Otimização estrutural de pás de turbinas eólicas de material compósito via funções de base radial

Abstract

Wind turbines are machines capable of transforming wind kinetic energy into electricity. The turbine blades are responsible for converting aerodynamic forces into torque for the generator. In most cases, blades are made of composite materials and their stiffness and behavior affect the turbine performance, on the other hand, the material and layup orientation influence the blade stiffness. By taking advantage of the bendtwist coupling that is offered by the layup sequence of composite materials, it is possible to produce a passive control of the pitch angle in small size wind turbines. The passive control aims to improve the blade performance for higher wind speeds than the nominal design speed. Within this context, this research aims to obtain the best layup sequence of a wind turbine blade made of laminated composite material using metamodeling based optimization. For this, the finite element method is used to model the mechanical behavior of the wind turbine blade. The model is parameterized in order to simulate several layup configurations and to evaluate the behavior of the blade undergoing different wind speeds. From the simulation responses of the finite element model, a surrogate model is generated, using radial basis functions. Ply orientations are adopted as input variables, and the power coefficient of the wind turbine as output. In possession of the metamodel, an optimization process is carried out to find the layup sequence that offers the highest power coefficient. The optimized layup sequence is compared with other layup configurations, as well as the surrogate model responses with the finite element model itself. The results demonstrate that the optimization methodology using metamodels is efficient, moreover, the passive control of the pitch angle of wind turbine blades can be applied and generates a better performance for the turbine, using for that the bend-twist coupling of composite materials.