Construction and Application of Power Grid Fault Handling Knowledge Graph
Fan Li, Yingjie Tian, Yun Su, Naiwang Guo, Jun Hua Gao, Haiyan Pan · 2023
The fault handling emergency plan of power grid has great instructive significance for the quick and proper emergency disposal when the failures or accidents occur. To solve the problem of poor application effectiveness of the current fault handling emergency plan, a new method to construct the knowledge graph for the fault handling based on the BERT-BiLSTM-CRF model is proposed. In response to the problem of insufficient data available for deep learning model training and high data labeling costs in the power grid, the knowledge extraction module utilizes Bert pre-training method to construct the entity recognition model to improve the performance. The experimental results show that F1-score reaches 86.75% and the accuracy rate reaches 95.81%. Finally, the Neo4j graph database is adopted for highly visual management of the knowledge graph. When power grid faults occur, the fault handling knowledge graph can assist dispatchers to improve the capabilities of power grid's emergency disposal and the level of dispatch intelligence.