The application of rock-filled concrete (RFC) technology in high dams and other large-scale projects is on the rise. Given the elevated seismic fortification intensity at certain project sites is relatively high, there emerges a pressing need for enhanced structural design consideration for RFC dams, necessitating comprehensive research on the dynamic performance of RFC. This study employs a scaled-down testing approach to conduct a series of uniaxial and triaxial compression tests on RFC specimens subjected to varying confining pressures and strain rates. The results indicate that the uniaxial strength of RFC is positively correlated with the strain rate, while it will decline to some extent as the size of coarse aggregates or the dimension of the specimen increases. Confining pressure significantly increases the dynamic strength of RFC. Under the same confining pressure, specimens containing larger aggregates demonstrate lower uniaxial strength but exhibit higher triaxial strength. This observation suggests the mesoscopic characteristics of the self-sustaining skeleton formed by coarse aggregates play a crucial role in the strength of RFC. Based on the test results, a failure criterion is established using the octahedral stress, thereby providing experimental evidence for the seismic design of RFC structures.

As key water resource projects, reservoirs and their efficient operation are crucial to the society. However, most reservoirs, domestic or international, have been formulating the operating rules based on historical statistical data, and the static rules for planning and design limit their capability of proactive responding to floods and droughts. Recent advancements in meteorological and hydrological forecasting have made forecast informed reservoir operation (FIRO) a research hotspot. This paper discusses the main factors that impact the accuracy of numerical weather forecasts, commonly used hydrological models, and rapidly developing artificial intelligence forecasting techniques. The review also covers the FIRO methods and their applications. Finally, we suggest the comprehensive use of various meteorological and hydrological forecasting products to achieve efficient water resource utilization－including global navigation satellite system (GNSS)-based precipitable water vapor, the development of FIRO methods focusing on the vapor-precipitation-runoff three lines of defense, and the construction of deep reinforcement learning operation models that account for the multi-blocking effects of cascade reservoirs.

The prediction of river runoff in a small or medium-sized catchment is constrained by the spatial distribution and density of its rain gauges and record length historical rainfall data. To enhance the accuracy of flash flood early warning and forecasting for such catchments, this study redefines the data structure of an hourly rainfall-runoff model based on the graph theory and the 2000-2014 data of the Shaxi River basin. We use graph neural networks (GNNs) to construct an end-to-end dynamic mapping model for its rainfall-runoff data, and predict its future hydrographs at different forecast periods, using Graph Convolutional Neural Network (GCN), Graph Attention Network (GAT), and Chebyshev Graph Neural Network (Chebnet) models. Mean Absolute Error (EMAE) is used as an evaluation indicator to compare the predictions for the next two hours with those by the Long Short-Term Memory (LSTM) models, Gated Recurrent Unit (GRU), and Artificial Neural Networks (ANNs). The results indicate that for this basin, the Chebnet and GAT models are superior in nonlinear data fitting capability for rainfall-runoff predictions at the forecast periods of one and two hours, improving prediction accuracy by 37.3% to 64.7% compared to LSTM and GRU. The Chebnet model exhibits stable performance in its runoff prediction of the next 15 hours, significantly reducing the impact of timeliness while improving accuracy and applicability. This study has achieved highly reliable predictions of river runoff, useful for early flood warning in small and medium-sized catchments.

Spiral vortex ropes in the draft tube under certain partial load conditions could be frequently observed in the model tests of Francis turbines. In addition to typical vortex rope rotation frequencies fv, pressure signals recorded at the draft tube wall monitor points often exhibit the rotational frequency fn and its higher-order harmonics (1 ~ 5)fn as distinct spectral components. Under certain operating conditions, these special frequency components show significant amplitudes, yet their underlying generation mechanisms remain controversial. This study examines the internal flow characteristics of a Francis turbine operating under typical vortex rope conditions. By combining the data from high-speed imaging with pressure fluctuation measurements, this study presents a systematic analysis on the effects of unit discharge, unit speed, and cavitation number on the vortex rope structure and pressure pulsation characteristics, focusing on analyzing the possible sources of the harmonics (1 ~ 5)fn. The results indicate that unit discharge significantly influences the amplitude of the vortex rope rotation at fv when the spiral vortex rope is visible, pressure pulsation amplitudes at fv increase substantially. In contrast, weaker effects of cavitation number and unit speed are cast on the amplitude at fv. Additionally, under certain specific conditions, a slender and stable cavitating spiral vortex rope forms in the draft tube, where the amplitude at fn sharply increases. This phenomenon is believed to be closely related to hydraulic system resonance triggered by this cavitating rope. The special harmonics at (1 ~ 5)fn are likely attributed to the elliptical cross-section of the vortex rope and its self-rotation, which is characterized by frequencies close to (1 ~ 5)fn.

