In order to accurately reflect the transient process and pressure wave propagation in an axial-flow nuclear main pump under accident conditions, a three-dimensional closed equivalent flow field model of the reactor primary circuit system was established. Based on computational fluid dynamics(CFD)methods, the transient pressure waves and hydraulic loads within the piping system during a main pump rotor seizure accident were investigated. The results indicate that rotor seizure induces a significant water hammer effect, causing local sudden pressure changes that propagate as pressure waves along the pipeline. The highest pressure peak(16.60 MPa)is monitored at the main pump inlet, while the lowest pressure trough(14.61 MPa)occurs at the pump vessel outlet. The pressure waves propagate from these locations along the pipeline and gradually dissipate. Significant hydraulic loads are generated at locations with changes in flow cross-section or direction changes. The maximum load 3.22×10⁶ N occurs at a bend near the main pump inlet and the minimum load 8.94×10⁵ N occurs at a bend near the reactor pressure vessel inlet. Spatially, the steam generator outlet exhibits higher pressure due to its proximity to the main pump inlet, and the top of the U-tube shows further increased pressure due to flow direction change. In contrast, the inlet section and resistance components expe-rience lower pressure, and the reactor pressure vessel inlet shows a more pronounced pressure drop due to its closeness to the main pump outlet. This study reveals the spatiotemporal evolution of pressure waves and hydraulic loads within the system under rotor seizure conditions, providing important engineering insights for safety design and transient dynamic analysis of reactor primary circuits.

To investigate the influence of sand parameters on the flow characteristics within a centrifugal pump, numerical simulations were conducted using the Eulerian multiphase flow model under solid-liquid two-phase flow conditions. A filtered turbulence model was incorporated to refine the local time-homogenized turbulence closure. The effects of sand volume fraction and particle size on the solid-liquid flow field and associated hydraulic losses were systematically examined. The results indicate that as the sand concentration increases, the head and efficiency of the centrifugal pump decrease by an average of 3.7% and 4.5%, respectively. The region of high sand volume fraction near the pressure side of the blade expands, and particle accumulation becomes more pronounced in the middle and lower sections of the flow passage. Consequently, flow losses intensify along the pressure side of the blade and along the front and side regions of the volute. In contrast, an increase in sand particle size leads to an average reduction in pump head and efficiency of 2.5% and 3.7%, respectively. The distribution of sand particles within the impeller becomes more uniform, with a reduction in localized high-volume fraction zones. Although the radial distribution pattern of particles remains largely unchanged, the particle velocity tends to decrease. Additionally, flow losses along the blade pressure side and the volute are noticeably reduced. These findings provide theoretical support for the hydraulic design of centrifugal pumps handling solid-liquid two-phase flows.

To tackle the problems of substantial energy loss being converted into heat during the operation of liquid-ring pumps, which causes a temperature rise of the medium and consequently leads to a decline in pump performance, the simulation combined with experimental test, the RBF neural network surrogate model and NSGA-Ⅱ genetic algorithm were used to conduct a flow-thermal coupled multi-objective optimization analysis of the liquid-ring pump′s blade profile. The results show that the thermal equilibrium state of the flow in the pump startup process lags behind the equilibrium state of the hydraulic characteristics for a long time, and the source terms contributing to the increase in thermal energy include compression power and dissipation loss terms. The dissipative power is primarily concentrated on the outer rim of the impeller, the backside of the blade outlet, as well as the inner wall of the casing within the suction and compression zones. The regions with significant compressive power are mainly located in the gas compression zone and the exhaust zone. The prediction errors of the constructed thermal-efficiency surrogate model and vacuum-efficiency surrogate model are both within 5%. In the Pareto frontier of multi-objective optimization, the thermal energy of the model with the minimum thermal energy is 13.80% lower than that of the original model, and its blades feature a larger outlet setting angle β₂ and a smaller wrap angle φ. The efficiency of the model with optimal efficiency has improved by 7.97% compared to the original model, and its blades feature a larger outlet angle β₂ and wrap angle φ. The vacuum of the vacuum optimal design is 7.32% higher than that of the original design, and its blades show smaller outlet angle β₂, and larger wrap angle φ. With the increase of outlet angle β₂ and the decrease of wrap angle φ, the thermal energy gradually decreases, and there is a certain positive correlation between thermal energy and vacuum degree. The research results can provide a certain theoretical support for performance optimization of liquid-ring pumps.

