Neural Network Enabled Impedance Identification for VSC and Implementation in SimuNPS
Yongping Duan, Dayou Lu, Dengke Gao · 2025
The impedance identification work for VSC enables the stability analysis in designing and operation of modern power system. This paper proposes a funnel-shaped fully connected neural network method combined with transfer learning for impedance identification. Through training, a basic model is obtained, and through transfer learning, impedance identification for different controllers or equipment or working conditions can be supported. The proposed method is developed as a function into the commercial software SimuNPS for the convenient application in real life. Numerical experiments validate the accuracy of the method.