This study presents an improved YOLOv8n-based detection method to address the issue of false detections of dam cracks that is caused by low-quality surveillance images, limited effective samples, and interference from complex backgrounds. This model is trained using a dataset comprising 193 real-world crack images featuring complex engineering backgrounds, and enhanced by modifying the mosaic data augmentation mechanism and incorporating negative sample training targeted at the objects that were often falsely detected. Numerical experiments demonstrate that under small-sample training conditions, the YOLOv8n model achieves a mean Average Precision (mAP) of 89.2%, meeting the requirements of general engineering applications. After negative sample training, the mAP increases to 92.5%, and the false detection rate is reduced by 10.1%, providing an effective solution to the false detection problem in complex background scenarios. Our findings indicate that the YOLOv8n model is well-suited for dam surveillance images of suboptimal quality, and that the negative sample training strategy significantly improves detection accuracy. This approach offers a novel solution to crack identification in hydraulic projects, practically significant for engineering applications.

Missing value processing is an important foundation for analysis of dam safety monitoring data. Traditional methods for handling the missing values of a dam often use a single type of machine learning models for prediction and interpolation, ineffective in integrating the advantages of multiple types of machine learning models. This article integrates multiple classic machine learning and deep learning algorithms into a strong learner within the framework of ensemble learning. To address the issue of weight allocation to each model, we develop a new critic stacking (CS) weight allocation method so that we can construct a dam monitoring data interpolation hybrid model based on CS ensemble learning. The results show that compared to single base learners and traditional Stacking ensemble models, this CRITIC-Stacking ensemble learning method reduces the RMSE index by an average of 72.7% and 58%. This indicates that the method can fully leverage the predictive advantages of various machine learning models, and the improvement of weight allocation can also improve the predictive accuracy of ensemble learning models, thus providing a new solution for handling missing values in dam monitoring data and constructing prediction models.

Cracking, as one of the most common defects in concrete dams, weakens the integrity and durability of dam structures; crack detection has been a crucial task in the operation and maintenance management of concrete dams. Aimed at the drawbacks of traditional image-processing techniques in crack detection－such as substantial manual intervention and limited generalization ability, this paper presents a semantic segmentation model of dam cracks that incorporates multi-scale features and attention mechanisms. This model uses ResNet-50 as its backbone network for integrating the Path Aggregation Network to recycle shallow features, and makes use of the mechanisms of channel attention and spatial attention. These mechanisms enhance the model's ability to identify critical features, thus effectively improving its segmentation accuracy. Then, based on its semantic segmentation results, the digital image technology is adopted to quantify the geometric characteristics of cracks, including area, length, average width, and maximum width. Tests on a crack image dataset show this new model achieves a crack segmentation Intersection over Union of 82.02% and an F1 score of 90.12%; Quantification results of geometric characteristics exhibit an excellent agreement with the real values and a satisfactory accuracy. Thus, our method demonstrates significant potential for application in crack detection and geometric characteristics quantification for concrete dams.

During hydraulic transition, the highest water level in a pump station forebay determines its design top elevation, and the lowest determines its design bottom elevation. If the designs are not correct, it will overflow or be empty. Determining its extreme water levels is crucial to ensuring the station’s safe operation for the water supply project. Based on the KBM method, this paper derives explicit formulas for calculating the forebay’s extreme water levels through solving a nonlinear dynamic equation that describes the variations in its water level in a long-distance water supply system. We verify the accuracy of these formulas using numerical calculations, and conduct a sensitivity analysis. The results show that relative to the MOC numerical simulations, the error is less than 1.0% for the highest level formula, and less than 2.5% for the lowest level formula. These formulas have simple forms and are convenient for engineering design.

