Stainless Steel Design Guide — AISC DG27 and AISC 370

Stainless steel is not just "steel that doesn't rust." Its structural behavior differs from carbon steel in three fundamental ways: (1) the stress-strain curve is non-linear with no well-defined yield plateau, (2) cold-working during fabrication (bending, rolling, stretching) increases yield strength substantially — an effect that carbon steel design ignores, and (3) the buckling curves are more generous because stainless steel's gradual yielding provides post-buckling reserve strength that carbon steel does not have.

The governing design standards are AISC Design Guide 27 (Structural Stainless Steel) and AISC 370 (Specification for Structural Stainless Steel Buildings, to be incorporated into AISC 360 in the next cycle). Both use the same limit-states framework as AISC 360 but with material-specific modifications. This guide covers the non-linear material model, cold-worked strength enhancement, stainless-specific buckling provisions, welding of stainless steel, and a worked example comparing a stainless column to a carbon steel equivalent.

PRELIMINARY — NOT FOR CONSTRUCTION. This guide is for educational reference. All designs must be independently verified by a licensed Professional Engineer.

The non-linear stress-strain curve

Carbon steel has a linear elastic response up to a well-defined yield point, then a yield plateau, then strain hardening. Stainless steel has no yield plateau — it transitions smoothly from elastic to plastic, and the stress-strain relationship is commonly modeled with the Ramberg-Osgood equation:

ε = σ/E + 0.002 × (σ/F_y)^n

Where:

• ε = total strain
• σ = applied stress
• E = elastic modulus (approximately 193 GPa for austenitic stainless, vs. 200 GPa for carbon steel)
• F_y = 0.2% offset yield strength
• n = strain-hardening exponent

The exponent n characterizes the sharpness of the yield knee. For austenitic stainless steels (Types 304/304L, 316/316L), n ≈ 5–8. For duplex stainless steels (2205), n ≈ 8–12. For carbon steel, n → ∞ (no gradual yielding). A lower n means more gradual yielding, which means:

• Column and plate buckling capacities are higher because the tangent modulus E_t degrades more slowly as stress approaches F_y.
• Plastic moment redistribution is more complete because the strain at the tension edge continues to increase without fracture.
• The shape factor (Z/S) behaves differently — the Ramberg-Osgood curve at n = 5 shows significant hardening beyond the 0.2% offset.

Common stainless steel grades for structural use

The yield-to-tensile ratio F_y/F_u ranges from 0.4 (annealed 304) to 0.72 (duplex 2205). Carbon steel's ratio is typically 0.70–0.85. The lower ratio for stainless means more ductility and strain capacity before fracture — beneficial for seismic applications where large plastic strains may be required.

Cold-worked strength enhancement

This is the least appreciated aspect of stainless steel design. When stainless steel is plastically deformed at room temperature — during roll-forming, press-braking, or stretch-forming — its yield strength increases substantially. The AISC DG27 explicitly allows designers to use the post-cold-worked strength in the final member, provided the cold work is documented and the ductility requirements are still met.

The strength increase from cold working depends on the amount of plastic strain and the alloy:

Typical strength gains from cold forming:

The mechanism is work hardening: plastic deformation creates dislocations that impede further dislocation motion, effectively increasing the stress required for continued plastic flow. Austenitic stainless steels (304, 316) work-harden much more than duplex grades because their face-centered cubic crystal structure has more active slip systems.

Design approach (AISC DG27 Section 3.2.2)

The designer may use either:

1. Annealed properties — conservative, always acceptable. Use the mill-certified annealed F_y.
2. Cold-worked strength at corners — for press-braked or roll-formed shapes, use enhanced F_y at the corner regions (within a distance 5t from the bend tangent). The flat portions of the cross-section use annealed properties.
3. Average enhanced strength — for members predominantly cold-worked (e.g., cold-rolled structural tubing), use a weighted average of the corner and flat-region strengths.

The weighted-average approach per AISC DG27 Section 3.2.2:

F_ya = C × F_yc + (1 − C) × F_yf

Where:

• C = ratio of corner area to total cross-sectional area
• F_yc = enhanced yield strength at corners after cold work
• F_yf = annealed yield strength in flat regions

For a typical cold-formed stainless C-section, C ≈ 0.15–0.25, and the weighted average F_y is approximately 1.10–1.18 × the annealed F_y.

Buckling curves for stainless steel

AISC 370 provides separate column and beam buckling curves for stainless steel that are more generous than the carbon steel curves. This is because the gradual yielding (low n) means the tangent modulus E_t remains above zero near F_y, providing post-buckling capacity that carbon steel's elastic-perfectly-plastic model excludes.

