Canadian Column Design — CSA S16-19 Clause 13.3 Axial Compression

Comprehensive reference for steel column design per CSA S16-19. Covers the SSRC multiple column curve system (n-factors for W-shapes, HSS, and fabricated sections), effective length factors for braced and sway frames, section classification effects on Cr, and biaxial bending interaction.

Related pages: CSA Effective Length (K Factor) | CSA Combined Loading | CSA Compact Section Limits | Column Capacity Calculator


The CSA S16 Column Formula — SSRC Multiple Column Curves

CSA S16-19 uses the Structural Stability Research Council (SSRC) multiple column curve approach, which is fundamentally different from the single-curve approach of some other codes. The factored axial compressive resistance is:

Cr = phi x A x Fy x (1 + lambda^(2n))^(-1/n)

Where:

The parameter n controls the shape of the column curve in the transition region between squash (lambda = 0) and Euler buckling (lambda > 2). Higher n values produce a more favourable curve with higher capacity at intermediate slenderness because the section has lower residual stresses.

n-Factor Selection — CSA S16 Table 4

Section Type Manufacturing n Curve Residual Stress Level
W-shapes (hot-rolled) Hot-rolled + air 1.34 SSRC 2P Moderate (0.3 Fy)
HSS — round or square (hot-formed) Hot-formed + air 2.24 SSRC 1P Low (0.15 Fy)
HSS — cold-formed Cold-formed 2.24 SSRC 1P Low (0.15 Fy)
Double angles, back-to-back Hot-rolled 0.69 SSRC 3P High (0.5 Fy)
Welded built-up I-section (deep fillet welds) Fabricated 1.34 SSRC 2P Moderate-High
Pipe (API 5L) Seam-welded 2.24 SSRC 1P Low
Welded box sections Fabricated 0.69 SSRC 3P High
Heavy welded H-sections (flame-cut) Fabricated 0.69 SSRC 3P Very High (0.5 Fy)

The profound effect of n on column capacity is most visible at intermediate slenderness (KL/r = 60-100). At KL/r = 60 in 350W:

The 30% capacity difference between an HSS and a fabricated box section at the same KL/r is entirely due to the residual stress pattern — the geometric properties (area, radius of gyration) are identical.

Cr Values — W-Shapes (350W, n = 1.34)

KL/r lambda Cr/A (MPa) Cr for A = 15,000 mm^2
20 0.216 334 4,507 kN
40 0.431 293 3,956 kN
60 0.647 236 3,186 kN
80 0.862 179 2,417 kN
100 1.078 134 1,809 kN
120 1.294 99 1,337 kN
150 1.617 68 918 kN
200 2.156 39 527 kN

KL/r = 200 represents the maximum slenderness for main compression members (Clause 13.3.2). At this slenderness, the column retains only approximately 11% of its squash load capacity.

Effective Length Factor K — Clause 13.3.3

The effective length factor K accounts for the rotational and translational restraint at the column ends. For typical Canadian building columns:

Braced frame (non-sway): 0.50 <= K <= 1.00 **Sway frame (unbraced):** K >= 1.00, determined from alignment charts (Jackson & Moreland nomograph)

For a braced frame, K is determined by the stiffness ratio G = sum(Ic/Lc) / sum(Ig/Lg) at each end of the column:

End Conditions Theoretical K Design K Typical Application
Both ends fixed 0.50 0.65 Column with deep beams both sides
Fixed-pinned 0.70 0.80 Interior column, pinned base
Both ends pinned 1.00 1.00 Lean-on column in braced frame
Fixed-free 2.00 2.10 Cantilever column (flagpole)

The "design K" values include an allowance for the fact that connections are never perfectly rigid. A nominally "fixed" connection provides approximately 90% rotational restraint, which translates to a K of 0.65 rather than the ideal 0.50.