During the operation of a concrete dam, various uncertainties－such as sudden events, natural disasters, and changes in human management－are possible to impose an impact on it, potentially deviating its structure deformation from the conventional patterns. An accurate identification of such changes is crucial for raising the level of concrete dam warning and forecasting. This paper presents an intelligent method for identifying dam deformation under uncertainties. First, we use a spatial clustering method to categorize measurement points that are located in different regions of the concrete dam structure but share certain similarity. Then, a fuzzy clustering (Gath-Geva) algorithm is used to segment a multivariate time series into different phases, allowing its data points to belong to multiple periods based on the membership degree, to measure the homogeneity of segments and detect changes in its hidden structure. Last, we use a fuzzy decision algorithm based on the cluster compatibility criteria to determine the number of segments required, and adopts the principal component analysis (PCA) to identify the number of principal components, further improving the accuracy of the Gath-Geva algorithm. This intelligent method has been applied in a case study of a concrete arch dam structure to identify the changes hidden in the time series of its displacement measurements. Comparison of its results with those of single-period data shows that it is effective in extracting sudden anomalous changes during the operational phase of the dam, and that it is a valuable approach for assessing the operational conditions of concrete dams.

Aiming at the issue of the random motion characteristics of water droplets in a wind field, we formulate a hypothesis that the steady-state deformation of water droplets is ellipsoidal, and uses white noise to describe the randomness of the windward area caused by changes in relative wind speed and water droplet motion posture. A stochastic differential equation is worked out for the motion of ellipsoidal water droplets; their motion shape and drift distance are tested to verify the correctness of this new model. We apply it to calculations of the dispersion coefficients of water droplets under different conditions. The results show that for ellipsoidal droplets in motion, their random effects－produced by the random forces of a large number of air molecules and the changes in their windward area－are directly proportional to their Froude number.

Focusing on the application of forecast-based scheduling in optimal decision-making for complex urban water supply network systems, this study formulates a conceptual framework for integrating hydrological and rainfall forecast information. And we develop an optimal decision-making model for complex urban water supply networks, targeting at dual objectives－to maximize reservoir group safety and economic efficiency, and considering holistically the practical constraints, such as reservoir structural integrity, downstream flood control requirements, water treatment plant intake capacity, maximum allowable pipeline flow rates, and urban water supply reliability. This model has achieved a success in application to Ningbo's municipal water supply network system, validating its operational effectiveness. This case study demonstrates it can effectively adjust scheduling strategies and coordinate reservoir operations based on the hydrological conditions forecasted prior to flood events. It generates scheduling schemes that maintain a high operational safety level (100% water supply guarantee rate) while achieving an water resource utilization rate of 41.3% in flood periods, meeting the design requirements completely. This study helps design and optimize the complex urban water supply network systems.

To improve the accuracy of clearing price estimation in electricity markets with multiple power sources under source-load uncertainty, this paper presents a day-ahead market clearing price interval prediction model. This model considers the uncertainty of generation-side outputs characterized by multidimensional stochastic distributions and the impact of load-side demand response, aimed at minimizing the total generation cost with the participation of hydropower, thermal, wind, and solar power, based on the physical clearing mechanism of electricity markets and the IEEE 30-bus power system. Its solutions give predictions of the interval of locational marginal price and generation commitments, taking source-load uncertainty into account. Case studies show its improvement on bidding efficiency and generation revenue; compared with traditional point forecasting methods, our new model achieves a 4.09% increase in overall economic efficiency and a 0.47% increase in expected revenue. The results verify its feasibility and effectiveness, and provide reliable decision-making support for generators in formulating stepwise bidding and output strategies in the day-ahead market.

