A Knowledge Extraction Method Based on Deep Learning Applied to Distribution Network Fault Handing Assistant Decision

Abstract

In order to effectively solve the problem of a large number of semi-structured and unstructured text information associations in the distribution network, the fault handling efficiency of the distribution network is effectively improved. Taking the text of distribution network fault handling plan as the research object, a deep learning model based on knowledge graph is established, and knowledge graph technology is used to assist staff in making fault handling decisions. First, combined with the characteristics of the power grid fault handling plan text, multiple deep learning models are used for knowledge extraction, forming a bottom-up knowledge graph data layer: In order to prevent word segmentation errors, the TextCNN model based on word vector is used to classify the text of the plan; the LR-CNN model is used for named entity recognition, thereby solving the conflict problem of candidate words; on the basis of the above, the BiGRU-Attention model is used to extract entity relation. The effectiveness of the above knowledge extraction method is verified by experiments. Finally, the knowledge graph of fault handling information of distribution network is constructed by using the Neo4j graph database, which is used for intelligent information retrieval and auxiliary fault diagnosis.