Data Driven Urban Power Grid Disaster Prediction under Extreme Weather Conditions

In recent years, extreme weather has caused significant damage to the new urban power system, posing unpredictable risks to the safe and stable operation of the new urban power system. In order to ensure the power safety of important urban loads and residential users, it is urgent to achieve high-precision, high-efficiency, and high-performance prediction of power outage losses in urban power grids under extreme weather conditions. In order to predict urban power grid damage, optimize power flow risk scheduling and coordinate heterogeneous resource allocation, this paper processes disaster characteristic variables from three dimensions: power grid equipment factors, disaster weather factors, and geographical and natural factors. The SSA-XGBoost machine learning algorithm is applied to establish an intelligent model for urban power grid damage prediction. Taking the power outage caused by typhoon “Haikui” in a city level power grid as an example, the effectiveness of the integrated intelligent prediction model is verified, and the comprehensive prediction accuracy reaches over 90%. In the future of industrial digitization and intelligent transformation, the data-driven intelligent prediction model for urban power grid disaster losses will play an important scientific guiding significance and reflect the engineering application value for power system disaster prevention and reduction.