After the Deneng Xiangjiang Hydropower Station replaced the 4-blade runner with a 5-blade runner, the unit started experiencing severe vibrations, which eventually led to coupling vibrations in the powerhouse structure and posed significant safety hazards. This paper explores the vibration source and employs Finite Element Analysis (FEA) and Computational Fluid Dynamics (CFD) methods for analysis. Finally, modal verification through actual measurements validates the relevant conclusions. The study reveals that the resonance is caused by the multiple frequency relationship between the runner blade rotational frequency and the natural frequencies of the powerhouse structure. The research on the vibration issue at this power station hold great practical value.

In the condition monitoring of pumped storage units, vibration velocity signal is often integrated into the displacement signal, and then calculate characteristic values for unit evaluation. Taking the horizontal vibration velocity signal of upper bracket of a case unit under stable working condition as an example, the method of integrating velocity signal into displacement signal and the setting of filter parameters are studied. Furthermore, vibration signal in transient working condition is adopted for verification. The results indicate that a high-pass filter shall be used when the vibration velocity signal is integrated into the displacement signal under stable conditions, and the cut-off frequency of the high-pass filter can be set as per 0.2 times of the rotational frequency. Under transient conditions, there is a risk of over estimating unit vibration when using unfiltered displacement signals integrated from velocity signals to calculate characteristic values for unit safety assessment.

Aiming at the two failure modes of hydraulic steel gates, the relative deformation failure under the normal service state and the bending shear composite strength failure under the limit state of bearing capacity, the corresponding failure probability analysis models were established respectively. The influence of different parameter distribution types on the failure probability of two failure modes of outcrop steel gates and deep-hole steel gates in hydraulic steel gates was explored through Monte Carlo simulations. The results show that under the same failure mode, when the coefficient of variation of the variable itself remains unchanged, the failure probabilities obtained by different distribution types are relatively close. Compared with the outcrop gate, the deep-hole gate has a higher failure probability and is more prone to instability. But the type of parameter distribution has little influence on the failure probability of deep-hole gate and has great influence on the outcrop gate. When the random variables of each parameter follow the logarithmic normal distribution, it will have a relatively maximum failure probability. It is easy to overestimate the reliability of hydraulic steel gates by considering only the normal distribution.

In respect to the network security situation and attacks of new types, the needs of trusted computing of applications in SCADA for pumped storage power station was analyzed. Based on the trusted hardware platform and operating system, a system structure of trusted applications in SCADA was proposed, and the key techniques such as trusted distributed services and trusted behavior analysis were discussed. The results show that the proposed design realizes the trusted applications in SCADA, and provides active defense mechanism for SCADA, and adapts to the development of prospective SCADA applications for pumped storage power station.

After a blackout occurs resulting from an extreme disaster, the active distribution network with distributed generation and energy storage equipment can perform service restoration by island partition and network reconfiguration. To solve the service restoration problem for the active distribution network with uncertain wind power output, this paper proposes a two-stage service restoration model based on robust stochastic optimization. In the first stage, the event-wise ambiguity set of wind power output is constructed based on the historical data and the model which minimizes the outage cost is solved by the robust stochastic optimization method. In the second stage, the energy storage and controllable loads in the distribution network are used to track the wind power output, so as to optimize the scheduling during the fault recovery of the distribution network, which minimizes the outage cost and the total power loss. Finally, the superiority of the proposed the model and strategy is verified by the simulations of a modified IEEE 33-bus distribution system case.

Based on the daily precipitation data of 135 weather stations from the Yangtze river delta city during 1951-2017, spatio-temporal patterns of the extreme precipitation of the Yangtze river delta agglomeration were analyzed by the linear trend analysis, M-K trend test, wavelet cycle analysis and reverse distance weight interpolation methods. The extreme precipitation indexes were defined from the aspects of extreme precipitation amount, frequency, and intensity. The results show that except for a decrease in the continuous drought index of -0.11 d/a, the extreme precipitation indexes showed an increasing trend. The average period of change in extreme precipitation indicators was in the range of 15 a and 25 a, and the characteristic timescale was 40-50 a; The high-value area of the annual mean spatial distribution of maximum daily precipitation, maximum daily precipitation, continuous moisture index, strong rainfall, heavy rain days and annual rainy day precipitation presented in the southeast areas of the Yangtze River Delta urban agglomeration, which were distributed the opposite parts to the persistent drought index. However, the high-value areas of ordinary daily precipitation intensity were in the western part of the Yangtze River Delta urban agglomeration; The high-value areas of the changing trends of the spatial distribution and the maximum precipitation in the Yangtze River Delta urban agglomeration were all in Nanjing, Zhenjiang, Changzhou and Shanghai. The high-value areas of the spatial distribution change trend of ordinary daily precipitation intensity and annual rainy daily precipitation change trend were in Nanjing, Wuxi and Suzhou. The high-value areas of continuous moisture index, continuous drought index and heavy rain days were all in Hangzhou, Shaoxing and Wenzhou. The research findings can provide references for the prevention of rainstorm waterlogging disasters in the Yangtze River Delta Urban Agglomeration.

