| 引用本文: | 黄 简,杨 程,冯天波,等.面向风电机组运维的知识图谱构建研究与应用[J].电力系统保护与控制,2024,52(8):167-177.[点击复制] |
| HUANG Jian,YANG Cheng,FENG Tianbo,et al.Research and application of knowledge graph construction for wind turbine operation and maintenance[J].Power System Protection and Control,2024,52(8):167-177[点击复制] |
|
| 摘要: |
| 规模化风机高频运维信息愈发呈现数据高维、类间互联、规模攀升的特点，传统人工孤岛式故障检修运维模式难以适应人机料法环一体式发展进程。针对风机全环节、全要素的主动运维要求，提出知识规则、主从设备、规范条例一体融合的风电机组运维知识图谱构建方法。利用图论文本关键词提取算法(TextRank)完成对风电专业运维文本的实体识别与关系抽取，用以提高特征词的提取精度。采用Neo4j图数据库构建风电安全管理规程图谱及风电设备运维图谱，实现多元数据的互联与可视，进而实现风电运维信息的智能化查询。应用上述方法构建了629个实体、742条关系类型的风机知识图谱。数据查询试验表明：该方法的精确率及召回率等主要指标均在89%以上，较传统数据库方法平均提升了6.5%。该方法建立了运维要求可视表达和类间任务关联。运维大数据的有效查询，将有助于节省双碳战略风电运维力量，提高运维智能化水平。 |
| 关键词: 风电机组运维 知识图谱 TextRank算法 Neo4j图数据库 |
| DOI：10.19783/j.cnki.pspc.230950 |
| 投稿时间：2023-07-25修订日期：2023-10-08 |
| 基金项目:国家自然科学基金面上项目资助(52177185) |
|
| Research and application of knowledge graph construction for wind turbine operation and maintenance |
| HUANG Jian1,YANG Cheng1,FENG Tianbo2,SUN Ning3,LI Jiawen1,YU Hengwen1,CUI Haoyang1 |
| (1. College of Electronics and Information Engineering, Shanghai University of Electric Power, Shanghai 201306, China;
2. State Grid Shanghai Electric Power Company Information and Communication Company, Shanghai 200122, China;
3. Training Center of State Grid Shanghai Municipal Electric Power Company, Shanghai 200438, China) |
| Abstract: |
| The high-frequency operation and maintenance information of large-scale wind turbines increasingly presents the characteristics of high data dimension, interconnection between classes and scale increase. It is difficult to adapt he traditional artificial island-type fault repair and maintenance mode to the integrated development process of a man-machine, material, method and environment. For active operation and maintenance of all links and elements of wind turbines, a method to construct a knowledge graph of operation and maintenance of wind turbines is proposed. This integrates knowledge rules, master-slave equipment, specifications and regulations. TextRank is used to complete entity identification and relationship extraction of wind power professional operation and maintenance text to improve the extraction accuracy of feature words. The Neo4j database is used to construct wind power safety regulation and equipment diagrams to realize the interconnection and visualization of multivariate data, and then realize the intelligent query of wind power operation and maintenance information. In this paper, 629 entities and 742 relationship types of wind turbine knowledge graphs are constructed using the above method. The data query test shows that the accuracy rate and recall rate of the proposed method are all above 89%, which is 6.5% higher than the traditional database method. This method establishes the visual expression of operation and maintenance requirements and the task correlation between classes. Effective inquiry of massive equipment information is helpful to saving the maintenance force of the wind power in the two-carbon strategy and improve the intelligent level of the operation and maintenance. |
| Key words: wind turbine operation and maintenance knowledge graph TextRank algorithm Neo4j diagram database |
