Data-Driven Adaptive Neural Network Additional Damping Controller for SSCI Suppression of DFIG-Based Wind Farms

In this article, a data-driven adaptive neural network additional damping controller (DDANN-ADC) is proposed to regulate rotor-side converters of a DFIG-based power system to suppress sub-synchronous control interaction (SSCI). Firstly, a back propagation (BP) intermediate variable observer is designed to construct a dynamic model of DFIG-based wind farms based on real-time input–output measurement data. Subsequently, a modified cost function is developed for a BP online controller to generate a target control law, thereby contributing additional damping to the DFIG-based power system. The proposed DDANN-ADC can effectively utilize limited data generated during the control process to achieve online system identification and precise control of the system. Then, the stability of DFIG-based power system under the proposed DDANN-ADC is demonstrated with the Lyapunov function. Finally, simulation results reveal that the proposed DDANN-ADC methodology outperforms the traditional method with better adaptability and robustness under different operational conditions.