基于语义理解和AI的电力设备信息检索方法  

Power equipment information retrieval method based on semantic understanding and AI

作　　者：佘俊 罗勇 余少锋 廖崇阳 SHE Jun;LUO Yong;YU Shaofeng;LIAO Chongyang(Information and Communication Branch,Power Generation Company of China Southern Grid,Guangzhou 511442,China;Western Maintenance and Test Branch,Power Generation Company of China Southern Grid,Xingyi 562400,China)

机构地区：[1]南方电网公司调峰调频发电有限公司信息通信分公司,广东广州511442 [2]南方电网公司调峰调频发电有限公司西部检修试验分公司,贵州兴义562400

出　　处：《电子设计工程》2022年第22期89-92,98,共5页Electronic Design Engineering

基　　金：南方电网公司调峰调频发电有限公司科技项目(STKJXM20190134)。

摘　　要：传统的电力设备信息检索方法易受到电力系统中非结构化信息的影响,导致其存在信息查全率和查准率较低的缺陷。对此,提出了基于语义理解和AI的电力设备信息检索方法。使用加权法计算电力设备信息特征权重,以此为依据提取特征信息。根据语句类型和词语描述对象间的约束函数,设计基于语义理解的电力设备信息处理流程,通过扩展后的查询子串查询非结构和结构信息。构建基于语义理解的AI检索模型,从语义量化机制角度完成电力设备信息检索。实验表明,该方法信息查全率高于80%,查准率接近100%,证明其实现了设计预期。The traditional power equipment information retrieval method is easy to be affected by the unstructured information of the power system,which leads to the defects of low information recall and precision.A method of power equipment information retrieval based on semantic understanding and AI is proposed.The weighted method is used to calculate the feature weight of power equipment information,and the feature information is extracted.The constraint function between objects is described according to the statement type and words,and the information processing flow of power equipment is designed based on semantic understanding,and the unstructured and structural information is inquired by the extended query substring.The AI retrieval model based on semantic understanding is constructed to complete the information retrieval of power equipment from the perspective of semantic quantization mechanism.The experimental results show that the information recall rate of this method is higher than 80%,and the accuracy rate is close to 100%,which proves that this method achieves the design expectation.

关 键 词：语义理解 AI智能 电力设备信息 信息检索 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]