CELE Strength of Materials — Columns and BucklingSummary

A short compression member fails by **crushing** when axial stress σ = P/A reaches the material's yield or ultimate strength. A long, slender member fails much earlier by **elastic buckling**—a lateral deflection instability triggered at the critical (Euler) load, well below material strength. The distinction hinges on the **slenderness ratio** KL/r. Short columns (low KL/r) are stress-governed; long columns (high KL/r) are stiffness-governed.

For a slender, elastic, pin-ended column of length L, modulus E, and second moment of inertia I (least axis), the critical (buckling) load is: **P_cr = π²EI/L²**. The critical stress is: **σ_cr = π²E/(L/r)²**, where r = √(I/A) is the radius of gyration and L/r is the slenderness ratio. The formula is valid only while σ_cr remains below the proportional limit (typically 0.5σ_y in steel due to residual stresses), i.e., for genuinely slender columns above a threshold slenderness C_c = √(2π²E/σ_y).

Real columns have various end restraints, not just pinned ends. The **effective length** L_e = KL accounts for these: substitute L_e into Euler's formula: **P_cr = π²EI/(KL)²**. The K factor (0 < K ≤ 2.0) depends on boundary conditions: pinned–pinned K = 1.0; fixed–fixed K = 0.5 (theoretical, 0.65 design per NSCP/AISC); fixed–pinned K = 0.7 (theoretical, 0.80 design); fixed–free (cantilever) K = 2.0 (theoretical, 2.10 design). Fixing the ends reduces L_e, raising capacity exponentially (since P ∝ 1/(KL)²). A fixed–fixed column carries ~4× the pin-ended one (K² ratio = 0.25).

The dimensionless slenderness ratio **KL/r** segregates columns into three regimes: (1) **Short columns** (KL/r low, typically < 50 for steel): fail by yielding/crushing; design formula is P = σ_y A. (2) **Intermediate columns** (moderate KL/r, ~50–130 for steel): fail by inelastic buckling; require empirical formulas (Rankine, NSCP/AISC curves). (3) **Long (slender) columns** (KL/r > C_c ≈ 125 for typical steel): fail by elastic (Euler) buckling; Euler formula applies. The transition boundary C_c = √(2π²E/σ_y) is material- and proportional-limit dependent.

The Rankine–Gordon formula bridges short and long buckling regimes: **P_cr = σ_y A / [1 + a(L_e/r)²]**, where a is a material constant (~1/7500 for steel with σ_y ≈ 250 MPa). For short columns (L_e/r → 0), P_cr → σ_y A (crushing); for long columns (L_e/r → ∞), P_cr ∝ 1/(L_e/r)² (Euler-like decay). The parabolic denominator smoothly transitions between regimes. This formula was standard in older codes and remains a board-exam favorite because it captures the real (inelastic) behavior of intermediate columns, which Euler alone misses.

Modern codes (NSCP 2015, AISC 360-16 LRFD) use a continuous flexural-buckling stress F_cr based on the elastic buckling stress F_e = π²E/(KL/r)². For compact sections, the critical stress is: **F_cr = (0.658)^(λ_c²) F_y** if λ_c ≤ 1.5 (inelastic), or **F_cr = 0.877/λ_c² F_y** if λ_c > 1.5 (elastic), where λ_c = √(F_y/F_e) is the column strength parameter. This unified approach avoids the discrete C_c threshold and automatically interpolates between inelastic and elastic regimes. The nominal buckling strength is P_n = F_cr A.

A column **always buckles about its weakest axis**, i.e., the axis with the **least** moment of inertia I_min. For a rectangular section, if I_x > I_y, the column buckles in the y-direction. For a W-section, buckling is typically about the weak (y) axis unless bracing prevents it. The radius of gyration r = √(I/A), so the smallest I gives the smallest r and the largest slenderness ratio KL/r for the same K and L. Thus, to find the buckling load, always compute KL/r using r_min (based on I_min). Using the larger I is a common exam pitfall that yields non-conservative (dangerous) results.

