Transient Stability Assessment of Power Systems Based on Semi-Supervised Active Learning

Machine learning method has a broad application prospect in the power system transient stability assessment field. Still, there is a problem that it needs to use a large number of labeled samples to train the model, and time domain simulation takes a long time. Combining the semi-supervised learning method with the active learning method, a power system transient stability assessment method based on semi-supervised active learning is proposed to solve the problem. A clustering algorithm is used to select the initial samples, ensuring the initial model’s performance. Semi-supervised learning makes full use of the unlabeled sample pool and accelerate overall learning process. The simulation results on IEEE 10-machine, 39-bus system show that this method requires fewer labeled samples than traditional active learning methods, further improving assessment efficiency.