interaction equations

When a column carries bending in addition to axial compression (a beam-column), the AISC Chapter H interaction equations limit the combined demand. This is the general case for columns in moment frames, where lateral loads and frame continuity induce bending moments.

When does this apply? Any structural member with P_u > 0 and non-zero bending moments. Common scenarios include:

• Columns in moment frames (bending from lateral forces)
• Columns with eccentric connections (moment = P * eccentricity)
• Corner columns with beams framing in from two directions
• Columns supporting continuous beams (moment transfer through connections)

Equation H1-1a (when P_r / P_c >= 0.2):

P_r/P_c + (8/9) * [M_rx/M_cx + M_ry/M_cy]  <=  1.0

Equation H1-1b (when P_r / P_c < 0.2):

P_r/(2*P_c) + [M_rx/M_cx + M_ry/M_cy]  <=  1.0

Where:

Second-order effects are mandatory. The required moments M_rx and M_ry must include the amplification from P-delta effects. AISC Chapter C provides two approaches:

1. B1-B2 amplifier method — Amplify first-order moments using B1 (member P-delta, Eq. C2-2) and B2 (story P-Delta, Eq. C2-3). This is the traditional hand-calculation approach.

2. Direct second-order analysis — Run a geometric nonlinear (P-Delta) analysis in structural analysis software. The analysis model directly captures the amplification. This is the preferred method for complex frames.

Biaxial bending. When both M_rx and M_ry are non-zero (corner columns, columns at re-entrant corners), both terms appear in the interaction equation. This significantly reduces available capacity compared to uniaxial bending. A column that is adequate for P + M_x alone may fail the interaction check when a small M_y is added.

Practical design tip. For initial sizing of beam-columns, first size for axial alone (phi*P_n >= P_u with D/C around 0.70-0.80). This leaves 20-30% of the interaction equation for the bending terms. Then verify the full interaction check. If bending demands are high (M > 0.5 * M_c), start with a section selected for bending (using Z_x) and then verify axial capacity.

Run this calculation

• Column Capacity Calculator
• Section Properties Database

Related references

• K-Factor Guide
• Column K-Factor
• How to Verify Calculations
• Column Buckling Reference
• Effective Length Factors
• Frame Analysis Methods
• Composite Column
• Steel Buckling

Disclaimer

This page is for educational and reference use only. It does not constitute professional engineering advice. All design values must be verified against the applicable standard and project specification before use. The site operator disclaims liability for any loss arising from the use of this information.

Design Resources

Calculator tools

• Column Capacity Calculator
• Steel Buckling Calculator
• Steel Column Base Design Calculator

Design guides

• Column Capacity Worked Example
• Column Buckling Guide
• Column Buckling Calculator Guide
• EN 1993-1-1 Column Buckling Worked Example