Neurocontrolled Prediction of Blade Position in Wind Generators

Controlling the position of wind turbine blades is a challenging task due to their highly nonlinear behavior and susceptibility to external disturbances from varying wind conditions and other weather phenomena. This challenge becomes even more complex when dealing with floating marine turbines, which are additionally influenced by currents and tides. To address this issue, it becomes essential to characterize the wind, including its speed and direction. By employing a predictive neural network model, it becomes possible to forecast the wind direction and effectively harness the wind’s potential. In this project, a machine learning-based network model is introduced. A dedicated real-time database was created to train the Artificial Neural Network (ANN). The practicality and effectiveness of this neural model are demonstrated. Subsequently, the predicted data from the neural network are utilized to control the pitch system of the aerodynamic model using actuators. The outcomes of this control approach were satisfactory, as a significant portion of the wind’s potential was effectively utilized, resulting in a considerable increase in energy generation. This approach proves applicable to large-scale wind turbines, providing a promising solution for harnessing wind energy efficiently.