Koopman Operator-Based Prediction of Short-Term Voltage Stability

In this study, a novel method for predicting shortterm voltage stability(SVS) is developed. In the proposed method, the dominant Koopman operator is extracted from the initial datasets of measured voltage by using dynamic mode decomposition (DMD). By employing the dominant Koopman operator as the basis, the trends of short-term voltage are predicted. The SVS is evaluated by determining the Lyapunov exponent (LE) of the predicted voltage trends. Since the negative influence of oscillation on LE is eliminated, the proposed method is highly accurate with the ability to rapidly predict the SVS. Numerous numerical simulations demonstrate the effectiveness of the proposed method.