ISO 19902 Fixed Steel Offshore Structures

ISO 19902 is the international reference standard for fixed steel offshore platforms — jacket structures, compliant towers, and gravity-based platforms with steel topsides. Written under ISO Technical Committee 67 (Oil and gas industries), it covers structural design, fabrication, installation, in-service inspection, and structural assessment. If you work in offshore oil and gas outside a purely Norwegian regulatory context, ISO 19902 is likely the primary structural standard — and it is explicitly referenced by both DNV and NORSOK as an acceptable alternative basis for design.

1. Scope and Applicability

ISO 19902:2007 + Amendment 1 (2020) applies to:

• Fixed steel offshore structures: jacket platforms, jack-up legs, compliant towers, and mono-towers
• All phases: design, fabrication, installation, operation, and decommissioning
• Environments: open sea, arctic, and moderate-climate locations
• Target: primary and secondary structural steel members and connections

ISO 19902 does not cover floating structures (FPSO, semi-submersible, TLP) — those fall under ISO 19904-1 and DNV-ST-F101 / DNV-OS-C101 for floating applications.

2. Structural Design Principles

2.1 Limit State Philosophy

ISO 19902 uses a Load and Resistance Factor Design (LRFD) approach — the same philosophy as DNV-OS-C101 but with different partial factor values calibrated to global offshore practice:

The four limit states are:

• ULS (Ultimate Limit State) — 100-year return period environmental loads
• ALS (Abnormal Limit State) — 10 000-year environmental or accidental (dropped object, blast, fire)
• FLS (Fatigue Limit State) — cumulative damage over design life
• SLS (Serviceability Limit State) — deflections, vibrations, equipment operability

2.2 Structural Member Design

ISO 19902 §13 covers tubular and non-tubular member design. For tubular members (the dominant form in jacket structures), utilisation checks are required for:

• Axial tension and compression (including column buckling)
• Bending (strong and weak axis)
• Shear and torsion
• Combined loading (unity check format): U_c = (f_a/F_a) + (f_b/F_b) ≤ 1.0

Column buckling uses an effective length factor K that accounts for end fixity — for jacket bracing K = 0.8 (both ends pinned in practice), for legs K = 1.0 to 1.5 depending on the deck-to-leg connection stiffness.

3. Tubular Joint Design — Punching Shear

The most distinctive and technically demanding aspect of ISO 19902 is its treatment of tubular joint capacity. Jacket frames use welded tubular joints (K, T, Y, X-joints) where chord and brace walls interact through complex stress fields. ISO 19902 §14 defines the primary method.

3.1 Joint Classification

3.2 Punching Shear — The Governing Failure Mode

The governing failure mode in most tubular joints is punching shear — the brace wall "punches" through the chord wall. ISO 19902 §14.3 checks this using the nominal load approach:

Key geometric parameters governing tubular joint strength:

3.3 Joint Can — Thickened Chord at the Joint

When punching shear capacity is insufficient, the standard solution is a joint can — a locally thickened section of chord wall at the joint location. ISO 19902 §14.4 permits the joint can thickness T_c to replace T in the capacity formula if:

• The joint can extends at least 0.25D beyond the brace footprint on each side
• The wall thickness transition from chord to can does not exceed 1:4

4. Worked Example: K-Joint Punching Shear Capacity Check

The §14 formula above is compact but every term hides a decision. Walk one K-joint through the full capacity check so the chord-load coupling, joint-can sizing, and unity-check sequence become concrete — and so the typical "Q_f caught me late" story has a number attached.

Joint: K-joint at the upper bracing panel of a fixed steel jacket. Chord: D = 1200 mm, T = 30 mm, S420 (f_y = 420 MPa). Two diagonal braces: d = 700 mm, t = 20 mm, both at θ = 45° to chord, gap g = 80 mm. Brace 1 carries axial tension P₁ = 4500 kN; Brace 2 carries balancing axial compression P₂ = 4500 kN (K-joint loads balance). Chord carries N_chord = 6000 kN axial compression and M_chord = 1800 kNm in-plane bending under the governing ULS combination.

