Novel UUV Shape Optimization Design Based on Causal Relationships and Knowledge-Assisted Methods

To reduce the computational resources consumed during the shape optimization design process of novel unmanned underwater vehicles (UUVs) and to accelerate design efficiency, this study proposes a novel UUV shape optimization design method based on the causal relationships and knowledge assistance derived from conventional UUVs. A two-stage optimization mathematical model for conventional UUVs is established, aiming to minimize the drag coefficient, by using the conventional UUV drag coefficient causal relationship diagram and classifying design variables through correlation analysis and sensitivity analysis. On this basis, a correspondence between the design variables of conventional UUVs and novel UUVs is established using physical a priori knowledge, which maps to a two-stage optimization model for the novel UUV. The optimal drag coefficient solution from the conventional UUV’s design variable relationships is then transferred to guide the optimization of the novel UUV, enabling a rapid optimization design of its shape. Results indicate that the optimization design method based on causal relationships and knowledge assistance significantly accelerates the optimization iteration process of the novel UUV, reducing the drag coefficient by 42.9%.