Setting a panel outside the soilbags-constructed retaining wall may affect the behavior of the wall. Therefore, a full-scale test of the soilbags-constructed retaining wall with fixed panels and with unfixed panels were conducted to explore the influence of the panel on the deformation mode and lateral earth pressure of the wall. The results show the walls with fixed and unfixed panels deformed laterally in a cantilever-like manner, but the lateral earth pressure acting in the fixed-panel wall is greater than that in the unfixed-panel wall. A calculation formula for calculating the earth pressure acting on the soilbags-constructed retaining wall under limited displacement was established based on the linear elastic constitution theory, and then it was used to analyze the relationship between the lateral earth pressures and lateral strain. It is found that distribution of the lateral earth pressure of the soilbags-constructed retaining wall with unfixed panel is close to the calculated active earth pressure as a result of the poor deformation constraint of the unfixed-panel, which indicates that the flexible characteristics of the soilbags-constructed retaining wall can be fully utilized by the wall with unfixed panel, so as to effectively reduce the earth pressure behind the wall.

Early and timely defect detection is essential to ensure the safe operation of hydraulic concrete structures. The deep learning-based computer vision method does not require complex manual feature engineering, and can automatically determine the category of structural defects in remote images, overcoming the shortcomings of traditional manual vision that are labor-intensive, subjective and prone to errors. Inspired by this, this paper proposes a deep learning-based defect detection method, which introduces attention mechanism into the ResNeXt50 network to adaptively recalibrate the channel-level feature responses, so that the model can pay more attention to the defect information in the image and enhance the feature extraction ability. Test results show that the proposed method can achieve an F1 score of 88.0%, and realize a good classification effect for common concrete defects.

The geometric characteristics of landslide are closely related to the rainfall infiltration process, which is the key internal factor to control the stability of rainfall-induced landslides. There are many rainfall-induced landslides in the Three Gorges Reservoir area, and it is of great significance to deeply reveal the body shape characteristics of rainfall-induced landslides (landslide whose stability is sensitive to rainfall in hydraulic engineering are generally referred to as rainfall-induced landslides) for the identification and prevention of landslide disasters in the reservoir area. In this paper, the data of 275 rainfall-type sedimentary landslides in the Three Gorges Reservoir area were collected, and their body shape characteristics were statistically analyzed. The rainfall-type landslides in the area are narrow and long, thin in thickness and gentle in slope. Combined with the numerical simulation of the rainfall infiltration process of typical landslides, it is found that the landslides of this type have a larger rain-receiving area and infiltrated rainwater is more likely to infiltrate the sliding bed, which makes the effect of rainfall more significant, and is not conducive to the stability of the slope under rainfall conditions.

Baishuihe landslide is a typical representative of bedding landslide in the Three Gorges reservoir area. However, the research on its formation evolution process and deformation characteristics after water storage is not deep. On the basis of field geological investigation and the latest survey data and GPS monitoring data, this paper makes a detailed study on the rock and soil structure of Baishuihe landslide and its deformation characteristics after impounding. The results show that the evolution process of Baishuihe landslide has gone through the early stage of landslide, the rock landslide and the accumulation landslide. At present, Baishuihe landslide is in the accumulation landslide stage; Baishuihe accumulation landslide is the product of vertical superposition of landslide accumulation bodies caused by the destruction of the second-stage rock landslide, and there is a clay layer in the middle. Corresponding to the water level of 135 m and 156-175 m, the deformation of Baishuihe landslide can be divided into two stages. The deformation characteristics of Baishuihe landslide are gently inclined steps and steeply inclined steps. The research results provide the corresponding decision-making suggestions for the treatment of Baishui River landslide, and also provide the direction for the later treatment of geological disasters with similar structural characteristics.

In order to explore the feasibility of using waste concrete to dispose expansive soil and make it as roadbed filler, the waste concrete and expansive soil were mixed evenly according to the ratio of 96 : 4, 92 : 8, 88 : 12, 84 : 16, 80 : 20 and 76 : 24. The boundary moisture content, linear shrinkage rate, expansion force and compaction tests were carried out. The influence of curing time was considered in unconfined strength and CBR tests to explore the improvement effect of the selected materials. The experimental results show that the liquid limit and plastic index decrease first and then increase with the increase of waste concrete content, and the minimum value appears when the content is 12 %, while the plastic limit increases slowly. The linear shrinkage rate increases first and then decreases with the increase of dosage, and the linear shrinkage rate is the smallest when the dosage is 12 %. The expansion force under different curing periods decreases slowly with the incorporation of concrete. When the dosage is 12 %, the dry density of the treated expansive soil is greater than that of the plain soil, and the optimum moisture content is less than that of the plain soil. Considering various factors, 12 % is determined as the best dosage. Under this dosage, the unconfined compressive strength of the mixed soil can reach the peak value, and the bearing capacity of the treated soil is about 5 times that of the plain soil. The treated expansive soil can be used as roadbed filler.

