Search papers, labs, and topics across Lattice.
This paper introduces a machine learning-assisted adjoint-based shape optimization framework for ship hulls with complex propulsion systems, specifically Voith Schneider Propellers (VSP). A Conditional Variational Autoencoder (CVAE) is used to create a surrogate model of the VSP's time-averaged flow field, replacing computationally expensive time-resolved propeller simulations. Results demonstrate that the surrogate model significantly reduces computational cost while maintaining accuracy, and the optimized hull designs achieve over 8% reduction in resistance compared to designs optimized without considering the propulsion system.
Ignoring complex propulsion systems in ship hull optimization can lead to worse designs, but a CVAE-based surrogate model unlocks efficient adjoint optimization achieving an 8% resistance reduction.
Adjoint-based shape optimization of ship hulls is a powerful tool for addressing high-dimensional design problems in naval architecture, particularly in minimizing the ship resistance. However, its application to vessels that employ complex propulsion systems introduces significant challenges. They arise from the need for transient simulations extending over long periods of time with small time steps and from the reverse temporal propagation of the primal and adjoint solutions. These challenges place considerable demands on the required storage and computing power, which significantly hamper the use of adjoint methods in the industry. To address this issue, we propose a machine learning-assisted optimization framework that employs a Conditional Variational Autoencoder-based surrogate model of the propulsion system. The surrogate model replicates the time-averaged flow field induced by a Voith Schneider Propeller and replaces the geometrically and time-resolved propeller with a data-driven approximation. Primal flow verification examples demonstrate that the surrogate model achieves significant computational savings while maintaining the necessary accuracy of the resolved propeller. Optimization studies show that ignoring the propulsion system can yield designs that perform worse than the initial shape. In contrast, the proposed method produces shapes that achieve more than an 8\% reduction in resistance.
