Abstract:The transient power angle stability and transient voltage stability in the power system are crucial for the safe operation of the system, while the increasing proportion of renewable energy access and the electrification of core equipment make traditional transient stability assessment methods difficult to apply. Therefore, this article proposes a multi task transient stability assessment model based on Transformer. This model utilizes Transformer"s self attention mechanism to break the serial computing structure of traditional recurrent neural networks, parallelly calculate the interrelationships between various time points in the transient process, quickly and efficiently extract temporal features, and uses gated multitasking learning methods to simultaneously evaluate transient voltage and transient power angle stability issues, improving model calculation accuracy and efficiency. This article verifies the effectiveness of the proposed method through simulation and testing in IEEE 10 machine 39 node systems and systems containing new energy.