Stepwise concept selection and performance–effort refinement of marine composite stiffened panels under axial compression

Stiffened composite panels are widely used in marine lightweight structures, but their design is challenging because performance depends on stiffener type, geometry, laminate layout, fabric type, and interface properties. Relevant limit states require different modeling approaches: buckling-related stiffness degradation and ply failure can be assessed with layered shell models, whereas detailed interface debonding requires local modeling. A parametric global-local finite element framework, validated against published experiments for the three stiffener families, is used for a performance–effort selection and refinement procedure for top-hat-, blade-, and double-T-stiffened composite panels under axial compression with constant dimensions and equal mass. The contribution is its structured application to early-stage concept comparison and refinement rather than a general optimization procedure. Concept selection combines shell-based structural indicators with a normalized manufacturing-effort score. The selected baseline concept is refined through Latin-hypercube geometry variation, fabric comparison, and geometric fine-tuning. For the investigated case with fixed dimensions, equal mass, selected material data, and retained laminate concept, the four-top-hat configuration achieved the highest performance–effort efficiency. During refinement, the minimum indicator load increased from 218.88 kN to 281.94 kN, corresponding to a 28.81% increase. The results demonstrate improved performance within the design space investigated while limiting local cohesive analyses.