Section Classification Effect on Column Capacity

For Class 3 sections (slender elements in compression), CSA S16 Clause 13.3.1(b) requires using an effective area Aeff rather than the gross area A:

For Class 3 webs in axial compression: The web effective width: b_eff = (750 / sqrt(Fy)) / (h/w) x h (mm)

For a W310x39 in 350W under axial compression: h/w = 44.0, Fy = 350 MPa. 750/sqrt(350) = 40.1. b_eff = (40.1/44.0) x h = 0.911 x h (about 9% of web is ineffective).

The effective area reduction for Class 3 sections ranges from 5-15% for lighter W-shapes. For columns, the web is typically the governing element for Class 3, even when the flange remains Class 1 or 2.

Biaxial Bending Interaction — Clause 13.8

For columns subject to both axial compression and bending about both axes, CSA S16 Clause 13.8.2 provides:

Cf/Cr + 0.85 x Mfx/Mrx + 0.60 x Mfy/Mry <= 1.0 (for Class 1 and 2 sections)

The 0.85 and 0.60 coefficients reflect the lower probability of simultaneous peak loading for secondary bending moments. For Class 3 sections under combined loads, the coefficients are 0.85 for both axes.

More rigorous interaction (Cross-section strength, Clause 13.8.3): Cf/Cr + (Mfx^2 + Mfy^2)^0.5 / Mr <= 1.0

This is appropriate when bending moments are comparable in magnitude about both axes and derive from the same load source (e.g., a corner column loaded by beams in two directions).

Worked Example — W310x107 Interior Column

Problem: An interior column W310x107 (CSA G40.21 350W) at the ground floor of a 6-storey office building. Unbraced length L = 4.2 m. Kx = 1.0 (pinned-pinned in major axis), Ky = 1.0 (pinned-pinned in minor axis — braced by slab in that direction but conservatively assumed). Factored axial load Cf = 2,800 kN. Minor-axis bending from unequal beam spans: Mfy = 60 kNm.

Section properties — W310x107: A = 13,600 mm^2 | rx = 136 mm | ry = 77.5 mm | Zy = 450 x 10^3 mm^3 (minor axis plastic modulus)

Step 1 — Section classification: Flange: b/2tf = (191/2)/13.0 = 7.35 <= 7.75 (Class 1 for 350W) Web (compression): h/w = (311-2x13)/10.5 = 285/10.5 = 27.1 <= 40.1 (Class 1 for axial) Section is Class 1 — full area effective.

Step 2 — Slenderness: KL/rx = 1.0 x 4200 / 136 = 30.9 (major axis — apparently more favourable) KL/ry = 1.0 x 4200 / 77.5 = 54.2 (minor axis governs — always check both)

Weak axis governs for virtually all W-shape columns unless the column is braced differently in each direction.

Step 3 — Cr (axial): lambda = 54.2 x sqrt(350 / (pi^2 x 200000)) = 54.2 x sqrt(350/1,973,921) = 54.2 x 0.01333 = 0.722

n = 1.34 (W-shape) lambda^(2n) = lambda^(2.68) = 0.722^2.68 = 0.722^2.68

0.722^2 = 0.521. 0.521^1.34 = 0.521^(1.34) = exp(1.34 x ln(0.521)) = exp(1.34 x (-0.652)) = exp(-0.874) = 0.417.

Cr = 0.90 x 13,600 x 350 x (1 + 0.417)^(-1/1.34) = 4,284,000 x 1.417^(-0.746) = 4,284,000 x 0.759 = 3,252,000 N = 3,252 kN > 2,800 kN. OK.

Step 4 — Moment resistance (minor axis): Mry = phi x Zy x Fy = 0.90 x 450 x 10^3 x 350 / 10^6 = 141.8 kNm > 60 kNm. OK.

Step 5 — Interaction (Clause 13.8.2): Cf/Cr + 0.85 x Mfx/Mrx + 0.60 x Mfy/Mry <= 1.0

2800/3252 + 0.85 x 0 + 0.60 x 60/141.8 = 0.861 + 0 + 0.254 = 1.115 > 1.0. FAILS.

