Prediction of the power ratio and torque in wind turbine Savonius rotors using artificial neural networks

-The power factor and torque of wind turbines are predicted using artificial neural networks (ANNs) based on experimental that are collected over seven prototype vertical Savonius rotors. Unlike horizontal-axis turbines, in verticalaxis turbines rotation speed is low and torque is high. Therefore, this device could be used for local production of electricity. In this research, the rotors having different features in the wind tunnel and the tests are repeated 4 to 6 times for reducing error. All experiments are done on six blades in different Reynolds number and wind speed varied from 8 to 14 m/s. Input quantities for the prediction in neural network are Reynolds number and the tip speed ratio (TSR). Rotor's power factor and torque were simulated in different Reynolds numbers and different angles of blade in proportion to blowing wind in a complete rotation. The simulated Results were compared with the corresponding experimental data shows that the simulation has the capability of providing reasonable predictions for the maximum power of rotors and maximizing the efficiency of Savonius wind turbines. According to results, increasing Reynolds number leads to increase of power ratio and torque. For all examined rotors, maximum and minimum amount of torque happens in angle about 60 and 120, respectively. Key-Words: -Vertical Savonius rotors, Neural networks, Wind tunnel, Power factor, Torque, Tip speed ratio