Research On Transient Stability Evaluation Of Energy Internet Based On Gate Graph Neural Network

With the development of internet technology,the internet of energy has penetrated into people's work and life.The power grid is an important part of the internet of energy,so the transient stability analysis of the power system has great significance to the safe and stable operation of the internet of energy.In this study,the Gated Graph Neural Network is used to evaluate the transient stability of the power system and infer the types of events that cause the power system to instability.First,through the Conditional Generative Adversarial Network(CGAN),unstable sample data is generated to balance the stable and unstable samples.The sample classification balance can make the GGNN model evaluation more accurate.Secondly,the balanced data graph is structured to form a data set,and the graph structured data set is used as the training test set to train the transient stability assessment model of the power system based on GGNN.Finally,push the test data to the trained GGNN model to judge the transient stability of the power system,and when the result is instability,output the events that caused the system to instability.Through experiments,it can be concluded that compared with other methods,the method proposed in this study has better accuracy and can determine the cause of system instability.Using Power Factory software to perform time-domain simulation calculations on the New England 39-bus system,stable and unstable samples under 5 different events are obtained.The obtained samples are used as a data set to perform offline training and online testing on the CGAN and the transient stability assessment model based on GGNN.Experimental results show that CGAN can generate data conforming to the distribution,solve the problem of data imbalance,and improve the performance of the transient stability assessment model.Using the data set balanced by the CGAN model to train the transient stability assessment model based on GGNN,and set different training and test set ratios.Experiments show that no matter which scale the model is under,the accuracy rate is higher than that the transient stability assessment model based on CNN,and the highest accuracy rate is 92.09%.When the evaluation result of the system is unstable,the GGNN model is used to determine the most likely event that caused the system.Experiments show that it is feasible to judge events that cause system instability,and the accuracy rate is 94.82%.