CELE Steel & Timber Design — Steel Compression MembersSummary

For an ideal column with pinned ends and no initial imperfections, the elastic buckling stress is F_e = π²E/(KL/r)². Here, E is Young's modulus (200,000 MPa for steel), KL is the effective length (K is the effective-length factor, L is the unsupported length), and r is the radius of gyration about the buckling axis. The member becomes unstable at this stress regardless of yield strength when the slenderness ratio KL/r exceeds the transition threshold. This is the cornerstone concept: a slender, high-quality steel column may buckle at a stress far below F_y if it is too long relative to its cross-sectional dimensions.

The boundary between inelastic and elastic buckling is (KL/r)_transition = 4.71√(E/F_y). Equivalently, this occurs when F_y/F_e = 2.25. For E = 200,000 MPa: For F_y = 248 MPa (Grade A36), transition ≈ 133.7; for F_y = 345 MPa (Grade A572-345), transition ≈ 113.4; for F_y = 415 MPa (Grade A572-415), transition ≈ 103.6. Columns with KL/r below this threshold are stocky and governed by inelastic buckling formulas; those above are slender and governed by elastic buckling. This is the key decision point in every column problem.

When KL/r ≤ 4.71√(E/F_y), buckling occurs in the inelastic range where strains are appreciable but the member has not yet yielded uniformly. AISC 360 / NSCP 2015 uses the formula F_cr = [0.658^(F_y/F_e)] × F_y. The exponent base 0.658 is an empirical constant derived from column test data and accounts for residual stresses, mill imperfections, and initial out-of-straightness. As F_y/F_e increases (shorter, stockier columns), the exponent becomes larger, and 0.658 raised to a larger power yields a smaller value, correctly reducing F_cr from F_y. At the transition (F_y/F_e = 2.25), the formula gives F_cr ≈ 0.5F_y.

When KL/r > 4.71√(E/F_y), buckling occurs elastically (like an Euler column). The critical stress is F_cr = 0.877 × F_e. The 0.877 factor (not 1.0) accounts for initial out-of-straightness per AISC assumptions (L/1000 maximum initial bow). Without this reduction, thin, long columns would be overpredicted. In this regime, F_cr drops rapidly with increasing KL/r (proportional to 1/(KL/r)²), so slenderness dominates design.

Design axial strength is P_n = F_cr × A_g, where A_g is the gross cross-sectional area. For LRFD (Limit States Design, used in PH), the design strength is φ_c × P_n with resistance factor φ_c = 0.90. For ASD (Allowable Stress Design), the nominal strength is divided by Ω_c = 1.67 (equivalent to φ = 0.60). In Philippine practice (NSCP 2015), LRFD is the primary method; ASD is provided as an alternative. The φ_c = 0.90 reflects the relatively high confidence in column buckling predictions (lower uncertainty than, e.g., connection failure).

The effective length KL accounts for boundary conditions. K = 1 for pinned-pinned (theoretical); K = 2 for fixed-free cantilever; K = 0.5 for fixed-fixed (ideal case, rarely achieved); K ≈ 0.7–0.8 for fixed-pinned (typical in buildings with strong lateral bracing); K > 1 for unbraced frames. The effective length is the distance between inflection points (zero-moment points) in the buckled shape. AISC / NSCP provide alignment charts to estimate K for frames; for standard problems, K is usually given or can be found in tables. A common error is forgetting that K > 1 in sway (unbraced) frames.

NSCP 2015 / AISC recommends a practical upper limit of KL/r ≤ 200 for compression members in buildings. This limit ensures reasonable stiffness and avoids excessive lateral deflections under small transverse loads (e.g., wind or vibration). Columns with KL/r > 200 become very flexible and may be prone to drift, oscillation, or second-order effects that push them outside the scope of simple buckling formulas. Some jurisdictions (e.g., for main load-bearing columns) may impose KL/r ≤ 150 or tighter limits.

