| 基于知识图谱的电力设备故障诊断方法 |
 点此下载全文 |
| 引用本文：王宁.基于知识图谱的电力设备故障诊断方法[J].计算技术与自动化,2026,(1):132-138 |
| 摘要点击次数: 255 |
| 全文下载次数: 245 |
|
|
| 中文摘要:现有的中文实体识别方法大多准确率较低，难以从中文技术文献中提取知识来构建高质量的知识图谱。因此提出了基于知识图谱的电力设备故障诊断方法，通过BERT-BiLSTM-CRF模型，建立电力设备故障的知识图谱。首先通过BERT模型对动态词向量进行预训练，然后引入BiLSTM-CRF模型，综合考虑文本的全局和局部特征，考虑数据标签的隐藏序列规则，提取实体之间的语义关系。最后，提取的知识以三元组的形式存储在Neo4j数据库中，并以图的形式进行可视化。以电力设备故障诊断的中文技术文献为实验对象，实验结果表明，该模型比传统方法更准确地识别和提取中文实体，从而更容易构建全面、准确的电力设备故障中文知识图谱。 |
| 中文关键词:知识图谱；故障诊断；BERT-BiLSTM-CRF 实体识别 |
| |
| Power Equipment Fault Diagnosis Method Based on Knowledge Graph |
|
|
| Abstract:Most of the existing Chinese entity recognition methods have low accuracy, and it is difficult to extract knowledge from Chinese technical literature to construct high-quality knowledge graph. Therefore, this paper proposes a fault diagnosis method of power equipment based on knowledge graph, and establishes the fault knowledge graph of power equipment through BERT-BiLSTM-CRF model. Firstly, BERT model is used to pre-train dynamic word vectors, and then BiLSTM-CRF model is introduced to comprehensively consider the global and local features of text and the hidden sequence rules of data labels to extract the semantic relationships between entities. Finally, the extracted knowledge is stored in the Neo4j database in the form of triples and visualized in the form of graphs. The experimental results show that the model can identify and extract Chinese entities more accurately than the traditional methods, so that it is easier to construct a comprehensive and accurate Chinese knowledge graph of power equipment fault diagnosis. |
| keywords:knowledge graph fault diagnosis BERT-BiLSTM-CRF entity recognition |
| 查看全文 查看/发表评论 下载pdf阅读器 |
位访问者,今日访问