Optimization Method for Real-time Control Rules of Gates and Pumps Based on Cluster Analysis

　　The coastal cities are under the double pressure of extreme heavy rainfall and the topographic obstruction caused by the tidal current of the estuary. The risk of flooding is becoming increasingly prominent. However, the traditional real-time control rules for sluice gates and pumps mostly take the location of the project as the reference section for regulation, and it is difficult for the internal areas of the complex river network to respond quickly to the regulation. To solve this problem, this paper takes the central area of Cangshan District in Fuzhou City as the research area, uses InfoWorks ICM to construct a city hydrological and hydrodynamic model and conducts verification. Then, a hierarchical clustering framework composed of DBSCAN spatial clustering and X-Means feature clustering is proposed. Firstly, the regulation zones are divided, and then the key control sections within each zone are adaptively identified based on the total value of the surface inundation characteristics of the overflow section of the river channel, and the real-time control rules for the sluice gates and pumps are formulated with the new sections as the reference. Finally, the performance of the measured typhoon assessment method is evaluated. The results show that the hierarchical clustering framework clusters into three regulation zones, and each zone clusters into one key control section. Under the "Haikui" typhoon, compared with the original rules, the average total water level of the three key control sections decreased by 6.53%, 10.66%, and 12.38% respectively, and the peak water level was reduced. Meanwhile, the proportion of the submerged area with water depth exceeding 1 meter decreased from 15.25% to 12.38%, and the maximum submerged depth decreased by 0.49 meters. It effectively improves the regulation ability of the sluice gates for regional flooding, providing new technical support for
　　coastal cities to cope with extreme heavy rain and floods.