Step 1 — Geometric parameters

• β = d/D = 700 / 1200 = 0.583
• γ = D / 2T = 1200 / 60 = 20
• τ = t/T = 20 / 30 = 0.667
• ζ = g/D = 80 / 1200 = 0.067 (small positive gap K-joint)
• θ = 45°, sin θ = 0.707

Step 2 — Q_u: joint strength factor

For a K-joint under balanced axial load per ISO 19902 §14.3 Table 14.3-1:

Gap factor Q_g = 1 + 0.2(1 − 2.8ζ)³ for ζ ≥ 0.05 → Q_g = 1 + 0.2 × (1 − 0.187)³ ≈ 1.107. So Q_u = 23.2.

Step 3 — Q_f: chord load influence factor (the catch)

Per §14.3.4, Q_f = 1 − λ · A · B², where A is the chord stress utilisation and B accounts for joint type:

• Chord axial stress: σ_ax = 6000×10³ / (π × 1200 × 30) ≈ 53.1 MPa
• Chord bending stress: σ_b = 1800×10⁶ × (D/2) / I; I ≈ π/8 × D³ × T = π/8 × 1200³ × 30 ≈ 2.04×10¹⁰ mm⁴; σ_b ≈ 53.0 MPa
• Combined chord utilisation: A = (σ_ax + σ_b) / f_y = 106.1 / 420 = 0.253
• For K-joint axial load, λ = 0.030, B = 1.0
• Q_f = 1 − 0.030 × 0.253 × 1.0² = 0.992

Low chord utilisation (~25%) → Q_f stays close to 1.0. The danger is when chord utilisation climbs past 50%: at A = 0.6, Q_f for a T-joint axial check (λ = 0.30) drops to 1 − 0.30 × 0.6 × 1.0 = 0.82, an 18% capacity haircut applied late in the design when the chord has already been sized.

Step 4 — Punching shear capacity

Step 5 — Design capacity and unity check

Apply the partial factor γ_R,j = 1.05 per §14.3.6:

• P_Rd = 12,300 / 1.05 ≈ 11,700 kN
• UR = P_1,Ed / P_Rd = 4500 / 11,700 = 0.385

UR < 1.0 with significant margin → punching shear PASSES for both braces (loads balance, identical capacity).

Step 6 — Joint can sensitivity

Suppose the chord wall thickness was T = 25 mm (not 30 mm) — a designer tightening for steel weight:

• γ = 24, Q_u = (16 + 1.2 × 24) × 0.524 × 1.107 ≈ 26.0
• P_u = 26.0 × 0.992 × 420 × 25² / 0.707 ≈ 9,600 kN
• P_Rd = 9600 / 1.05 ≈ 9,150 kN
• UR = 4500 / 9150 = 0.492 — still passes, but a 28% capacity reduction for a 5 mm chord wall reduction

The T² term in P_u means a 17% reduction in chord wall thickness gives a 25–30% reduction in joint capacity. This is why joint cans (locally thickened chord at the joint per §14.4) are routinely specified — the chord wall is sized for its general member check, then the joint can locally upsizes to T_c where punching shear governs.

5. Fatigue of Tubular Joints

ISO 19902 §16 defines the fatigue methodology for tubular joints. It uses an S-N approach with hot-spot stress — the same fundamental method as DNV-RP-C203 but with different S-N curves calibrated to ISO data.

4.1 Stress Concentration Factors (SCF)

Hot-spot stress at a tubular joint is obtained by applying SCFs to the nominal brace stress:

4.2 ISO 19902 S-N Curves

ISO 19902 uses two S-N curves for tubular joints:

• T-curve: for tubular joints — log(N) = 12.476 − 3·log(Δσ) for N ≤ 10⁷, then slope m=5
• T'-curve: for improved joints (grinding, weld toe peening)

The T-curve is directly comparable to DNV-RP-C203's T-curve (both derived from the same experimental database). For through-thickness cracks or welds in complex geometries, more detailed assessment may require fracture mechanics.

6. Pile and Foundation Design

ISO 19902 §6 covers the geotechnical design of pile foundations, which are the primary lateral load-resisting system for jacket platforms in soft clay and medium-dense sand.

