Chaotic self-governing particle swarm optimization for marine propeller design

Due to many antithetical design parameters and complex fluctuating underwater conditions, marine propeller design has been one of the researchers' challenging problems. Recently, meta-heuristic algorithms have become one of the highly efficient solutions for solving such complex engineering problems. However, due to the meta-heuristic algorithm's stochastic nature, they are unreliable for industrial applications such as marine propeller design. Therefore, for the sake of having a robust meta-heuristic optimizer, in this paper, the conventional particle swarm optimization (PSO) algorithm is improved by modified chaotic self-governing groups of particles (MGPSO). To approve the efficiency of the designed algorithm, this paper first investigates MGPSO's performance on six challenging benchmark functions. Then the MGPSO is used to design the marine propellers optimally. To this aim, two targets, viz. maximize the propeller efficiency and minimize its cavitation, which conflict with each other, are considered as the fitness function. In this regard, the propeller's chord length and thickness are considered two main design parameters. The adverse effects of uncertainties in design parameters and operating conditions on efficiency and cavitation also are considered. In this regard, MGPSO is evaluated against the recently proposed benchmark algorithms such as ALO and BBO. First, the results indicated that MGPSO could find an exact true Pareto optimal front with a uniformly distributed approximation. The results also show that the propeller with 5 or 6 blades with rotation speeds between 180 and 190 RPM will have the best performance in the trade-off between efficiency and cavitation.