Propeller Optimization Design Based on Dakota and Potential Flow Methods

An efficient propeller design framework facilitates automated propeller design, enhancing the propulsion efficiency of underwater vehicles and ships with lower time and labor costs, increasing working hours, and reducing operational costs. Based on the lifting-line theory, this paper automates the generation of geometric parameters for propeller blades. Using these parameters as optimization variables within the Dakota framework, with constraints like the thrust coefficient and efficiency as the objective function, the performance is predicted via multiple invocations of the potential flow method program potHydro. The optimization algorithm iteratively adjusts the variables until convergence is achieved, yielding the optimal design. Integrating actual design requirements, this paper completes the automated design of the propeller, optimizing through the framework to achieve up to a 16.9% efficiency increase in the self-designed propeller. To validate the applicability of the optimization frame work, the constraints were modified to optimize the P4119 prototype propeller, resulting in a maximum efficiency increase of 9.13% for the P4119 propeller.