Jianxun Zhao, Yongqiang Sun, Yuzeng Shu, Lijuan Gao, Yufei Liu, Yong Zhai
Gray clustering methods; Supervisory control and data acquisition;Multi-source heterogeneous; Smart grid; Data processing
Existing smart grid monitoring models exhibit inefficiencies in data processing and lack flexibility when handling massive data, failing to reflect real-time data conditions. To address these limitations, this study introduces gray clustering methods to optimize a smart grid monitoring model utilizing multi-source heterogeneous monitoring in- formation from Supervisory Control and Data Acquisition systems. The gray clustering approach reduces model complexity by simplify- ing intricate systems and minimizing subjective influences on evalu- ation metrics. Experimental results demonstrate that the proposed model achieves a Mean Absolute Percentage Error of 2.273%, out- performing the model based on a cloud model combined with a deep Q network and a graph neural network (4.281%), dynamic flow-by- flow load balancing technology (4.874%), and smart grid monitoring models based on Internet of Things (5.461%). When processing 2,500 message units, the proposed model reduces data storage gaps by 243 compared to the model based on a cloud model combined with a deep Q network and a graph neural network, and by 124 and 395 compared to dynamic load balancing and models based on the Inter- net of Things, respectively. The study provides effective data refer- ences for real-time power load monitoring and enhances operational flexibility in smart grid systems.
Important Links:
Go Back
