Research on Early Warning Methods for Electrical Safety Risks in AC and DC Systems

With the development of renewable energy and urbanization, the risk of electrical fires has increased. This paper constructs a three-layer fault disaster risk warning model through multi-sensor information fusion technology. Firstly, system signals are obtained through different types of electrical and physical sensors at the information layer. Next, fault features are extracted to determine the type of fault at the feature layer. Then Analytic Hierarchy Process (AHP) and Deep Neural Network (DNN) algorithms are applied to establish the electrical safety risk evaluation model at the decision layer. Finally, the risk evaluation model is applied to warn the electrical safety risks in AC/DC power supply systems separately. The model calculation results show that AHP is suitable for simple and linear problems while DNN is suitable for complex and nonlinear problems.