An Application Of Multidisciplinary Design Optimization To The Optimization Of The Hydrodynamic Performances Of Underwater Robots

Traditionally, the optimization of underwater robot’s hydrodynamic performances mainly adopts the solitary discipline optimization design, which ignores the coupling effect between different hydrodynamic performances disciplines. It would lead to low efficiency and time consumption in the design. Furthermore, the global optimization solution cannot be guaranteed after the optimization process. This paper applies a complex engineering system theory into the optimization of underwater robot’s hydrodynamic performances. Research works in this paper are summarized as follows.(1) This paper discusses the background and significance of the research project, and addresses the state-of-the-art of research progresses of underwater robot hydrodynamic performance design. Combined with the design processes of underwater robot, the limitation of the traditional design had been described. The feasibility and necessity of the application of Multidisciplinary Design Optimization (MDO) to the optimization of the hydrodynamic performances of underwater robots had been shown.(2) The Collaborative Optimization (CO), one of the MDO’s methods, is adopted. To remove the defects caused by the consistency constraints of CO, hybrid optimization strategy is proposed. The hybrid optimization strategy includes Particle Swarm Optimization (PSO) and Modified Method of Feasible Directions (MMFD). From the numerical simulation results, the modified CO based on hybrid optimization strategy not only improves the defects but also guarantees the global optimization solution.(3) The Computational Fluid Dynamics (CFD) is used in numerical test. The optimization results of hydrodynamics are compared with the trial data. The numerical test for forward moving and oblique towing had been conducted in the SUBOFF models, including SUBOFF Bare Hull, SUBOFF Bare Hull with Fairwater and SUBOFF Fully Appended. The results of the numerical tests are compared with the trial data. The results are quite coincident and the forecasting purpose can be achieved. Therefore, it can be stated that the computational domain, meshing and numerical methods used in this paper is reliable.(4) To deal with the contradiction between accuracy and efficiency, the approximate model based on Optimal Latin Hypercube design (Opt LHD) and Radial Basis Functions (RBF) is proposed in the modified CO framework. The approximation model method can not only simplify the system, but also meet the requirement of design accuracy in the design stage.(5) An automatic optimization of Isight integration platform is set up based on CO framework on sub-disciplines, i.e. the rapidity and the manoeuvrability. Softwares and author-edited-programming are integrated in the platform. SUBOFF is taken as a verification model. Through the platform, an automatic design and analysis of the hydrodynamic performances of underwater robots can be achieved. The optimization makes underwater robot into a more streamline pattern, also both the size and the edge of the fairwater are optimized. Optimization results also provided several feasible designs.