Optimal design of Savonius wind turbine blade based on support vector regression surrogate model and modified flower pollination algorithm

Abstract

Blade shape has a significant effect on the wind-capturing ability of the Savonius wind turbine. This paper proposes an intelligent optimization method for optimizing the shape of the blade of the Savonius wind turbine. This method firstly uses the cubic Bezier curve with four design parameters (x1, y1, x2, y2) to characterize the complex blade shape, and then uses the Latin hypercube sampling method to sample some design schemes throughout the design space, and uses computational fluid dynamics simulation to evaluate the response value, i.e. the moment coefficient, of each scheme, and then uses the support vector regression surrogate model to describe the relationship between the design parameters and their response values, finally uses modified flower pollination algorithm to solve the surrogate model to obtain the optimal blade shape. After comparing and analyzing the optimized blade and the classical semicircular blade, it is found that compared with the classical semicircular blade, the optimized blade has a better wind-capturing ability. When the wind speed is 7 m/s and the tip speed ratio is 1, its average power coefficient Cp is significantly increased from 0.260027 to 0.277902 (about 6.87% higher). In addition, the aerodynamic performance of the optimized blade is also better than the classical semicircular blade at other tip speed ratios (TSRs = 0.6–1.2). It is shown that the wind turbine with the optimized blade has great potential in a practical application environment.