Repository logo

Browsing by Author "Pereira, Francisco Miguel Bastos"

Now showing 1 - 1 of 1
  • Computational Fluid Dynamics for Synthetic Data Generation
    Publication . Pereira, Francisco Miguel Bastos; Barata, José; Moura, Ricardo
    This dissertation explores the use of computational fluid dynamics (CFD) for generating synthetic data in the context of biomimetic propulsion. It contributes directly to the SABUVIS II project, which involves a swarm of autonomous underwater vehicles designed for military and scientific applications. The study focuses on designing, simulating and evaluating bio-inspired fin propulsion systems with the aim of producing synthetic datasets for future Digital Twin integration. The objectives of this work are to develop a parametrized fin profile that can be easily and quickly modified in terms of both shape and movement, whether its amplitude, frequency of oscillation or the type of movement, in order to create a model for simulations and data generation. This model should be capable of running di!erent simulations to generate a diverse set of results. The methodology followed three stages. First, a bibliographic review mapped the state of research in biomimetic propulsion, synthetic data, and digital twins. Second, a CFD workflow modelled fin-driven propulsion, including fin profiling, mesh independence, domain setup, solver configuration, and UDFs for sinusoidal and turtle-inspired motions. Third, parametric simulations assessed flow velocity, thrust, hydrodynamic forces, energy use, and propulsion e"ciency under controlled conditions. Results show both movement types generate biologically consistent thrust and flow. Sinusoidal motion produced smoother velocity fields and steadier thrust, while turtle-inspired motion yielded localized peaks suited for rapid manoeuvring. E"ciency depended on fin length, frequency, and amplitude, with a 50 cm fin at 2 Hz giving the best thrust-to-energy ratio. The datasets also support synthetic data repositories and future digital twin development of the SABUVIS propulsion subsystem.
    2025-09-22Master thesis Open access Show more