Power system situation awareness method and device based on hybrid learning network large model

The invention relates to the technical field of electric power system control, and provides an electric power system situation awareness method and device based on a hybrid learning network large model, and the method comprises the steps: obtaining the transient energy and operation states of a plurality of generator nodes in a target electric power system in real time; determining the current transient energy margin and time margin of the target power system according to the transient energy, and fusing the transient energy and the operation state into situation awareness comprehensive variable data; inputting the transient energy margin, the time margin and the situation awareness comprehensive variable into a pre-trained extreme learning machine to predict the stability of the target power system; wherein the input layer weight and the hidden layer deviation of the extreme learning machine are corrected based on an LM algorithm during pre-training and are globally optimized based on a particle swarm algori