5.1 Driven Pile Capacity

Pile axial capacity is the sum of skin friction and end bearing:

ISO 19902 gives α-values (adhesion factor for clay) that range from 0.5 to 1.0 depending on normalised shear strength s_u/σ'_v. These are slightly more conservative than the API RP 2GEO values for lightly overconsolidated clays.

5.2 Pile Group Effects

For jacket legs with multiple skirt piles or cluster piles, ISO 19902 §6.8 requires group efficiency calculations. The group capacity is typically 60–80% of the sum of individual pile capacities for closely spaced piles in soft clay — a reduction that significantly affects leg design for deep-water applications.

5.3 Lateral Load Capacity — p-y Analysis

Lateral pile-soil interaction is assessed using p-y curves (lateral soil resistance vs lateral pile displacement). ISO 19902 Annex A provides p-y formulations for:

• Soft clay (Matlock 1970 formulation)
• Stiff clay (Reese formulation)
• Sand (API/Reese formulation with depth-dependent initial modulus)

7. Assessment of Existing Structures

ISO 19902 §21 is uniquely valuable for operators of ageing platforms: it defines a fitness-for-service (FFS) assessment methodology that allows structures designed to older standards to be evaluated against current criteria without necessarily requiring major structural modifications.

6.1 Reserve Strength Ratio (RSR)

The key metric for platform structural adequacy is the Reserve Strength Ratio:

RSR is determined by a pushover analysis — a nonlinear collapse analysis that loads the structure to failure, identifying the weakest members (typically K-joint cans or piles) and the sequence of plastic hinge formation.

6.2 Platform Assessment Triggers

ISO 19902 §21.3 lists triggers requiring a formal structural assessment:

• Design life extension beyond the original design basis
• Topsides load increases exceeding 5% of original operating weight
• Significant corrosion, damage, or degradation of structural members
• Change of service (e.g., change from unmanned to manned operations)
• Updated metocean criteria (revised 100-year Hs from new hindcast)
• Revised inspection findings showing fatigue crack or corrosion damage

8. In-Service Inspection

ISO 19902 §20 defines the in-service inspection programme. Key points:

The inspection programme is risk-based: high-consequence, high-utilisation joints receive more frequent and more detailed inspection. ISO 19902 §20.5 permits inspection interval extension where inspection history shows consistently clean results.

9. Cross-Reference Map

10. ISO 19902 vs DNV-OS-C101: Key Differences

11. Common Pitfalls and Errors

• Classifying KT-joints as pure K-joints without decomposing into K + T components — the T-component can be the governing failure mode and is easy to miss if joint classification is done visually rather than by load path analysis
• Ignoring Q_f during early scantling design — chord loads from global analysis often change significantly between concept and FEED; always run joint checks with the final chord utilisation, not a placeholder of 0
• Using ISO 19902 resistance factors without matching ISO 19902 load factors — the partial factor system is calibrated as a pair; mixing ISO resistance factors with NORSOK load factors produces unconservative results
• Pile skin friction in sand using outdated API RP 2A WSD β-values — ISO 19902 Annex A uses unit friction limits from CPT-based methods that differ from the older API WSD approach; applying API WSD in an ISO 19902 regime can be non-conservative for dense sand layers
• Assessment triggers ignored during brownfield modifications — a topsides weight increase of >5% over original design basis requires formal structural assessment per ISO 19902 §21.3; this is commonly overlooked in brownfield hook-up scopes
• RSR calculated for extreme wave alone — ISO 19902 §21 requires the RSR check to include the combined effect of wave + current + wind; wave-only pushover can overstate RSR by 10–20%
• Flooded member detection omitted from inspection programme — a flooded chord or brace loses its design buoyancy contribution and may have internal corrosion; FMD is a mandatory inspection item per ISO 19902 §20 yet is frequently dropped to reduce dive time costs
• Hot-spot stress extrapolation at wrong reference points for non-tubular joints — ISO 19902 defines extrapolation distances specifically for tubular joints; applying the same 0.4t/1.4t distances to gusset plate or stiffened panel welds is incorrect and must use plate S-N class methods instead