Design of experiments for efficient fatigue life prediction of ship structures

Fatigue damage in ship structures is a critical issue affecting operational safety and long-term reliability. Traditional fatigue analysis methods, including direct fatigue and spectral approaches, often incur high computational costs due to the large number of load scenarios. This paper presents a case study integrating Design of Experiments (DOE) techniques – specifically Taguchi's orthogonal array and the Plackett–Burman method – into fatigue analysis to reduce computational effort while maintaining accuracy. A representative ship model was analyzed using the Finite Element Method (FEM) under various operational and environmental conditions. Additionally, a simplified fatigue analysis was employed as an independent cross-check to confirm that the full-factorial baseline is not biased by scenario selection. A full-factorial design generated 108 FEM scenarios, reduced to 36 (Taguchi) and 22 (Plackett–Burman). Taguchi achieved 93% agreement with the full factorial analysis, while the Plackett–Burman achieved 84% agreement. These findings demonstrate that DOE techniques offer significant advantages as efficient, cost-effective and reliable tools for fatigue-life predictions in marine structures.