PSI - Issue 77

João Nuno Silva et al. / Procedia Structural Integrity 77 (2026) 657–664 Joa˜o Nuno Silva et al. / Structural Integrity Procedia 00 (2026) 000–000

660

4

2.4. EN 17149-3

The EN 17149 European standard establishes harmonized procedures for the strength assessment of railway ve hicle components. It is structured in three parts: Part 1 (General), Part 2 (Static Strength Assessment), and Part 3 (Fatigue Strength Assessment Based on Cumulative Damage). Part 3 provides a rail-specific methodology applicable to welded and unwelded components made of steel or aluminium, expressed in terms of stress-life (S–N) resistance and cumulative damage. For parent materials, the reference condition adopts a stress ratio of R = − 1 with a 97.5% survival probability at 10 6 cycles. Fatigue resistance is derived from the normal and shear stress components calculated from material properties (e.g., ultimate tensile strength) and adjusted by modifying factors. These include an anisotropy factor to reflect direction-dependent behaviour in rolled or extruded products, a surface-roughness factor accounting for the interaction between topography and grade, and material-type fatigue factors. Where components are cast, a casting quality factor is applied and calibrated with the inspection method. For welded joints, the reference stress ratio is R = 0 . 5 and fatigue classes are defined at 2 × 10 6 cycles. The resistance is modified by factors representing manufacturing quality, geometrical configuration, post-weld treatment, and inspection regime. Thickness e ff ects are introduced via a penalty relative to standard references (25 mm for steel, 10 mm for aluminium). The influence of tensile residual stresses is handled with a residual-stress factor that distinguishes stress-relieved and as-welded conditions; post-weld improvements such as TIG dressing or toe grinding are credited through validated enhancement factors. The weld quality level is linked to performance classes defined in EN 15085-3 standard, and the inspection class adjusts resistance according to inspection intensity. The standard defines S-N curves for fatigue class and stress type. Damage may be quantified either with classic or modified Miner’s rule. The design procedure incorporates partial safety factors to address uncertainty: γ L acts on the load spectra according to data confidence, while the resistance factor , γ M , is decomposed into γ M , S (consequence of failure), γ M , I (inspection regime), and γ M , V (validation by testing or reference standards). The assessment begins with the evaluation of the single component utilization ( U c ) derived from the Equation 9. These component utilisation factors, corresponding to the direct and shear stress components, are combined to form the multiaxial utilisation factor( U f ). Fatigue resistance is verified when the condition U f < 1 is satisfied. For parent material, the general combination rule is defined in Equation 10. However, if the stress state is proportional to direct components of identical sign, or if a critical-plane procedure is applied throughout, the simplified form is applied using the Equation 11. For welded joints, the orientation is fixed by the weld geometry. In this case, U f is obtained from the maximum normal stress components and the shear stress component, as expressed in Equation 12.

∆ σ eq , N D ∆ σ D

(9)

U c = γ M

U f , Multi = U c ,σ x 2 U f , Multi = U c ,σ x 2

+ U c ,σ y 2 + U c ,σ y 2

+ ( U c ,σ x · U c ,σ y ) + U c ,τ xy 2 − ( U c ,σ x · U c ,σ y ) + U c ,τ xy 2

(10)

(11)

U f , Multi = max( U c ,σ ⊥

) 2 + U

∥

c ,τ

2

; U c ,σ

(12)

2.5. EN 1993-1-9

The EN 1993-1-9 standard (Eurocode 3, Part 1-9) provides a unified framework for the fatigue design of steel structures based on stress-life (S–N) assessment and linear damage accumulation. The standard supports three stress approaches for welded details: nominal, structural (hot-spot), and e ff ective notch stress, and supplies reference detail categories (FAT classes) defined at 2 × 10 6 cycles with 95% survival probability. Variable-amplitude load histories are