Situation Assessment Model Based on Attention Mechanism Data Fusion for the New-Type Power System

The network security situation assessment of the new-type power system is a technology that uses the current power system data to judge the overall network security state and predict the future security state. The challenge nowadays is how to capture the network security situation from massive amounts of data and give the situation value and prediction with low latency. In response to the problems of the original network security situation assessment methods, such as excessive manual intervention and reliance on expert experience, this paper proposes a data fusion situation assessment method based on the attention mechanism, and entrusts the situation assessment task to the deep neural network to solve. Firstly, the CNN model and the BiLSTM model are used to extract the local features and context features of different types of abstract data, then the feature fusion method based on the attention mechanism is used to obtain the fusion vector, and finally, the one-dimensional convolutional neural network is used to assess the network situation. And the Focal loss function is used in the training process to further handle the problem of data imbalance. This method can achieve an accuracy close to that of expert assessment with less manual intervention.