A data‐driven approach for new perspectives on the development of small‐scale H‐Darrieus turbines

• Ano de defesa: 2023
• Autor(a) principal: Santos, Gabriel Bertacco dos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Vertical‐axis wind turbine; Computational fluid dynamics; Sensitivity analysis; Aerodynamic optimization; Generative adversarial network; Turbinas de eixo vertical; Dinâmica dos fluidos computacional; Análise de sensibilidade; Otimização aerodinâmica; Rede neural generativa
• Link de acesso: https://hdl.handle.net/11449/250827

Wind energy has emerged as an attractive alternative to the current fossil fuel‐based energy mix. In this context, small‐scale H‐Darrieus vertical‐axis wind turbine (VAWT) combine interesting characteristics for harvesting wind energy in urban‐like conditions. Still, H‐Darrieus turbines are reported to experience relatively low aerodynamic efficiency, especially when compared with horizontal‐axis wind turbines (HAWTs) of equal scale. Even though several devices have been proposed to increase the aerodynamic performance H‐Darrieus turbines, the literature seems to overlook the potential of specifically designed airfoil shapes and blade geometries. In part, this is a direct consequence of the cost of optimization studies for H‐Darrieus turbines. So far, available alternatives rely on limiting either the design exploration or the search algorithm capabilities, which is surely a suboptimal approach for such a complex problem. To overcome the limitations of traditional approaches, we propose here a data‐driven analysis framework that mostly gravitates toward reducing the total computational cost for the optimization of H‐Darrieus turbines. For that, we use computational fluid dynamics (CFD) simulations along with sensitivity analysis, metamodeling, and optimization strategies. Additionally, airfoil shapes are parameterized using a deep generative adversarial network (GAN). The results show that the proposed analysis framework considerably reduces the number of model evaluations required for a complete analysis and optimization. The airfoil parameterization allows expanding the bounds of the latent design space to easily explore novel airfoil designs. The optimized geometries can increase the aerodynamic performance of the turbine by up to 20 % when compared with the NACA 0015 and the NACA 0021—two common airfoil shapes used in H‐Darrieus turbines. Interestingly, the optimized geometries were found outside the original bounds of the design space, further confirming that the search for novel airfoil designs may open the way for better aerodynamic performance of small‐scale H‐Darrieus turbines. To this end, data‐driven strategies may be an interesting approach to indicate new perspective for future development.