CFD 기반 자항 해석의 격자 민감도 정량 평가

Abstract

This study performs resistance and self-propulsion simulations for four ship types using computational fluid dynamics(CFD) and evaluates the impact of grid resolution on prediction accuracy and numerical convergence. A total of 56 cases were analyzed by applying coarse, medium, and fine mesh systems. Convergence behavior was assessed based on the Grid Convergence Index(GCI) methodology. Among the 56 cases, 42 were included in the GCI analysis, showing order of convergence(p) greater than 1.5 and grid uncertainty(UG) below 3% in most conditions. In particular, when over 10 million cells were used, simulation results showed consistent trends across different speeds within the same hull form, indicating robust convergence and reliability. Although a few cases were excluded due to irregular convergence behavior, their impact on the overall tendency was negligible. The findings provide practical guidelines for grid arrangement and uncertainty quantification in CFD-based ship performance predictions.