When an axial load P is applied at eccentricity e (offset from the neutral axis), bending moments develop, and the maximum compressive stress becomes: **σ_max = (P/A)[1 + (ec/r²) sec(L_e/(2r)√(P/(EA)))]**, where c is the distance to the extreme fiber. The secant term grows as P approaches the buckling load, and even small eccentricities can produce large stresses. For stocky columns (small L_e/r), the simpler superposition **σ = P/A + Mc/I** with M = Pe is adequate. For slender eccentric columns, the secant formula must be used or iterative methods applied (e.g., NSCP/AISC interaction formulas in combined compression and bending).

Real columns are not perfectly straight (initial crookedness δ₀ ~ L/500 to L/1000) and contain residual stresses from fabrication (e.g., welding, cold rolling). These imperfections lower the buckling load below the ideal Euler prediction. For steel, residual stresses are typically on the order of 0.1–0.3σ_y and cause the material to begin yielding before P reaches P_cr (Euler), leading to inelastic buckling. The C_c transition and the NSCP/AISC formulas (which use 0.5σ_y as the proportional limit, not σ_y) implicitly account for these imperfections. This is why Euler over-predicts capacity for intermediate columns.

If a column is laterally braced (restrained against deflection) at intermediate points, the unbraced length L used in KL/r is the distance between braced points, not the full height. For example, a 10 m column with lateral bracing at 2 m intervals has an unbraced length of 2 m. This greatly reduces the slenderness ratio and increases capacity. NSCP 2015 specifies where bracing must be placed (e.g., near the top, every floor height, at roof/floor levels). In the design of compression members in trusses and frames, bracing is critical to economical design.

Timber and concrete columns are designed using codes specific to those materials. Timber (NSCP 2015 NDS provisions) uses the slenderness ratio Le/d (where d is the least dimension) and applies reduction factors for intermediate buckling. Concrete (ACI 318) uses the slenderness ratio kL_u/r (where L_u is the unsupported length, k is the effective-length factor, r is the radius of gyration, and slenderness limit ~22 for pinned, ~45 for fixed) and applies a reduction factor φ_c to account for buckling and material variability. Both codes provide tables and charts for rapid design. Steel uses AISC 360 and NSCP 2015 (which adopts AISC for steel).

The slenderness ratio KL/r determines which failure mode governs: if KL/r < ~50, use crushing formulas (P = σ_y A); if 50 < KL/r < C_c, use Rankine or NSCP inelastic buckling formulas; if KL/r > C_c, use Euler or NSCP elastic formulas. This hierarchy avoids using the wrong theory.

Euler's formula P_cr = π²EI/L² assumes pinned ends (K = 1.0). To account for fixed, partial-fixed, or free ends, substitute L_e = KL: P_cr = π²EI/(KL)². The K value (theoretically 0.5 to 2.0, design 0.65 to 2.1) bridges ideal theory and practical restraint.

The buckling load P_cr is divided by the cross-sectional area to yield critical stress σ_cr. In ASD (older codes), this is compared directly to allowable stress (e.g., F_allow ~ σ_y / safety factor). In LRFD (NSCP 2015, AISC 360), the nominal buckling strength P_n = F_cr × A is factored (e.g., φ_c × P_n) and compared to factored demand (1.2 dead + 1.6 live). Understanding both is important for comprehensive design.

Rankine's single parabolic formula is easy to apply by hand but is an approximation. Modern codes use piecewise (inelastic + elastic) or continuous unified functions (e.g., the 0.658^(λ_c²) form in AISC). All aim to capture the smooth transition from crushing (short) through inelastic buckling (intermediate) to elastic buckling (long), accounting for imperfections.

Small load eccentricities create bending moments M = P × e that amplify stress. The secant formula (nonlinear in P) accurately models this; for stocky columns, the simpler combined-stress approximation σ = P/A + Mc/I suffices. For columns under combined axial and bending loads, NSCP/AISC interaction formulas (P/P_allow + M/M_allow ≤ 1.0 or similar) provide a practical design check.

Higher E (steel vs. wood) and lower σ_y shift C_c = √(2π²E/σ_y) to larger values, expanding the elastic (Euler) regime. Timber has lower E, so C_c is smaller and more timbers fall into the intermediate or short range. This explains why timber column formulas emphasize the intermediate range.

Bracing at intermediate points shortens L in KL/r, reducing the slenderness ratio. Since P_cr ∝ 1/(KL)², even modest bracing intervals dramatically raise capacity. This relationship is exploited in efficient structural design (e.g., floor-by-floor bracing in tall frames).