Abstract:In order to improve the accuracy of transient stability predictions based on convolutional neural networks (CNN) and demonstrate multi-angle auxiliary decision-making information, such as transient stability classification and margin, etc., this paper proposes a multi-task model for transient stability prediction using multi-level feature fusion based spatial pyramid pooling convolutional network. Firstly, the short-time disturbed trajectories of each generator can be obtained by phasor measurement units (PMUs), and a three-dimension information matrix may be constructed using these trajectories. Furthermore, the space pyramid pooling layer and multi-level pooling layer could be employed for extracting multi-scale and multi-level feature based on the CNN, which can obtain more comprehensive and detailed information of input data. Finally, the multi-task learning model with classification and regression will be built by hard parameter sharing. It may achieve transient stability classi?cation, critical generators identification and stability margin prediction. The case studies have been carried on an IEEE 145-bus system and an IEEE 39-bus system with the help of Matlab R2020a and PST 3.0 software. The results in comparison with those results from the conventional methods show that the proposed methodology is valid, and it shows the applicability when the information of PUMs is incomplete or contains noise.