Computational fluid dynamics driven surrogate model to predict hydrodynamic and acoustic properties of propeller boss cap fins

A novel method is proposed to predict the hydrodynamic and acoustic properties of propeller boss cap fins (PBCF) based on computational fluid dynamics (CFD) and surrogate models. The propulsive performance and radiated noise of the B-series propeller equipped with PBCF under open-water and hull-propeller coupling conditions are predicted and analyzed. The transient calculations are performed to generate accurate sample data for the surrogate model based on the Unsteady Reynolds Averaged Navier-Stokes and Ffowcs Williams- Hawkings equations. The response surface, polynomial, and Kriging models are used to learn sample data and output predictions. The relationship between inputs and outputs can be established from local data, which enables to obtain arbitrary response results in the global range. The hydrodynamic performance and radiated noise are compared for configurations with and without PBCF. PBCF improves the open-water efficiency by more than 1.5% at medium to high advance velocities. For the self-propulsion efficiency, over 5% improvement is achieved under ideal working conditions. In addition, PBCF has a directional effect on the radial and axial radiated noise, with better noise reduction in the axial direction. The difference between the axial and radial spectrums is significant, especially near the first blade passing frequency.