In response to the challenge of capacity mismatch in the multi-energy complementary systems located in the complicated terrain areas of the plateau, this paper focuses on the optimization of capacity allocation for hydroelectric-wind-photovoltaic-pumped storage system in the study area of the middle Yarlung Zangbo River basin. We use the Ward's hierarchical clustering algorithm to the field of spatial layout optimization for wind and solar resources development, and construct a refined 8760-hour simulation framework for four-dimensional hydroelectric-wind-photovoltaic-pumped storage. This method breaks through the time resolution limitation of traditional monthly-scale planning. And to verify its applicability, we examine a study cases of the under-construction Jiexu hydropower station, and the planned Yongmu pumped storage power station within the basin. Based on high-precision meteorological data sets, fine evaluation of the wind and solar energy resources in the study area indicates the middle reaches have a development potential of 23.5 million kW for wind power and 238 million kW for solar energy. Our predictions show new energy consumption can be regulated significantly by a hydroelectric-wind-photovoltaic-pumped storage multi-energy complementary system that is based on the two power stations of Jiexu and Yongmu. Wind-solar complementarity enhances the consumption rate; the pumped-storage power stations can mitigate the volatility of wind-solar outputs and increase power grid stability. Furthermore, if considering the flexibility of the demand side, the installed capacity ratio of new energy to pumped-storage power stations would increase significantly.

Trial pit test is a standard method for evaluating rockfill compaction in dam construction. The advent of 3D laser scanning technology offers a rapid, accurate alternative to the traditional water filling method for measuring the pit’s volume. This study develops a new method that integrates data acquisition, registration, and fast volume calculation using 3D laser scanning. Tested against the standard model and a laboratory experiment, the method captures the irregular distribution of rockfill particles effectively; Optimizations in large-scale in-situ tests have achieved volume measurement errors roughly within 5%. Field applications in dam projects further validate the method's efficiency, significantly reducing inspection time and having the potential to replace traditional methods.

As increasingly pumped-storage power stations are built, designing the upper reservoir will encounter more complicated geological issues, and better evaluation on cement grouting projects is demanded. However, the invisibility and complexity of the underground project make the current grouting theory difficult to guide the grouting practice effectively. Artificial intelligence technology has significant advantages in addressing problems characterized by fuzzy constitutive relationships, bringing a new trend in geotechnical engineering field in big data era. But its main inputs are from geological datasets that are usually too small because of difficulties in practical collection limited severely by technical means, time constrains, and costs. Instead, the grouting data, which can be easily acquired in large amounts, is not being fully utilized. This paper presents a new grouting quality evaluation method based on descriptive statistical analysis and spatial statistical analysis on grouting data. It is based on the deep mining of information contained in the 4350 groups of grouting datasets collected from 1104 grout holes and a small amount of geological data from the Wuyue Pumped-Storage Power Station project. The results show that the descriptive statistical characteristics of grout consumption can be used to identify fresh bedrock, general fractured rock mass, densely fractured rock mass, and rock masses that features a higher risk of substandard grouting quality. The grout consumption is correlated with geological conditions to a certain extent, but their impact varies in different grouting sequences due to spatial variations in geological features and different fracture fillings. Neither grouting quality nor grouting efficiency can be evaluated only based on the consumption in the first order grouting. Compared to the limited data from inspection holes, massive grouting data in high density help evaluate grouting quality more comprehensively. Full mining of the information hidden in grouting data is a new approach to full utilization of massive grouting data, which would greatly promote the intelligent development of grouting engineering.

This paper presents a vibration prediction method of pumped storage units based on adaptive features and an optimized kernel extreme learning machine (KELM) to reduce the impact of the nonlinear, non-stationary characteristics of vibration signals on the accuracy of vibration predictions. First, we use improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) to decompose a vibration signal and generate the intrinsic mode components of different frequencies. And, an autoencoder is used to extract the features of these components adaptively and capture their key features dynamically. Then, a KELM prediction model is developed to predict each component separately, using a modified DEIHHO algorithm to optimize its regularization parameter and kernel parameter. Finally, the final prediction result of unit vibration is obtained by superadding the predictions of all the components. Comparison with previous experimental data shows our new method is better in vibration prediction of pumped storage units and improves the accuracy effectively.

