Designing a Tool for Automating Submarine Hull CFD Simulations

Significant advancements in computational resources have enabled the use of sophisticated techniques such as CFD in conjunction with optimization methods during the early stages of project design. However, in naval engineering, optimization techniques are often limited to parametric models based on empirical methods. Furthermore, the high cost of tools, the complexity of modeling, and the numerous iterations required often make the use of CFD unfeasible. This paper presents a design methodology that integrates submarine hull modeling into an optimization method that uses CFD to determine the vessel's total drag. The hull is described by mathematical functions that characterizes its geometry, with the CFD model built using the Open-Source library OpenFOAM. The domain discretization, simulation input parameters, and mesh generation are iteratively adjusted until quality criteria and the y+ value are within established limits. The flow simulation takes place in a fully submerged domain, using the simpleFoam solver coupled with the k-omega SST turbulence model. A Python code is capable of automating the entire process of parametrization and simulation analysis, obtaining the hull drag with the following input parameters: total length, diameter, parallel mid-body length, and stern and bow curvature. Compared to the experimental results of the series 58 hull model 4165, the developed simulation presented errors of less than 6.9% for all Reynolds numbers analyzed, as well as consistent results with the imposed parametric variations. As an application example, a parametric study is presented, where the mesh quality analysis was conducted manually, with no significant problems found in the automated approach.