Abstract: To address the issues of strong concealment and lack of timely warning for electrical fires, an intelligent early monitoring system for electrical fires integrating digital twin technology has been designed. The system is built upon a "cloud-edge-end" integrated architecture, and deeply incorporates the five-dimensional digital twin model theory. It integrates key technologies such as multi-dimensional sensing, artificial intelligence, and big data analytics to achieve full-process intelligence from data acquisition and edge analysis to cloud fusion. Front-end devices collect multi-dimensional operational data from the electrical system. The edge computing module enables real-time processing and early warning responses. The cloud platform integrates the digital model of the power distribution system, twin data, an electrical insulation aging prediction model, and a risk assessment model, comprehensively displaying the operational status and potential risks of the electrical system. Fire risk distribution is intuitively presented through a "unified risk map", achieving precise characterization of the digital twin model for a three-level power distribution system. This approach enhances proactive perception and intelligent prevention and control capabilities for electrical safety, providing comprehensive safety assurance for electrical systems.

Key words: electrical fire, digital twin, early-warning system, five-dimensional model, fire prevention