Shape optimization using an open-source framework: the bulbous bow case

Shape optimization is a very time consuming and expensive task, especially if experimental tests need to be performed. To overcome the challenges of geometry optimization, industry is increasingly relying on numerical simulations. These kind of problems typically involve the interaction of three main applications: a shape morpher, a multiphysics solver and an optimizer. We present a shape optimization framework entirely based on open-source tools, where an ad-hoc shape morpher routine is built by calling the high-level functions of the MiMMO library, the multiphysics simulations are performed using OpenFOAM and the optimization loop is controlled by DAKOTA. To demonstrate the usability and flexibility of the proposed framework, we test it in a practical case related to the naval industry, where we aim at optimizing the shape of a bulbous bow in order to minimize the hydrodynamic resistance. To tackle this problem, we first validate the solver and calibrate the numerical model for a reference geometry of which experimental results are available. After having found the ideal mesh and solver parameters, we set up the optimization loop. In particular we investigate the effect of protrusion and immersion of the bulb through Surrogate Based Optimization. Finally, we present other examples of optimization scenarios for the bulbous bow case (e.g. effect of sailing speed), we highlight the logic behind the choice made, we give guidelines on how to deal with some problematic issues encountered during the SBO (e.g. sampling, building techniques, testing and infilling, effect of noise, use of machine learning techniques to detect early anomalies) and we compare the SBO with high fidelity optimization approaches.