
Dario Vucinic
Surrogate model based optimization of wing sails for racing applications.
Rel. Mauro Bonfanti, Matteo Mastorakis. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Meccanica (Mechanical Engineering), 2025
![]() |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (3MB) |
Abstract: |
The search for an alternative propulsion system for marine transport has returned back to its roots, to a technology that has been known for thousands of years, wind propulsion. The need to reduce emissions and costs of transporting goods has revived the idea of applying wing sails to ships, starting with the experience of high-performance racing sailboats, that have been experimenting with this concept for 20 years already. The purpose of this work is to provide an innovative framework for wing sail design optimization, known in literature as Surrogate Based Design Optimization. The idea is to reduce the computational burden of high-fidelity fluid dynamic simulations, that would be too impractical to use coupled with an evolutionary algorithm for design space exploration and substituting them with a surrogate model trained on as few samples as possible. To account for the complex dynamics and behaviours that a wing sail would create on the whole boat, a static model is implemented in the workflow to guarantee the analysis of just feasible geometries. The geometry of the wing sails is described with 9 parameters in total, and they are initially sampled to create a Design of Experiments (DOE) table through a Latin Hypercube Sampling technique. These initial samples are then simulated with 2D unsteady CFD simulations to extract CL, CD and CM values for each design and used as training dataset for a Neural Network. The training of the Neural Network is carried out using the k-fold cross validation technique and the evaluation of MSE (Mean Squared Error) and MAE (Mean Absolute Error). A 6 Degreed of freedom static model is used to evaluate the equilibrium conditions and to find the combination of control parameters that maximises forward speed and at the same time guarantees equilibrium of the system. The optimal geometry of the wing sail is explored and found by a genetic algorithm, that includes in an inner optimization loop the search for control parameters that guarantee maximum performances for that particular geometry, and the overall best design will be the one that guarantees higher VMG (Velocity Made Good) on two buoys. |
---|---|
Relatori: | Mauro Bonfanti, Matteo Mastorakis |
Anno accademico: | 2024/25 |
Tipo di pubblicazione: | Elettronica |
Numero di pagine: | 93 |
Soggetti: | |
Corso di laurea: | Corso di laurea magistrale in Ingegneria Meccanica (Mechanical Engineering) |
Classe di laurea: | Nuovo ordinamento > Laurea magistrale > LM-33 - INGEGNERIA MECCANICA |
Aziende collaboratrici: | Politecnico di Torino |
URI: | http://webthesis.biblio.polito.it/id/eprint/36718 |
![]() |
Modifica (riservato agli operatori) |