Research on Power Network Fault Event Extraction Based on Hybrid Neural Network

In order to obtain event information of power grid faults in time, assist electric power workers to solve fault problems, make quick decisions, and reduce economic losses, a power grid fault event extraction model based on deep learning is proposed. The model is mainly composed of two Parts are composed of fault detection model and event role extraction model. Firstly, the power-related text is encoded based on the RoBERTa pre-training model. The fault detection model uses the BLSTM model to further extract the text features to obtain the specific fault category of the power public opinion text. Secondly, the event role extraction model uses the BLSTM-CRF model Extract the text features to get the event roles contained in the text. Finally, the power grid fault event information contained in the text data is obtained through the joint extraction of fault detection and event roles. Experimental tests show that the proposed model has better performance in grid fault event extraction results and accuracy.