The large-scale integration of doubly fed induction generators (DFIG) has led to increasingly highlighted stability issues in renewable energy connected system. The multi-operation impedance identification of DFIGs is crucial for analyzing the stability of the grid interconnection system. However

existing impedance identification methods rely on extensive data training and repeated parameter adjustment

resulting in a lack of physical meaning in the mapping relationship between the model and the actual impedance characteristics. Besides

these methods are insufficient to identify other DFIGs with different control structures and parameters. To address this issue

this paper proposes a unified multi-operation impedance modeling method of DFIG based on physical information in frequency domain characteristics. First

the paper derives the similarities and differences of the impedance characteristics of different DFIGs

and constructs the physical information of the unified impedance model. Then

the neural network model for impedance identification is established based on the prior physical information

which can enhance the ability to identify different DFIGs. Furthermore

the regularization constraints of DFIG are embed into the loss function to enhance the physical significance of the model. Finally

experimental verification confirms that the proposed method can accurately identify the impedance characteristics of DFIGs with different control structures and parameters

while significantly reducing the required amount of training data.