Applicability of larger particle size aggregate in asphalt concrete used as impervious panel material helps reduce material costs while improving its mechanical properties. This paper presents a method that increases the maximum particle size Dmax of its concrete aggregate from 16 mm to 26.5 mm and 31.5 mm. Optimal mix parameters are obtained through mix proportion tests for three sets of asphalt concrete with different maximum particle sizes; Their splitting, uniaxial compression, tensile, and slope flow tests are conducted. The performances are compared and the influence of aggregate size on mechanical properties is examined. The results demonstrate the maximum particle size has a significant impact on both its mix proportion parameters and properties. With the maximum particle size increasing, the gradation index increases, while the specific surface area of aggregate and the asphalt-aggregate ratio both decreases (from 7.0% to 6.2%), reducing material costs effectively. Appropriate increase in the maximum particle size helps improve splitting strength (an increase of 4.02%) and compressive strength (an increase of 14.07%) of asphalt concrete, and reduce the slope flow value. However, when the maximum particle size is too large, tensile strength decreases (a reduction of 5.78%), brittle failure dominates the failure mode, and toughness and deformation adaptability are weakened. This study would lay a theoretical basis for further research on large particle size aggregates used in impermeable layer asphalt concrete to build pumped storage power station panels.

To accurately identify the safety hazards at hydropower construction sites in real time, this paper combines the channel attention mechanism and spatial attention mechanism, improves and applies the YOLOv8 algorithm, and develops an automatic identification method of safety hazards in hydropower construction based on the dual attention mechanism. First, based on the YOLOv8 network framework, we construct a channel attention mechanism to highlight key features adaptively, strengthen dynamically the expression of image features of hidden danger areas, and suppress the influence of background noise. Then, a spatial attention mechanism is built that helps weight important regions, reduce background interference, and optimize feature fusion. It allows to adjust attention adaptively, enhance local detail capture and the positioning accuracy, improve the multi-scale target detection ability, and enhance the spatial feature representation ability of the model. Finally, we verify the accuracy and reliability of the model through a case study of an ongoing construction project. The results show that the proposed method identifies the hazards effectively against the interference in the construction site through the attention mechanism, and achieves an accuracy rate of up to 86.2%, better than previous identification models, thereby improving the dynamic management, prevention and control of hydropower construction safety hazards.

High-resolution spatiotemporal precipitation data are critical for accurate hydrological simulations. This paper presents stochastic simulations of the hourly precipitation processes at 142 meteorological stations across the Huang-Huai-Hai Plain, using a Hyetos stochastic precipitation model that couples the Bartlett-Lewis rectangular pulse model with a daily precipitation adjustment algorithm. This model is evaluated comprehensively on its performance using metrics such as basic statistical characteristics, extreme precipitation indices, and intra-day wet-dry characteristics. Results reveal the model accurately simulates the means and standard deviations of hourly precipitation but underestimates its skewness coefficients. While it captures the intensity and frequency indices of certain extreme precipitation events, it underestimates the annual maximum 1-hour precipitation and extreme precipitation intensities. Its biases are evident in simulating intra-day wet-dry characteristics, particularly for those long-duration events. The findings would be useful to precipitation-runoff simulations in the study region.

For gap-graded sand and gravel soils under seepage flow, internal erosion is caused by fine particles migrating through void channels between coarse particles. The migration leads to the redistribution and deformation of the soil skeleton, thereby threatening the safety and stability of earth and rock dams, dykes and so on. Particle morphology along with its correlation with other parameters is one of the most important influences on internal erosion to soil structures. In this study, a custom-developed soil internal erosion set-up is used to conduct experimental tests on three types of gap-graded sand and gravel soils with varying particle morphologies under different hydraulic gradients, focusing on the macroscopic evolution characteristics of soil internal erosion. Using the computational fluid dynamics-discrete element method (CFD-DEM) coupling approach, we consider coarse particle morphology with different sphericity and fine particle content to examine their combined effects on internal erosion from the perspectives of force chains, contact forces, and coordination numbers. The findings indicate that with a fixed content of fine particles in the soil, the higher the sphericity of its coarse particles, the more significant is the number and magnitude of its fine particles that are lost due to internal erosion. The self-locking effect of non-spherical particles enhances its resistance to seepage failure. Additionally, the sphericity of coarse particles is inversely related to the average coordination number. These findings lay a basis for assessing the internal erosion risk of gap-graded sand and gravel soils.