Transient stability evaluation method of electric power system based on graph attention network

Abstract

The invention discloses a transient stability evaluation method of an electric power system based on a graph attention network, and the method comprises the steps: obtaining operation data which comprises an operation data matrix X constructed by power grid real-time operation data measured by a PMU and an adjacent matrix A constructed according to a power grid operation topology; and inputting the operation data into a power system transient stability evaluation model for evaluation to obtain a stability result. The power grid topology is embedded into the evaluation model in an adjacent matrix mode, so that the difference of transient characteristics under different topologies is effectively considered, and meanwhile, the mobility of the model is improved; through a pooling layer in front of the classifier, the model is still applicable under the condition that the system scale is changed, the problem that the model needs to be redesigned and trained under the condition that the system scale is increased or decreased is avoided while the evaluation performance is ensured, and the model is not possessed by a previous deep learning transient stability evaluation model.

Description

technical field [0001] The invention relates to deep learning transient stability evaluation of power systems, in particular to a power system transient stability evaluation method based on a graph attention network. Background technique [0002] Fast and accurate power system transient stability assessment is of great significance. The power system transient stability assessment technology based on the deep learning model has attracted attention. Through the learning of a large number of samples, the mapping relationship between the observable operating characteristics of the system and the stability results can be established, so as to quickly perform stability assessment. The current common deep learning models, such as CNN, SAE, DBN, etc., cannot effectively consider the influence of the power grid topology when performing feature extraction. The feature extraction effect of high-dimensional operating data is not ideal enough. When the scale of the power grid changes, ...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to overcome the above-mentioned deficiencies in the prior art, and provide a power system transient stability evaluation method based on graph attention network, so as to solve the problem that it is difficult to effectively consider the power grid when extracting features of the power system transient stability evaluation model based on deep learning. The problem of topology change can better mine the dynamic correlation between adjacent nodes and embed it in the high-dimensional feature aggregation process, which can significantly improve the application performance of the transient stability evaluation model

Images