Accurate and fast assessment of power system transient stability is essential for maintaining the safe and stable operation of the power grid. A transient stability assessment method based on temporal convolutional network (TCN) and adaptive graph convolutional network (AGCN) is proposed to enhance the accuracy of assessing power system transient stability. This method models the transient stability assessment as a sample space mapping problem. The amplitude and phase angle of the bus voltage before

during and after the fault are taken as the input. The time convolution network is used to extract the time series characteristics of the transient data

and the adaptive graph convolution network is used to process the topology information between the grid nodes to mine its spatial structure characteristics to realize the fast and accurate judgment of the transient stability of the system. In addition

in model training

the focal loss (FL) function is used as the objective function to improve the model tendency problem caused by the inherent class imbalance of transient samples and the error-prone problem of difficult-to-classify samples in the stable boundary region. Finally

the proposed method's effectiveness is validated through simulation analysis using the IEEE 39 and IEEE 145 bus systems.