Differential Flatness Using the Predictive Neural Network Control Law for Hybrid Power System

In this paper, a control design for a renewable energy hybrid power system that is fed by a photovoltaic (PV), Wind turbine (WT) and fuel cell (FC) sources with a battery (Batt) storage device is presented. The energy generated is managed through a nonlinear approach based on the differential flatness property. The control technique used in this work permits the entire description of the state's trajectories, and so to improve the dynamic response, stability and robustness of the proposed hybrid system by decreasing the static error in the output regulated voltage. The control law of this approach is improved using the predictive neural network (PNN) to ensure a better tracking for the reference trajectory signals. The obtained results show that the proposed flatness-PNN is able to manage well the power flow in a hybrid system with multi-renewable sources, providing more stability by decreasing the perturbation in the controlled DC bus voltage.