基于知识图谱的电力检修方案智能编制方法

Knowledge Graph-based Intelligent Programming Method of Electric Power Maintenance Scheme

为了解决传统电力检修方案编制不规范、电力检修效率低的问题,提出基于知识图谱的电力检修方案智能编制方法。采用改进传统知识图谱的创建方法,构建本体模型、实体模型和关系模型;利用关系将实体与实体、实体与属性相连接,组合成三元组存入图数据库内,构建电网拓扑结构的知识图谱;依据构建好的知识图谱,建立设备图元和电力检修数据库间的图库映射关系,生成电网检修基础数据;制定图形环境下的检修时间,采用启发式搜索法搜索设备映射库中的缺陷记录,并将其与电网检修知识库相关联,进行数据规范化及填补,完成电力检修方案智能编制。实验结果表明,该方法编制的数据报表更加工整规范,检修方案的设备整体搜索时间均在90ms以下,少于现有方法的220 ms以上,检索准确率在70%以上,高于其他方法的70%以下,召回率在50%以上,显著高于现有方法。该方法提高了检修效率,且合理性及可执行性好。

In order to solve the problem of nonstandard and low efficiency of traditional electric power maintenance plan,an intelligent programming method of electric power maintenance scheme based on knowledge graph is proposed.The ontology model,entity model and relationship model are constructed by improving the traditional knowledge graph method.The connections between entities and entities and between entities and attributes are established by means of relationship and are then combined to form a triplet stored into a graph database,thereby obtaining a knowledge graph of power grid topology structure.According to the constructed knowledge graph,the mapping relationship between equipment graphics and power maintenance database is established to generate the basic data of power grid maintenance.By determining maintenance time under graphical environment,defect records in the equipment mapping library are searched via the heuristic method.The equipment mapping library is associated with the power grid maintenance database to standardize and fill the data,and the intelligent programming of power maintenance scheme is realized.The experimental results show that the data report prepared by the proposed method is more standardized.The overall search time of equipment in the maintenance scheme(<90 ms)is less than ms msthat of existing methods(>220 ms).The retrieval accuracy(>70%)is higher than that of other methods(>70%).And the recall rate(>50%)is significantly higher than that of existing methods.The method improves the maintenance efficiency,and has good rationality and feasibility.