Due to the unbalanced development of various regions, in most regions, the distribution transformer of medium voltage line has not realized automatic monitoring, so the staff cannot monitor it in real-time. Therefore, an online monitoring system for the reactive power status of intelligent distribution transformers based on edge computing is designed. The system hardware structure configuration consisting of the master station layer, the edge layer and the acquisition layer is constructed, and the system software functional structure consisting of data processing, online monitoring, and system management is designed. On this basis, the continuous sampling data of the reactive power of the intelligent distribution transformer is collected, and based on edge computing, the reactive power operation characteristics of the intelligent distribution transformer are obtained and matched to judge the reactive state of the intelligent distribution transformer operation, so as to realize online monitoring of the reactive state of the intelligent distribution transformer. The experimental results show that the online monitoring accuracy and efficiency of the intelligent distribution transformer reactive state are high.

To reduce the error of single-phase ground fault and improve the security and stability of distribution network operation, an online monitoring method for single-phase ground fault in a medium voltage distribution network based on multi-measuring point information is proposed. Based on the analysis of single-phase grounding transient characteristics, the fault location is determined from the information of multiple measuring points according to the characteristics analysis results to realize the fault subsection location, identify the fault line section, and improve the location accuracy. Based on the fault location results, fault monitoring is realized through the “four-in-one” intelligent distribution terminal. The experimental results show that the accuracy of the monitoring results of this method is kept above 95%, and the monitoring error value of grounding fault is kept at about 1%, so the monitoring results have high reliability.

The current traditional low-voltage power line physical topology identification method mainly detects the line pulse signal to identify the line topology, leading to poor identification effects due to the lack of correlation analysis of node voltage. A data-driven low-voltage power line physical topology identification method is proposed in this regard. The correlation between the node timing voltages is analyzed by exploring the voltage law to determine the connection between low-voltage power lines. The node resistance values are also calculated to convert the physical topology identification problem into a linear equation problem. In the experiments, the proposed method is verified for identification accuracy. The analysis of the experimental results shows that the proposed method has a low root-mean-square error and better recognition accuracy when recognizing the physical topology of low-voltage power lines.