基于知识图谱的电力物资检测辅助优化方法

Auxiliary Optimization Method for Electric Power Material Detection Based on Knowledge Graph

电力物资检测业务累积数据的多源异构特性导致难以高效分析挖掘,传统物资检测手段过程繁琐,增加成本和工作量。针对电力物资中二次设备检测流程冗长、效率低的问题,选取命名实体识别与知识图谱技术结合的知识抽取、转化及应用方案,提出引入BERT预训练模型的改进方法抽取二次设备缺陷文本信息,结合知识图谱技术实现跨领域异构电力物资知识的表示与融合建模。最终搭建二次设备检测智能推荐系统,基于知识图谱与关联分析推理给出检测方案建议,实现了同业务需求下对电力二次设备检测流程的精简优化。

The multi-source,heterogeneous nature of the accumulated data from power material inspection operations makes efficient analysis and mining challenging.Traditional material inspection methods are cumbersome and time-consuming,leading to increased costs and workloads.This paper focuses on the lengthy and inefficient inspection process for secondary equipment in power materials.It proposes a knowledge extraction,transformation,and application scheme that integrates named entity recognition and knowledge graph technologies.It introduces an enhanced method using the pre-trained BERT model to extract textual information on defects in secondary equipment.Furthermore,it employs knowledge graph technology to represent and integrate heterogeneous power material knowledge across domains.Ultimately,an intelligent recommendation system for secondary equipment inspection is developed.Based on knowledge graph and association analysis reasoning,the system generates inspection plan recommendations,thereby streamlining and optimizing the inspection process for power secondary equipment under similar business requirements.