The medium and long-term runoff simulation can provide a scientific basis for the rational allocation of water resources, which is of great significance to the high-quality development of the basin. Based on the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and singular spectrum analysis (SSA) two-stage decomposition strategy, the monthly runoff simulation model ICEEMDAN-SSA-WOA-LSTM was constructed by using the long short-term memory network model (LSTM) optimized by the whale optimization algorithm (WOA). It was applied to the simulation of monthly runoff in Stone River Reservoir and compared with the single decomposed ICEEMDAN-WOA-LSTM, SSA-WOA-LSTM and the undecomposed WOA-LSTM models. The results show that the ICEEMDAN-SSA-WOA-LSTM model has the best simulation effect, and the three evaluation indexes in the calibration period and validation period are better than other models, with the root mean square error of 1.278 m³/s, the average absolute error of 0.893 m³/s, and the Nash efficiency coefficient of 0.985 in the validation period. The two-stage decomposition strategy model can significantly improve the accuracy of monthly runoff simulation and can be used for year-round incoming runoff simulation.

Simulation and prediction of runoff under climate change and land use scenarios are of great significance for the study of water balance and water resources planning and management. Taking the Qingshui River Basin in Zhangjiakou City as the research area, the global climate model GFDL-ESM2M and CA-Markov model were used to analyze and predict the meteorological data and land use in the Qingshui River basin, and the SWAT hydrological model was constructed to quantify the changes of water balance factors in the basin in 2025 under the joint influence of climate change and land use. The results show that under the three GHG emission scenarios, the rainfall in Qingshui River Basin increased significantly in 2025, the maximum temperature and average temperature under the RCP2.6 emission scenario decreased compared with 2015; The maximum temperature and average temperature under the RCP4.5 and RCP8.5 emission scenario increased, and the minimum temperature under the three scenarios decreased. From 2015 to 2025, arable land, woodland, grassland, water area and construction land changed by -6.24 %, -0.86%, 6.32%, 0.20% and 0.59%, respectively. Compared with 2014-2015, the water balance distribution of the watershed changed in 2025, and the peak monthly runoff occurred from July to September. The average annual runoff under the three discharge scenarios were 4.40 m³/s, 5.84 m³/s and 9.94 m³/s, respectively.

In order to provide a suitable means for the tidal prediction of the Changjiang Estuary and its adjacent area and for the open boundary conditions of tidal flow numerical model of the Changjiang Estuary, accuracy of NAO.99Jb was evaluated. The feasibility of providing open boundary conditions for tidal flow numerical model of the Changjiang Estuary was studied. The results show that the NAO.99Jb model can accurately simulate the propagation of the tide outside the Changjiang Estuary and its adjacent area; The NAO.99Jb has good accuracy outside the Changjiang Estuary, but has lower accuracy inside the Changjiang Estuary; It is feasible for using the tidal level predicted by NAO.99Jb as the open boundary condition of the tidal flow numerical model of the Changjiang Estuary, and the accuracy is good, which can be further popularized and applied.

In order to address the limitations in forecasting the maximum scour depth of conventional river bends, this study amalgamated the methodologies of isolated forest (IF) and gene expression programming (GEP). An IF-GEP model for predicting the maximum scour depth of river bends was established. The validation results demonstrate that the IFGEP prediction model surpasses existing formulations in terms of its accuracy on the test set. Moreover, it exhibits enhanced predictive performance compared to the traditional GS-SVR and RF models. Application of the prediction model to various rivers yielded remarkably close results to the actual measured values, affirming its strong predictive capability and robust generalization performance.