Physics Informed Neural Network based Online Impedance Identification of Voltage Source Converters

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review


The wide integration of voltage source converters (VSCs) in power grids as the interface of renewables causes the converter-grid interaction stability challenge. The black-box impedance of VSCs identified at the converter terminal is the key to facilitate the study of converter-grid interaction stability. However, since the limited impedance data amount in online measurement, the existing impedance identification methods cannot accurately capture characteristics of the impedance model in various operating scenarios with the changing profiles of renewables and loads. In this paper, a physics informed neural network based impedance identification is proposed to fill this research gap. The physics knowledge of the VSC is used to compress the artificial neural network, which can reduce the calculation burden of online impedance identification. Meanwhile, the two-steps impedance identification is developed with the inspiration of the transfer learning theory to further increase the online impedance identification efficiency. This method can significantly reduce the required data amount used in online impedance identification for the online stability analysis with the changing operating points. The case studies confirm the effectiveness of the proposed method.

TidsskriftI E E E Transactions on Industrial Electronics
Sider (fra-til)1
Antal sider1
StatusUdgivet - 2022


Dyk ned i forskningsemnerne om 'Physics Informed Neural Network based Online Impedance Identification of Voltage Source Converters'. Sammen danner de et unikt fingeraftryk.