Abstract
Multiaxial fatigue data from 233 welded test specimens taken from eight
different studies have been evaluated based on three published
interaction equations for normal and shear stress. The interaction
equations were obtained from SFS 2378, Eurocode 3 and International
Institute of Welding (IIW) recommendations. Fatigue classes for normal
and shear stress were obtained directly from the design guidance
documents. Additionally, mean fatigue strengths were determined by
regression analysis of bending only and torsion only data for different
specimen types. In some cases, the S–N slopes assumed by the different
standards were not appropriate for the test data. Specimens that showed
significantly different cracking locations or cracking mode between
bending and torsion were not easily correlated by the interaction
equations. Interaction equations work best in cases where both the
normal stress and the shear stress tend to produce crack initiation and
growth in the same location and in the same direction. The use of a
damage summation of 0.5 for non‐proportional loading as recommended by
IIW was consistent with experimental observations for tube‐to‐plate
specimens. Other codes used a damage sum of unity.
Original language | English |
---|---|
Pages (from-to) | 991-1003 |
Journal | Fatigue & Fracture of Engineering Materials & Structures |
Volume | 27 |
Issue number | 11 |
DOIs | |
Publication status | Published - 2004 |
MoE publication type | A1 Journal article-refereed |
Keywords
- multiaxial fatigue
- fatigue