Abstract
The increase in energy demand, together with growing energy-supply challenges, is expected to reshape the relationship between energy production and consumption in the coming decades. This transition increases the importance of small-scale and distributed renewable energy systems. Therefore, the Savonius rotor offers important advantages, particularly because of its simple design and suitability for urban environments. Hence, in the present study, a novel blade design methodology where the blade geometry is generated directly from the analytical equation of an ellipse and is fully defined by three parameters as the blade start angle , the end angle , and the angular parameter φ, which governs the blade depth. Afterwards, OpenFOAM simulations were conducted to evaluate the effects of these parameters and generate a dataset consisting of 157 cases for the subsequent surrogate-model development and optimisation. The resulting dataset was used to train an ANN surrogate model. The trained ANN surrogate model was coupled with a GA to perform surrogate-assisted optimization, and the predicted optimum was subsequently validated using CFD simulations. Computational results showed well agreement with literature. The result showed that the rotor efficiency increased by up to 45.32% at the highest investigated Tip Speed Ratio (TSR), with an average improvement of 32.75% over six different operating conditions for the same rotor diameter. In addition, the optimized configuration achieved a 22.32% improvement in coefficient over the reference configuration. In addition, the kinematic analysis demonstrated that increasing the blade-tip angle toward reduces the TSR dependence of the blade-normal component of the relative velocity.

This publication has 49 references indexed in Scilit: