Non-stationary hydrological frequency analysis is a critical scientific research issue for engineering hydrologic design under changing environments. Existing methods exhibit high uncertainty and often fail to account for the spatial dependences among data series of different stations. This paper proposes a regional non-stationary hydrological frequency analysis methodology based on hierarchical Bayesian framework and extreme hydrological regionalization. Firstly， the partitioning around medoids （PAM） clustering algorithm based on the F-madogram variogram is utilized to partition the basin into hydrologically homogeneous regions. And then regional non-stationary hydrological frequency analysis models based on hierarchical Bayesian are constructed. They can be classified into no pooling， partial pooling and full pooling models. Finally， a case study is conducted using the annual maximum 24 hour extreme rainfall data from the Xiangjiang River Basin. Results indicate that based on the spatial clustering algorithm adapting with extreme value theory， the Xiangjiang River Basin with 36 rainfall gauges is divided into three hydrological regions. For Region I， FMA Nino12 is identified as a climatic driver significantly positively correlated with most stations. Compared with the at-site non-stationary model， the partial pooling model exhibits superior performance， effectively capturing the regionally homogeneous response to climatic drivers while preserving individual site characteristics. Simultaneously， the regional models exhibit a great benefit with the uncertainty for regional parameters at each station reduced by about 10% to 35%. Moreover， as the return period increases， the non-stationary return periods consistently decrease， compared with the stationary conditions. For instance， at the 50-year return period， the reduction rate approaches 60% for some stations， indicating an increased probability of extreme rainfall events within the region driven by climatic factors. This methodology enriches the methodological framework of non-stationary hydrological frequency analysis and can provide a scientific basis for determining the design rainstorms in basins and formulating disaster prevention strategies.

To overcome the limitations of existing research in effectively addressing the high-dimensional and nonlinear hydraulic processes in open-channel sections of water transfer projects under water diversion disturbances during the icing period， and the high reliance on manual experience in actual operations， this study takes the Wangnou-ruwugou section of the Jiaodong Water Transfer Project as a case study to investigate real-time intelligent hydraulic regulation of gate-pump groups in open channels. Through one-dimensional hydrodynamic simulations of gate-pump groups， step disturbances were applied to the diversion flows of canal pools to reveal the coupling mechanism between hydraulic processes in open channels and the operational responses of gate-pump groups under various diversion disturbances， and thus determine safety thresholds for water diversion disturbances during the icing period. On this basis， a real-time intelligent regulation model for gate-pump groups in open channels was developed by coupling a hydraulic model with a deep reinforcement learning algorithm， which excels at handling nonlinear， high-dimensional problems and requires minimal modeling data. The robustness of the proposed model was validated under various operational conditions. Application results demonstrate that the derived scheduling strategy effectively raises and stabilizes the water level at control sections to the target ice-period regulation level while ensuring operational safety and reducing the gate adjustment frequency.