Intelligent Optimization of Fast Boat Hull Form for Resistance Reduction Using CFD and Surrogate-Assisted Algorithms

Abstract

High-speed fast boats operating at high Froude numbers experience rapidly increasing resistance due to coupled viscous and wave-making effects. This study proposes a surrogate-assisted hull-form optimization framework that combines Reynolds-averaged Navier-Stokes (RANS) CFD (SST k-ω) with Gaussian Process Regression (Kriging) to minimize the total resistance of a fast monohull while maintaining displacement and geometric feasibility. The hull geometry was parameterized using four variables: deadrise angle, prismatic coefficient, longitudinal centre of buoyancy (LCB), and bow entrance angle. A set of candidate designs was evaluated by CFD, a surrogate model was trained and validated (R² = 0.985, prediction error < 2%), and global optimization was carried out using Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). CFD verification of the best design shows a consistent resistance reduction across Fn = 0–0.75, with a maximum reduction of 14.4% at Fn = 0.65 compared to the baseline hull. The optimized hull exhibits reduced bow pressure peaks and delayed flow separation at the transom. The surrogate-assisted strategy reduces the number of CFD evaluations and lowers the overall computational effort by about 78% while preserving prediction accuracy. R^2=0.992.