The frequent occurrence of extreme rainstorms leads to an increase in urban flooding disasters, and flooding damage to power system equipment can cause large-scale power outages, affecting social production and life. This paper discusses the power system recovery and intelligent operation and maintenance equipment scheduling after extreme rainstorms triggering flooding disasters. By constructing a grid fault diagnosis alarm model based on decision tree algorithm, a hierarchical fault information diagnosis algorithm is applied to realize rapid power system recovery. The study utilizes ID3, CART, and C4.5 decision tree algorithms to construct the grid scheduling alarm model, and the experiments show that the C4.5 decision tree model has an accuracy of 96.91% on the training set, which is 2.66% higher than that of the ID3 algorithm; the weighted W-C4.5 decision tree’s accuracy rate for the “safe” state can reach 98.11%, and the recall rate for the “risky” state can be increased to 79.65%. The recall rate for “risky” states increased to 79.65%. The single-factor analysis shows that the load has the greatest influence over the number of risks, with an average K-value of 0.4578. This study provides theoretical foundation and technical support for power system recovery decision-making and intelligent operation and maintenance equipment scheduling under the flooding disaster, and is of great significance for improving the postdisaster recovery capability of urban distribution networks.
