Schwickardi, MarikeMarikeSchwickardiHarder, MathisMathisHarderRenken, FinnFinnRenkenBraun, MoritzMoritzBraun2026-01-092026-01-092025-1111th International Conference on Fatigue Design, FatDes 2025https://hdl.handle.net/11420/60537This paper presents a novel approach for fatigue assessment of welded joints by employing extreme value statistics to predict the most critical weld toe geometry in welded structures. The methodology integrates advanced laser line sensor technology for precise weld geometry measurements, allowing the identification of geometrical features that are most likely to contribute to fatigue failure. By applying extreme value statistical techniques, the approach emphasizes the criticality of localized weld toe imperfections, providing a more accurate prediction of fatigue life compared to traditional methods. This technique offers improved reliability in fatigue life estimation for large welded structures, particularly in high-stress applications, and enhances the safety and durability of engineering designs.en2452-321620256571Elsevierhttps://creativecommons.org/licenses/by-nc-nd/4.0/Geometrical Feature ExtractionLaser Line SensorsNon-Destructive TestingWeld ImperfectionsWeld Quality AssessmentTechnology::620: Engineering::620.1: Engineering Mechanics and Materials Science::620.11: Engineering MaterialsConference Paperhttps://doi.org/10.15480/882.1637010.1016/j.prostr.2025.11.00810.15480/882.16370Conference Paper