Artificial Neural Networks for Predicting the Electrical Power of a New Configuration of Savonius Rotor
The electrical power of new configuration Savonius turbine is estimated using multilayer perceptron neural network (MLPNN) and radial basis function neural network (RBFNN) based on the experimental data which have been collected for 108 vertical axis wind turbine. In this work, the experiments were conducted at air velocities ranging from 3 to 12 m/s in front of a low-speed subsonic wind tunnel. Based on the experimental results, the new configuration of Savonius resulted in a noticeable improvement in the power compared to that of the classical Savonius turbine. This study reveals that the MLPNN has the potential of predicting the mechanical power of a Savonius wind turbine with minimal prediction error scores compared to the RBFNN model.
KeywordsElectrical power Savonius turbine MLPNN RBFNN
- 8.El-Ghazali, A.: The influence of turbine geometry on the performance of c-section vertical axis wind turbine. Master’s thesis, Near East University (2016)Google Scholar