Urban underground comprehensive pipe galleries are gradually replacing traditional underground pipe galleries due to their advantages of strong safety, easy maintenance and monitoring. In order to monitor the power system facilities in the underground pipe gallery, and collect and process the collected power system data effectively, this paper designs and builds the overall hierarchical structure of the power detection system and the sub function modules of the power system monitoring and control system, and improves the corresponding data sensor layout scheme at each acquisition node. Then, combined with edge computing technology and cloud processing technology, it strengthens the processing and analysis level of the collected power data. Finally, based on edge cloud collaboration, a real-time monitoring model for the power system in the pipe gallery is developed. Convolutional neural networks are used to extract targets from power data. In order to enhance the ability to extract the main targets, attention mechanisms are integrated into the neural network. The model is trained using labeled multi time series data such as environmental temperature and humidity, protective layer grounding current, and fused data as datasets to infer the current state of the power system, achieve real-time monitoring of the power system status, and evaluate its operational status.
本站僅提供期刊文獻檢索。
