Multi-objective shape optimization of autonomous underwater vehicle by coupling CFD simulation with genetic algorithm

In this paper, a new method of multi-objective optimization of the shape parameters was proposed for the Autonomous Underwater Vehicle (AUV) with both towed and self-propelled, aiming to reduce drag and enhance motion stability during navigation. In this study, the hydrodynamics calculation of AUV was conducted by CFD method covering a wide range of Reynolds number between 105 and 107, and the optimization was performed by NSGA-II multi-objective optimization algorithm. Prior to deploying the optimization process, a sample library was created using the Latin hypercube sampling method to construct the Kriging approximation model. Thereafter, the model was used as the NSGA-II multi-objective optimization function, in which the range of design variables was considered to obtain the Pareto optimal solution. Furthermore, we also introduce and discuss the Pareto optimal points with special characteristics. In addition, the objective function values of these four optimal points were compared with the results of the initial model. The results showed that the drag at the four optimal points were reduced by 12.07%, 11.59%, 10.98% and 7.94%, respectively, and the overturning moments were reduced by 1.96%, 7.38%, 12.73% and 14.08%, respectively. The proposed multi-objective optimization method optimizes the shape of AUV effectively and provides valuable guidance for the design of AUV.