To improve the efficiency of high-pressure liquid recovery and utilization by turbines in the seawater desalination field, a forward-bent blade turbine was selected as the research object. With hydraulic efficiency at 0.6Q_d, 0.8Q_d, 1.0Q_d and 1.2Q_d operating conditions as the optimization objectives, combined with the matching performance of hydraulic components, five design variables of the volute and impeller were screened through Plackett-Burman experiments. A total of 100 sets of experiments were designed via optimal Latin hypercube sampling, and an intelligent optimization platform was established based on the Isight software. The multiple flow components optimization was accomplished by coupling the RBF neural network with the NSGA-Ⅱ algorithm. The research results indicate that the weighted average efficiency of the optimized turbine is increased by 2.198%, and the hydraulic efficiencies under the four operating conditions are improved by 0.170%, 1.990%, 3.230% and 2.370% respectively. The performance and stability of the turbine under both partial load and overload conditions are significantly enhanced, and the operating range of the high-efficiency zone is expanded. After optimization, the intensity and scope of the separation vortex induced by the mismatch between the volute outlet angle and the blade inlet setting angle are reduced, and the unstable flows such as impeller inlet backflow and flow separation are obviously mitigated. The vortex intensity near the volute tongue is decreased, the secondary flow is reduced, and the influence of rotor-stator interaction is weakened. This demonstrates that the matching performance between the volute and impeller of the optimized turbine is improved, and the flow field control in the rotor-stator interaction area is streng-thened. Meanwhile, due to the optimized flow field control, the flow channel energy loss and multi-operation condition energy dissipation are reduced.

In order to investigate the vortex dynamics characteristics in rotating cascades during cavitation, numerical simulations were performed using the detached eddy simulation(DES)model combined with the Zwart-Gerber-Belamari cavitation model, and experiments were carried out to verify the reliability of the numerical calculation method. Furthermore, the distribution laws of cavitation bubbles, streamlines, vortex structures, vorticity, vorticity transport terms, and turbulent kinetic energy in the impeller were analyzed. The results show that as the cavitation number decreases, high-velocity fluid is formed in the flow passage and disturbs the cavitation bubble interface. Cavitation bubbles and vortices are coupled with each other and block the flow passage, leading to a decline in the hydraulic performance of the rotating cascade. Furthermore, the growth of cavitation bubble volume promotes vortex generation, while the intense gas-liquid exchange at the vortex and cavitation bubble interfaces is a key factor affecting vorticity changes. The contribution of the Coriolis force term to vorti-city is relatively small, however, at low cavitation number, the contribution of the Coriolis force term to vorticity in the vortex structure region at the tail of cavitation bubbles increases significantly. As the cavitation number decreases, the region of high turbulent kinetic energy variation region and the turbulent kinetic energy loss per unit mass of fluid in the impeller gradually shift from the blade leading edge to the blade trailing edge. The main causes of turbulent kinetic energy loss include the collision between fluid and blades, cavitation development, and the vortex motion at the tail of cavitation bubbles during the unsteady cavitation stage. The research results can provide certain theoretical support for reducing the cavitation flow instability of hydraulic machinery.

To investigate the effects of air flow rate and noise on ducted fan performance, a turbine rotor was used to replace the traditional multi-blade centrifugal rotor.Five key structural parameters of the turbine rotor were optimized by combining the surrogate model with genetic algorithm, aiming to further improve the performance of the turbine fan. 100 sets of samples were generated by the Latin hypercube experimental design method, and the regression Kriging model was established by taking the maximum airflow of the turbine fan as the objective parameter and combined with computational fluid dynamics(CFD)simulations. Subsequently, a genetic algorithm was used to optimize the regression Kriging model. The results show that the regression Kriging model can accurately describe the relationship between the key design parameters of the rotor and the maximum airflow of the turbine fan. Specifically, the inlet diameter D_j of the turbine rotor and the outlet angle β₂ of the blade exert a significant influence on the maximum airflow of the fan, and with the increase of the value, the functional force of the rotor is enhanced. The blade bending degree f exhibits a moderate influence, while the number of blades N and the inlet angle β₁ of the rotor have little influence on the maximum airflow of the fan. After the optimization, the rotor increases the maximum airflow of the fan by 22.4% at the same speed, and reduces the noise by 4.00 dB at the same airflow conditions.

