|
| A PINN-VSG frequency control method for microgrids with high penetration of renewable energy |
| DOI：10.19783/j.cnki.pspc.260175 |
| Key Words:grid-forming converter physics-informed neural network virtual synchronous generator microgrid with high renewable penetration frequency control |
| Author Name | Affiliation | | SHEN Guangyu | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | DENG Wei | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | XIE Zijian | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | CUI Chenggang | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | XIAO Hao | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | YANG Yanhong | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China | | PEI Wei | 1. Faculty of Artificial Intelligence, Shanghai University of Electric Power, Shanghai 200090, China
2. Institute of Electrical Engineering of Chinese Academy of Sciences, Beijing 100190, China |
|
| Hits: 21 |
| Download times: 11 |
| Abstract:Virtual synchronous generators can provide inertia and damping support for microgrids with high renewable energy penetration. However, traditional fixed-parameter control strategies have limited adaptability under complex operating conditions. To address this issue, a parameter adaptive control method based on a physics-informed neural network (PINN) is proposed. First, the interaction mechanism among multiple virtual synchronous generators is analyzed. Angular frequency, the rate of change of angular frequency, and electromagnetic power at consecutive time instants are selected to construct time-series inputs, thereby achieving the online adjustments of the moment of inertia, damping coefficients, and secondary frequency regulation coefficients. Subsequently, the rotor motion equation is embedded into network training, and a joint loss function comprising a data-loss term and a physics-loss term is constructed to enhance the physical consistency of the network outputs. Finally, simulations and hardware-in-the-loop experiments are conducted to validate the proposed method. The results show that the proposed method achieves fast and smooth frequency recovery under different load disturbances and effectively suppresses overshoots and oscillations, demonstrating good adaptability and generalization capability. |
| View Full Text View/Add Comment Download reader |
|
|
|