A genetic optimisation-aided framework for the computational design and progressive damage modelling of filament-wound composite pressure vessels

Accurate prediction of the structural response of filament-wound composite overwrapped pressure vessels (COPVs) remains challenging because conventional winding design tools provide limited trajectory flexibility and oversimplified representations of fibre architecture, laminate thickness, and crossover regions. This study presents an optimisation-assisted computational framework that integrates filament-winding trajectory generation, geometric reconstruction, and progressive damage finite element analysis within a unified workflow. The methodology combines multi-segment geodesic construction, non-geodesic trajectory optimisation, geometric thickness reconstruction based on explicit band overlap, and automated transfer of spatially varying fibre orientation and laminate thickness to three-dimensional finite element models. Unlike conventional approaches, the proposed framework accurately reproduces winding crossover topology and complex laminate architectures while accommodating arbitrary vessel geometries and manufacturable winding patterns. Validation against three benchmark pressure vessels demonstrated excellent agreement with published numerical and experimental results, reproducing laminate thickness distributions, stress fields, progressive damage evolution, and burst locations. Predicted burst pressures differed by less than 2.0% from experimental measurements. The analyses further revealed that burst behaviour is governed by the interaction between hoop–helical load sharing, ply termination, laminate thickness transitions, and interlaminar shear concentrations. The proposed framework establishes a direct digital link between filament-winding design and structural performance prediction, enabling more reliable optimisation and virtual certification of composite pressure vessels.