Paddle mixers are widely used in coagulation processes, and by adjusting the placement angle of the blades, the water bodies in the sedimentation tank exhibits different flow structures. The mi-xing zone of a high-efficiency sedimentation tank was taken as the analysis object, and three different flow structure mixing modes were designed for the paddle mixers. Turbulent kinetic energy was selected as the evaluation index, and a technical method combining computational fluid dynamics software STAR-CCM+ with model experimental research was used to seek the optimal mixing scheme. Among them, the placement angles of the upper and lower blades in scheme 1 were both the +45°, while the placement angles of the lower blades in scheme 2 and the upper blades in scheme 3 were both the -45°, respectively. Based on the comprehensive CFD calculation results and experimental data, the evaluation of water treatment effectiveness of them are consistent, and the turbulent kinetic energy in the rotating domain of the blade is relatively high. The turbulent kinetic energy gradually decreases in the stationary domain, which is far from the blade. The turbulent kinetic energy in the middle water body between the two blades is higher than that in the upper and lower water bodies. Under the same stirring speed, the average velocity of flow and turbulent kinetic energy in scheme 3 are the highest, but scheme 2 has the highest average turbulent kinetic energy value under the same power consumption. The experimental results show that the turbidity removal rate of all schemes reaches its highest value under the treatment capacity of 10 m³/h, while the stirring speed is 60 r/min. The removal effect of scheme 2 is significantly better than scheme 1. The removal efficiency of scheme 2 can reach the maximum, 79.18%. The research results indicate that it is reasonable to select the turbulent kinetic energy value as the evaluation index to water treatment, either too small or too large turbulent kinetic energy will lead to poor effluent quality. Therefore, designing a reasonable mixing hydraulic component, optimizing the flow structure by controlling turbulent kinetic energy is of great significance for improving flocculation reaction and enhancing removal efficiency.

In order to effectively predict the pressure pulsation of the hydraulic turbine draft tube and take corresponding measures to reduce the pressure pulsation, a hybrid prediction model was proposed, which was based on complete ensemble empirical mode decomposition(CEEMD), the improvement of the bald eagle search algorithm(IBES), and the extreme learning machine( ELM)to predict the pressure pulsation signal of the draft tube. Firstly, the non-smooth hydraulic turbine draft tube pressure pulsation signal was split into several relatively stable submodal parts based on the CEEMD decomposition. Secondly, the decomposed submodal data were input into the ELM model for in-depth training and prediction analysis, and the initial weights and thresholds were optimised by IBES. Finally, the prediction result outputs of each sub-modality were superimposed to obtain the final prediction results of the pressure pulsation signals of the hydraulic turbine draft tube. The simulation results show that the proposed CEEMD-IBES-ELM prediction method can reduce the complexity of the prediction process with relatively low reconstruction error. In addition, compared with other models, this model demonstrates significant superiority in terms of prediction accuracy and stability, and has good potential for application.

To explore the influence of basalt fiber on the mechanical properties and durability of concrete, indoor rapid freeze-thaw tests were conducted to simulate the actual erosion environment. The evolution laws of mechanical properties of basalt fiber(BF)concrete with different volume fractions under freeze-thaw erosion in two media, namely clear water solution and 3%(mass fraction)NaCl+5% Na₂SO₄ composite salt solution, were analyzed. The deterioration laws of the microstructure of basalt fiber concrete were also analyzed by combining nuclear magnetic resonance and ultrasonic tests. Simultaneously, a life prediction model was established to predict the maximum service life of basalt fiber concrete structures. The research results show that the mechanical properties and microstructure of the specimens are much more damaged during freeze-thaw in composite salt solution than in clear water. The mechanical properties of concrete with basalt fiber are all improved, and the salt-freeze resistance of concrete is the best when the fiber volume fraction is 0.15%. The life prediction results of the grey prediction model and the Weibull model are roughly similar, among which the Weibull prediction model can accurately predict the service life of concrete. The research results can provide a theoretical basis for the study of concrete durability performance and the maintenance of hydraulic structures in cold and arid regions.