Column buckling (AISC 370 Section E3)

The critical stress F_cr for flexural buckling uses a modified curve:

F_cr = 0.658^(F_y/Fe) × F_y   for KL/r ≤ 4.71 × sqrt(E/F_y)
F_cr = 0.877 × F_e              for KL/r > 4.71 × sqrt(E/F_y)

Where F_e = π² × E / (KL/r)². This is the same functional form as AISC 360 Chapter E, but the resistance factor φ_c for stainless is 0.85 (vs. 0.90 for carbon steel). The 0.05 lower φ_c reflects the reduced statistical database for stainless member tests.

Lateral-torsional buckling (AISC 370 Section F3)

The LTB curve for stainless I-shapes uses the same three-zone format as carbon steel (plastic, inelastic LTB, elastic LTB) but with a different transition point:

L_p(stainless) = 1.76 × r_y × sqrt(E/F_y) × (1 + 0.25 × (F_y/F_u))

The additional term (1 + 0.25 × F_y/F_u) recognizes that the lower F_y/F_u ratio of stainless steel (≈ 0.4 for 304 vs. ≈ 0.77 for A992) provides more inelastic rotation capacity before local buckling initiates. This extends the plastic LTB zone by approximately 10–15% for annealed austenitic stainless.

Welding of stainless steel

Welding stainless steel requires different filler metals, different shielding gas, and different post-weld treatment than carbon steel:

Filler metal selection

Shielding gas

GTAW (TIG) of stainless requires 100% argon — no CO₂ or O₂ additions, which would oxidize the chromium and reduce corrosion resistance. GMAW (MIG) typically uses Ar + 2% O₂ or Ar + 2% CO₂ for spray transfer, but the oxygen content is kept low to minimize chromium oxidation.

Post-weld treatment

Three post-weld treatments are common for stainless structural welds:

1. Pickling and passivation (ASTM A380): Removes heat tint (chromium oxide discoloration) and restores the passive chromium oxide layer. Required for corrosive service.
2. Mechanical descaling (grinding, brushing): Removes weld slag and light heat tint. Suitable for non-corrosive interior applications.
3. Solution annealing (1,040°C for 304, water quench): Restores full corrosion resistance and annealed mechanical properties. Rarely practical for large structural assemblies due to distortion.

Critical note: Carbon steel tools (grinding wheels, wire brushes, clamps) must NEVER contact stainless steel. Carbon steel particles embedded in the stainless surface will rust and initiate pitting corrosion. Dedicated stainless-only tools are required.

Worked example: stainless column vs. carbon steel column

A 3.5 m long column in an exterior architectural feature carries a factored axial load P_u = 400 kN. The column is pin-ended about both axes (K = 1.0) and exposed to a coastal environment. Compare a 304 stainless steel rectangular hollow section to an equivalent carbon steel HSS section.

Option A: Stainless steel 304 RHS 150 × 100 × 5 mm

From AISC DG27 or manufacturer data for annealed 304:

• F_y = 205 MPa (annealed). Cold-worked average (corners enhanced): F_y ≈ 220 MPa.
• F_u = 515 MPa
• E = 193 GPa

Section properties (actual from supplier catalog):

• A = 2,370 mm²
• I_x = 7.59 × 10⁶ mm⁴, r_x = 56.6 mm
• I_y = 4.18 × 10⁶ mm⁴, r_y = 42.0 mm (governing for weak-axis buckling)
• b/t of flat width = (150 − 3 × 5) / 5 = 135/5 = 27.0

Check local buckling (AISC 370 Table B4.1a for HSS): λ_r = 1.40 × sqrt(E/F_y) = 1.40 × sqrt(193,000/205) = 1.40 × 30.68 = 43.0. b/t = 27.0 < 43.0 — the wall is non-slender. OK.

Column slenderness (weak axis): KL/r_y = 3,500 / 42.0 = 83.3. F_e = π² × 193,000 / 83.3² = 1.904e6 / 6,939 = 274.4 MPa. F_y / F_e = 205 / 274.4 = 0.747. F_cr = 0.658^(0.747) × 205 = 0.658^0.747 × 205 = 0.734 × 205 = 150.5 MPa. φ_c × P_n = 0.85 × 150.5 × 2,370 / 1,000 = 0.85 × 356.7 = 303.2 kN.

Utilization: 400 / 303 = 1.32 — FAIL. The section is too light.

Try 150 × 100 × 6 mm: A = 2,810 mm². I_y = 4.87 × 10⁶ mm⁴, r_y = 41.7 mm. KL/r_y = 3,500 / 41.7 = 83.9. F_e = 1.904e6 / 83.9² = 270.5 MPa. F_y/F_e = 205/270.5 = 0.758. F_cr = 0.658^0.758 × 205 = 0.731 × 205 = 149.9 MPa. φ_c × P_n = 0.85 × 149.9 × 2,810 / 1,000 = 0.85 × 421.2 = 358.0 kN.

Still fails. The light wall thickness and moderate F_y mean the column is buckling-governed.

