Design optimization of the podded propulsor considering pod housing drag and propulsion motor performance

The podded propulsor (POD) offers advantages such as high efficiency, low noise, and space-saving by integrating the propulsion motor and propeller. This paper presents an optimization methodology that simultaneously improves motor performance and pod housing drag through numerical analysis. The interaction between the motor and pod housing was analyzed, with parametric modeling conducted for both. The Latin hypercube sampling method generated samples of the motor and pod housing for simulation, and radial basis function neural networks were used to create approximate models for the motor performance and pod housing drag. The fitting R-2 of motor efficiency, motor power density, propeller hub drag, pod body drag, and pod strut drag of the approximate model reached 0.995, 0.978, 0.920, 0.972, and 0.999, respectively. Sensitivity analysis revealed that the stator outer diameter and fore taper angle are key factors influencing motor performance and drag, respectively. A genetic algorithm was used to optimize the POD with bi-objective and tri-objective focuses on minimizing drag, maximizing motor efficiency and power density. Pareto-optimal designs were validated through simulations. The results show that tri-objective optimization increased the overall efficiency by about 4.2% and enhanced motor power density by around 22%.