The column fails the interaction check due to the combined axial and bending. Solutions:

Alternatively, refine the K factor: If the column is part of a braced frame with feasible beam stiffnesses (W460 or larger framing into the strong axis), Kx may reduce to 0.80. The weak axis Ky remains at 1.0 (pinned base + simple beam connection in weak axis). The strong-axis KL/rx reduction does not change the weak-axis governing KL/ry.

Better approach: Re-evaluate Ky. If the floor slab provides some rotational restraint in the minor axis direction (common in Canadian construction where slabs bear onto the column web through shear studs): Ky = 0.85.

KL/ry = 0.85 x 4200 / 77.5 = 46.0. lambda = 46.0 x 0.01333 = 0.613. lambda^2.68 = 0.613^2.68 = 0.613^2 x 0.613^0.68 = 0.376 x 0.707 = 0.266. Cr = 4,284,000 x 1.266^(-0.746) = 4,284,000 x 0.838 = 3,590 kN. Cf/Cr = 2800/3590 = 0.780. Interaction: 0.780 + 0.254 = 1.034 — still marginal.

Final specification: W310x129, 350W, or provide a more refined capacity-design analysis.

Practical Canadian Column Notes

Storey-height column stepping: Canadian practice strongly favours stepping columns at 2-3 storey intervals (e.g., W310x158 at ground-2, W310x107 at 3-4, W310x74 at 5-6). This saves 8-15% of column tonnage and matches the load profile. Splicing at 1.2 m above finished floor is standard for erection access.

Column orientation: Orient W-shape columns with the web parallel to the strong-axis bending direction. In Canadian buildings, this typically means the web runs in the E-W direction of the structural grid, with beams framing into the flanges N-S and into the web E-W.

Base fixity: Canadian practice typically assumes pinned column bases (K = 1.0 at foundation) unless the base plate, anchor bolts, and foundation are specifically designed for moment transfer. Achieving meaningful base fixity requires anchor bolts outside the column footprint and a base plate designed as a moment connection — this is expensive and is avoided where possible.

Frequently Asked Questions

What is the maximum KL/r for columns in CSA S16? Clause 13.3.2 limits KL/r to 200 for main compression members and 300 for secondary members (bracing, struts in light framing). However, economical column design targets KL/r <= 80-100. Columns exceeding KL/r = 120 lose more than 60% of their squash load capacity and become uneconomical. For HSS columns, KL/r up to 120 is still reasonable because the n = 2.24 curve provides higher capacity at intermediate slenderness.

What n-factor should I use for a welded built-up I-section column? Per CSA S16 Table 4, a welded I-section with flame-cut flange plates and fillet-welded web-to-flange connections uses n = 1.34 if the residual stress level is moderate (typically, preheat and controlled cooling produce lower residual stresses). If the section uses submerged arc welded (SAW) heavy fillets without post-weld heat treatment, residual stresses can reach 0.4-0.5 Fy, and n = 0.69 applies. The fabricator should be consulted on the welding procedure to determine which n-factor is appropriate.

When does biaxial bending govern over uniaxial + axial? Biaxial bending governs for corner columns (beams framing in from two orthogonal directions) and for columns in the perimeter frame subject to major-axis bending from the frame action plus minor-axis bending from cladding eccentricity. The 0.60 coefficient on the minor axis in the Clause 13.8.2 interaction makes biaxial bending less penalising than might be expected — a 30% minor-axis utilisation only consumes 18% of the interaction equation capacity.


This page is for educational reference. Column design per CSA S16-19 Clause 13.3. Verify effective length factors, section classification, and n-factors against the CISC Handbook. All structural designs must be independently verified by a licensed Professional Engineer. Results are PRELIMINARY — NOT FOR CONSTRUCTION.