Research on grid multi-source model fusion system based on SG-CIM semantic distance
Xin He, Shijie Gao, Tingzi Pan, Wei Zhang, Fan Zhang · 2025
In order to solve the problems of data redundancy, naming conflict and structural inconsistency among multi-source system models such as SCADA, EMS, GIS, etc. in the electric power industry, this paper proposes a multisource model fusion method based on semantic distance based on the SG-CIM standard. The method defines the semantic distance between entities and attributes by constructing the SGCIM semantic graph model, and realizes inter-model similarity quantification by using the shortest path calculation of the graph structure. On this basis, combining naming similarity and structural similarity, the comprehensive similarity function is constructed to realize entity matching and complete attribute merging and relationship reconstruction. In this paper, a multisource fusion prototype system is designed and implemented, which has the functions of model parsing, semantic computation, fusion execution and visualization display, and supports inputs in various formats, such as CIM/XML, SCADA point list and EMS configuration model. The system is validated on the measured data sets of regional power grids, with a fusion accuracy of $96.45 \%$ and a processing speed of up to 200 entities/second, which significantly improves the automation and engineering usability of model fusion.