Sliding Mode Control and RNN-Based Power Optimisation for Small Wind Turbines in EVs

The demand for additional power sources to further improve energy efficiency and sustainability in electric vehicles has grown, encouraging innovation in renewable energy solutions. Among the most promising control approaches in this regard is the inclusion of wind energy systems to further increase the power generation in electric vehicles under various operational conditions. However, some approaches may struggle to adapt to turbulent and unpredictable wind conditions. This paper proposes a recurrent neural network to model uncertain wind turbine dynamics, embedded with a sliding mode control strategy to maintain optimal rotational speed. An online updating mechanism provides real-time updates to the RNN weights for efficient control. Simulation results show that the proposed controller outperforms conventional schemes for superior capturing of turbine speed under nonlinear conditions with system uncertainties and achieves higher power extraction, up to 40 times more than typical wind turbines in electric vehicles.