On the Limitations of Machine Learning (ML) Methodologies in Predicting the Wake Characteristics of Wind Turbines

Machine Learning (ML) algorithms have been more prevalent in recent years, and they are being used to tackle complicated issues across a broad range of fields. Wind energy is not an exception, as ML has recently applied wind turbine blade design, wake velocity turbulence intensity prediction, even farm optimization. The immense learning ability models enables them be trained predict regress complex relationship with high degree accuracy. However, data for testing often originate from the same rotor simulation training, just slight variations operating conditions. This research aims investigate generalizability ML-based prediction models, i.e., whether can correctly properties using different that was taken training. investigation's observation shows generalized model requires training multiple turbines wide In addition, advanced regularization, loss functions, methods focus on capturing physics (such Physics Informed Artificial Neural Networks (PINN) symbolic regression) utilized.