Adjoint-based shape optimization of a ship hull using a Conditional Variational Autoencoder (CVAE) assisted propulsion surrogate model
This work optimizes ship hull shape using adjoint methods but faces computational limits with complex propulsion systems like Voith Schneider Propellers (VSP). The solution introduces a Conditional Variational Autoencoder (CVAE) to create a machine learning surrogate model that replicates the time-averaged flow field. This allows for significant computational savings while maintaining accuracy for optimizing vessel resistance.