Renewable Power Generation Maximization for UAVs with Adaptive Sliding Mode Control

The demand for efficient and Fuel consumption alternative renewable energy sources in the aviation sector has led to the exploration of advanced technologies such as Ram Air Turbines (RATs). RATs are critical in applications requiring power generation in emergency scenarios that may happen to un-manned aerial vehicles (UAVs). However, the primary challenge lies in optimizing the performance of the RATs, such as power output and operational stability, under varying and unpredictable wind conditions. Traditional control methods often fail to adapt to these dynamic environments. This paper underscores the necessity of utilizing a recurrent neural network (RNN) to detect uncertain wind turbine dynamics; a sliding mode control rule is created in the suggested controller to monitor the optimal turbine rotation speed. Next, an online update technique is developed to provide real-time weight updates for the RNN, enabling control over maximum power extraction. According to simulation results, the suggested controller outperforms a traditional control approach in tracking the optimal turbine rotation speed and extracting the most wind power from RATs by 13 times, even in severe nonlinearities and system uncertainties.