A dynamic iterative method for hydrodynamic coefficients of flexible netting and its numerical validation

With the progressive expansion of marine fisheries equipment toward offshore and deep-sea environments, accurate characterization of the hydrodynamic behavior of netting structures has become a critical issue for ensuring structural safety and service reliability. Therefore, it is necessary to develop predictive methods for hydrodynamic coefficients applicable to different types of flexible netting structures. In this study, the effects of key parameters on the hydrodynamic response of a typical knotless cruciform mesh structure under two-way fluid–structure interaction (FSI) are investigated using computational fluid dynamics (CFD) and artificial neural networks. Based on the CFD results, a predictive model for the hydrodynamic coefficients of the cruciform mesh is established. Furthermore, an iterative computational strategy with dynamically updated hydrodynamic coefficients is introduced and applied to a single-layer flexible netting model to evaluate its overall hydrodynamic response. The results indicate that the Reynolds number and angle of attack are the dominant factors governing the hydrodynamic coefficients of the mesh. Compared with conventional approaches employing constant coefficients, the proposed neural network–based model improves prediction accuracy by approximately 9%, with more pronounced advantages under high-flow-velocity conditions. The proposed methodology provides an effective reference for refined hydrodynamic analysis and engineering design of large-scale flexible netting structures. • A BP neural network surrogate model was developed to predict drag and lift coefficients. • An iterative correction strategy dynamically updates net-twine hydrodynamic coefficients. • Net drag prediction is improved at medium-to-high inflow velocities.