A Short-term Voltage Stability Evaluation Method Combining Data-driven and Mechanism Criterion

In order to accurately and quickly assess the short-term voltage stability (STVS) of a power system, this paper utilizes a data-driven method to enhance the power current criterion (PCC) to form a real-time assessment method for STVS. First, the fixed voltage threshold adopted by PCC is taken as the parameter to be enhanced, and the labeling method of dynamic threshold and the data set generation scheme are proposed. Then, Gate recurrent unit (GRU) network is utilized to establish the mapping relationship between the voltage trajectory and the dynamic threshold, and a penalty loss function is designed based on the classification principle of the criterion to enhance the usability of the prediction threshold. Finally, a set of time-adaptive evaluation scheme for online applications is proposed. The scheme first utilizes PCC to monitor the system, and then combines the dynamic threshold provided by GRU to complete the evaluation. Compared with traditional methods, the proposed method has good accuracy, timeliness and interpretability. In particular, the method is directly oriented to the measurement data, and its monitoring scope can adapt to the observability of the power grid. Test results in CSEE-VC(88-2023) and actual power grid validate the effectiveness of the proposed method.