If a column cross-section has **slender elements** (flanges or webs with high width-to-thickness ratios), local buckling may occur before overall flexural buckling. AISC / NSCP handles this via the Q factor: Q = Q_s × Q_a, where Q_s accounts for stiffened elements (webs) and Q_a for unstiffened elements (flanges). When Q < 1, the effective yield stress is reduced: F_y,eff = Q × F_y. For compact sections (typical hot-rolled shapes), Q = 1. For slender sections (e.g., built-ups with thin plates), Q may be 0.6–0.9, significantly reducing capacity. The formulas for Q are in AISC Table B4.1 (NSCP equivalent).

Doubly symmetric sections (e.g., I-beams, box sections) buckle by **flexural buckling** (bending about one principal axis). Singly symmetric sections (e.g., channel shapes, T-sections) and unsymmetric sections may buckle by **torsional** or **flexural-torsional buckling**, where the member twists as well as bends. The critical stress is the smaller of the flexural-buckling stress (about x or y axis) and the torsional/flexural-torsional stress. AISC / NSCP provides formulas for these modes; singly symmetric sections require additional calculation. In practice, most building columns are doubly symmetric, so flexural buckling dominates; but the PE exam occasionally includes singly symmetric cases.

The critical stress equations depend on E = 200,000 MPa (constant for steel) and F_y (yield stress, varies by grade). Higher F_y shifts the transition slenderness to a smaller value, meaning high-strength steel benefits less from length than mild steel (per unit increase in length). For example, a column with KL/r = 120 might be inelastic in Grade A36 (F_y = 248) but elastic in Grade A572-415 (F_y = 415). This counterintuitive result—that stronger steel is more prone to elastic buckling for given geometry—is a key insight: buckling limitations mean you cannot always substitute high-strength steel without redesigning the section.

Standard boardwork steps: (1) Extract A_g, r (smallest radius), KL; (2) Calculate KL/r; (3) Calculate transition ratio 4.71√(E/F_y); (4) Compare KL/r to transition to pick inelastic or elastic formula; (5) Calculate F_e = π²E/(KL/r)²; (6) If inelastic, compute F_y/F_e, then F_cr = [0.658^(F_y/F_e)]F_y; if elastic, F_cr = 0.877 F_e; (7) Calculate P_n = F_cr × A_g; (8) Apply φ_c = 0.90 to get φ_c P_n (LRFD) or divide by Ω_c = 1.67 for ASD; (9) State final answer with units (kN).

The slenderness ratio is the primary input. It is compared to the transition ratio (4.71√(E/F_y)) to determine which of two F_cr formulas applies. This decision tree is the master control for all subsequent calculations.

Material properties (F_y via grade, E fixed at 200,000 MPa) set the transition threshold. Higher F_y lowers the transition, shifting more columns into the elastic regime. This explains why high-strength steel is less efficient for very slender members.

Boundary conditions (pinned, fixed, cantilever, or frame conditions) determine K. K directly multiplies L to form the effective length KL, which feeds into KL/r. Stronger bracing or frame action (K < 1) dramatically improves buckling resistance.

Once F_cr is found, nominal strength P_n = F_cr × A_g. The design strength φ_c P_n = 0.90 × P_n is the final deliverable. Omitting the 0.90 factor is a fatal error.

If the cross-section has slender elements, the Q-factor reduces the effective yield stress. This feeds back into the F_cr calculation, lowering the final answer. Skipping the Q-factor check overpredicts capacity.

Doubly symmetric sections have only flexural buckling (about x and y axes). Singly symmetric and unsymmetric sections must also check torsional and flexural-torsional modes. The minimum F_cr governs design.

The slenderness limit ensures not only buckling safety but also acceptable stiffness under service loads. Exceeding it signals the need for re-proportioning to avoid drift and vibration issues.

Unlike older codes that used separate formulas with discontinuities, the modern unified curve transitions smoothly. The 0.658 exponent and 0.877 factor are empirical fits that ensure continuity and match test data across all slenderness ranges.