Traditional prediction methods based on statistics or machine learning often struggle to effectively capture the complex mapping relationships between the displacement of concrete arch dams and various influencing factors. Thus, a novel deep learning prediction method was proposed. This method integrated densely connected convolutional networks(DenseNet)and gated recurrent unit(GRU)to form a DenseNet-GRU model, aiming to enhance the accuracy and generalization ability of deformation prediction for concrete arch dams. A typical concrete arch dam located in a certain region of China was selected as a case study, and deformation monitoring data from multiple measuring points were used for empirical analysis. The results indicate that the DenseNet-GRU model can accurately simulate the displacement deformation process of all monitoring points. Compared with other deep learning models, it demonstrates higher prediction accuracy and stronger generalization capabilities. This research provides an efficient and reliable prediction tool for dam safety monitoring and health management, and holds significant theoretical and practical implications for the advancement of dam safety management practices.

A mathematical model for calculating the hydraulic disturbance transient process of the water delivery and power generation system of a pumped storage power station equipped with a surge chamber was established based on the characteristic line method. The differences in the maximum output swing amplitude of the disturbed unit and the peak value of the surge in the surge chamber were investigated under two modes, namely frequency regulation and power regulation. The influences of the surge chambers geometric parameters including its location, the large well size, and the impedance orifice diameter on the regulation quality of the transient process under hydraulic disturbance were analyzed. The results indicate that, under the premise of satisfying relevant specifications of the power station, appropriately reducing the distance between the surge chamber and the powerhouse can significantly decrease the maximum output swing amplitude of the disturbed units while keeping the changes in other parameters relatively small. The surge peak value in the surge chamber under power regulation mode is greater than that under the frequency regulation mode, while the maximum output of the disturbed units under the frequency regulation mode is greater than that under the power regulation mode. In order to improve the regulation quality of the power station during the hydraulic disturbance transient process, appropriate diameters of the impedance orifice and the large well of the surge chamber should be selected, which can minimize the maximum output swing amplitude of the disturbed units and enable the rapid attenuation of the surge in the surge chamber.

In order to study the impact of aerated drip irrigation on water and nitrogen transport under varying soil bulk densities, an indoor soil box infiltration experiment was conducted, incorporating two irrigation methods(aerated drip irrigation and conventional drip irrigation), two nitrogen application levels(300 and 700 mg/L), and three soil bulk densities(1.30, 1.35, and 1.40 g/cm³). The experiment aimed to examine the patterns of water and nitrogen transport under different irrigation treatments. The interaction mechanisms between soil aeration and water-nitrogen dynamics were analyzed by using structural equation modeling. Results show that under aerated drip irrigation, water transport velocity in soil increases significantly. As soil bulk density increases from 1.30 g/cm³ to 1.35 and 1.40 g/cm³, soil moisture content generally declines by 7.23% and 6.49%, respectively, compared to previous lower density treatments. Compared to conventional drip irrigation, aerated drip irrigation increases nitrate nitrogen content by 10.96% on average, while reducing ammonium nitrogen content by 12.13%. Both nitrate and ammonium nitrogen decrease with increasing bulk density at the same soil depth. Furthermore, soil aeration gradually decreases with increasing soil bulk density, while aerated drip irrigation significantly enhances soil aeration. Structural equation modeling reveals that aerated drip irrigation effectively improves soil aeration, which is also significantly influenced by soil temperature, thereby affecting mineral nitrogen content. This study provides a theoretical basis for understanding the mechanisms of aerated drip irrigation and the transport of water and nitrogen under different soil bulk densities, and offers valuable insights into the role of soil aeration in regulating water-nitrogen distribution.

To build a lightweight tomato plant water stress recognition model for efficient monitoring of plant water status and to provide technical support for precise irrigation decision-making, a lightweight recognition model based on improved YOLOv11 was proposed. The core methodology included the novel design of a lightweight detection head(CDG)and the replacement of the standard convolutions in the C3k2 modules of the backbone and neck networks with the more computationally efficient partial convolution(PConv). This resulted in a new model named YOLOv11-C3k2_PConv-CDG. This approach aimed to significantly reduce both the model parameter count and computational complexity.Experimental results show that this model performs well in the task of monitoring tomato water stress. Compared to the baseline model, it maintains recognition accuracy while reducing giga floating-point operations(GFLOPs)by 0.8 G, decreasing model size by 0.8 MB, and shortening training time by approximately 2 hours.The proposed YOLOv11-C3k2_PConv-CDG model significantly improves efficiency while ensuring high accuracy, offering an effective new technical approach for the real-time and efficient monitoring of tomato water stress.