Non-proportional fatigue stress criteria in welded joints: a design code perspective

Abstract Welded joints often exhibit a significant reduction in fatigue life when subjected to non-proportional (NP) multiaxial loading conditions. Therefore, it is essential to evaluate how current design codes account for NP fatigue and whether their assessment criteria remain adequate. This study examines five NP fatigue life assessment methods from major design codes: IIW [1], DNV [2], ASME [3], FKM [4] and BSI [5] using a database of 22 experimental studies involving NP fatigue in welded steel joints. The evaluation is based on two structural stress assessment approaches derived from finite element analysis: the fine hotspot stress method, as recommended by IIW, DNV, FKM, and BSI, and the equivalent(-equilibrium) structural stress method (E(E)SS) method as proposed by ASME. These are also compared to two benchmark nominal stress-based criteria: maximum principal stress and von Mises stress. The results confirm that NP stress states lead to a notable reduction in fatigue life. While none of the evaluated design codes completely correct for this effect, the IIW and ASME method perform best overall in terms of prediction accuracy and consistency. Furthermore, the study proposes calibrated NP correction factors for each method based on the database. These factors yield near-perfect alignment between predicted and experimental fatigue life, offering a potential basis for improving current design practices.