Try 200 × 100 × 6 mm (increase depth): A = 3,410 mm². I_y ≈ 6.5 × 10⁶ mm⁴, r_y ≈ 43.7 mm. KL/r_y = 3,500 / 43.7 = 80.1. F_e = 1.904e6 / 80.1² = 296.8 MPa. F_y/F_e = 205/296.8 = 0.691. F_cr = 0.658^0.691 × 205 = 0.749 × 205 = 153.5 MPa. φ_c × P_n = 0.85 × 153.5 × 3,410 / 1,000 = 0.85 × 523.4 = 444.9 kN. Utilization: 400 / 445 = 0.90 — OK.

Option B: Carbon steel HSS 152 × 102 × 6.4 mm (A500 Gr C)

F_y = 345 MPa, E = 200 GPa. A = 3,190 mm², r_y = 37.6 mm. KL/r_y = 3,500 / 37.6 = 93.1. F_e = π² × 200,000 / 93.1² = 227.8 MPa. F_y/F_e = 345/227.8 = 1.515 > 1.0 → elastic buckling governs. F_cr = 0.877 × 227.8 = 199.8 MPa. φ_c × P_n = 0.90 × 199.8 × 3,190 / 1,000 = 573.6 kN. Utilization: 400 / 574 = 0.70.

Comparison:

The carbon steel HSS is 7% lighter and has 29% more capacity at a lower cost — but it would require a multi-layer marine coating system for the coastal environment, which adds $30–50/m² and requires periodic maintenance. The stainless option has higher material cost ($8–12/kg vs. $1.50–2.50/kg for carbon steel) but zero maintenance cost over a 50-year service life in a coastal environment. Life-cycle cost analysis typically favors stainless for exterior architectural applications with a design life exceeding 30 years.

Key takeaways

• Stainless steel has no yield plateau. The stress-strain curve is continuous and is modeled with the Ramberg-Osgood equation. The strain-hardening exponent n (5–8 for austenitic, 8–12 for duplex) determines the sharpness of the yield knee and affects buckling strength.

• Cold-working during fabrication increases yield strength by 10–25% at corners and bends. AISC DG27 allows designers to account for this enhancement using a weighted-average approach (corner area ratio × enhanced F_y + flat area ratio × annealed F_y).

• Column and LTB buckling curves in AISC 370 are slightly more generous than the carbon steel curves because the gradual yielding provides post-buckling reserve. The resistance factor φ_c = 0.85 for stainless (vs. 0.90 for carbon steel) reflects the smaller test database.

• Welding stainless requires alloy-specific filler metals (E308L for 304, E316L for 316, E309L for stainless-to-carbon-steel), 100% argon shielding gas for TIG, and post-weld pickling/passivation for corrosive service. Carbon steel tools must never contact stainless surfaces to avoid embedded iron contamination.

FAQ

When should I specify stainless steel instead of carbon steel?

Stainless is justified when: (1) the environment is corrosive (coastal, chemical plant, swimming pool, wastewater treatment) and the cost of coating maintenance over the design life exceeds the stainless premium; (2) the architectural specification requires a bare metal finish; (3) the member is inaccessible for maintenance (embedded in concrete, behind permanent cladding); or (4) hygiene requirements prohibit coatings that can chip or flake (food processing, pharmaceutical). For most interior building framing, carbon steel with paint or galvanizing is more cost-effective.

Why is the yield-to-tensile ratio F_y/F_u so low for stainless steel?

The low F_y/F_u ratio (≈ 0.40 for annealed 304 vs. ≈ 0.77 for A992 carbon steel) is inherent to the face-centered cubic crystal structure of austenitic stainless. F_y is defined by the 0.2% offset strain rather than a true yield point, and the extensive strain hardening between the 0.2% offset and the ultimate tensile strength (at 40–60% strain) provides large ductility. This is advantageous for seismic energy dissipation but means a larger cross-section is needed to meet the same strength demand at service loads.

Can stainless steel be galvanized for additional protection?

No. Galvanizing (hot-dip zinc coating) does not adhere reliably to stainless steel because the chromium oxide passive layer prevents the zinc-iron intermetallic reaction. Stainless steel is inherently corrosion-resistant and does not need galvanizing. If additional protection is desired, specify a higher-alloy grade (e.g., 316 instead of 304 for coastal, or duplex 2205 instead of 316 for high-chloride splash zones).

What is the difference between 304L and 316L?

Both are austenitic stainless steels with the same base composition (18% Cr, 8% Ni). 316L adds 2–3% molybdenum, which dramatically improves resistance to pitting corrosion from chlorides (salt water, de-icing salts). The pitting resistance equivalent number (PREN = %Cr + 3.3×%Mo + 16×%N) is approximately 19 for 304L and 25 for 316L. In coastal environments within 5 km of salt water, 316L is recommended. In interior or non-coastal exterior applications, 304L is sufficient.