Aiming at the nonlinear problem of regret value when regret theory is used to solve multi-attribute group decision making problems in hydraulic engineering, a multi-attribute group decision making method combined with grey correlation analysis method is proposed. Firstly, the regret theory was used to obtain the decision maker's perceived utility for each option and determine the regret decision matrix. Secondly, according to the nonlinear problem between regret values, the weight of attributes and the comprehensive regret value of each scheme were calculated by grey correlation analysis, and then the best scheme was obtained according to the ranking of the comprehensive regret value. Finally, the feasibility of the method was verified by case analysis. The results show that this method can solve the nonlinear characteristics and weight problems in regret theory, and make decision-making more reasonable when considering psychological factors.

Considering the importance of reasonably setting water prices for water transfer projects in promoting the healthy operation and sustainable development of the projects, taking water resources allocation project around Beibu Gulf as an example, this paper put forward a new idea to calculate the water price. Firstly, the influencing factors of the project water price were identified through sensitivity analysis. And then through in-depth consideration of the value of each factor and its time change process, the calculated water price is low before and high after, which is realistic and easy to bear as well as conducive to the connection between the water transfer price and the current water price in the receiving area. The results show that the design parameters of water price scheme are reasonable and the calculation results are feasible. This study can provide a certain reference for the water receiving area government in the cost accounting and water price formulation of this project.

In hydro-generator units, the thrust bearing oil tank has the largest volume and the internal structure is relatively more complex. Therefore, there are many oil mists in the oil tank, and there are often internal and external oil spills, which have a great impact on the safe and stable operation of the unit. Based on the basic theory of computational fluid dynamics, the calculation model of thrust bearing in a power station was established, and the number of pressure oil blade was changed. The ICEM was used to mesh the three-dimensional model, and Fluent was used to calculate. The CFD-POST was used to analyze the oil mist related parameters of thrust bearing oil tank. The results show that in the design of hydro-generator units, the number of pressure oil blade can be appropriately increased, so as to increase the temperature difference between the upper and lower blades, reduce the oil temperature below the blade and increase the pressure, so as to better inhibit the occurrence of internal oil rejection and effectively prevent the diffusion of oil mist due to low pressure.

The shaft system of hydropower unit has a significant impact on the stability of the unit. The degradation assessment of the shaft system can visually reflect the operating condition of the unit. This paper presents a method for assessing the degradation of the unit’s shaft system using instantaneous orbit feature image and conditional adversarial generative network (CGAN). Firstly, the vertical signals of each bearing were constructed as a complex signal, and the multivariate complex variational mode decomposition (MCVMD) method was used to process the signal and extract the instantaneous orbit features to construct the instantaneous orbit feature images. CGAN was used to construct the health model. The health model can fit the distribution of feature images in different operating conditions in healthy state and thus output health feature images. The healthy indicator was constructed using the differences between real and healthy images. The genetic algorithm was used to optimize the weights of multiple bearings in order to reduce the volatility of the comprehensive degradation curve in the healthy zone. The proposed method was tested on the unit's shaft system data and its validity has been proved.

To study the internal flow characteristics of bulb tubular turbine with different blade opening structures under large medium and small flow conditions, and to make theoretical support for the optimization of the flow passage components structure and the coordinated relationship, the research indexes were taken as three-dimensional streamline distribution, central meridian plane streamline distribution and central meridian plane pressure distribution. The results indicated that the maximum flow velocity existed at the runner of the bulb tubular turbine in the flow pattern analysis, and the flow pattern was chaotic, which was the focus of efficiency optimization. The reason of vortex in the draft tube was analyzed, and it could be optimized by reducing the dead water area. The internal pressure showed an overall decline tendency from the inlet channel to the draft tube. The pressure value at the runner discharge cone was the smallest. The pressure in the draft tube gradually increased along the flow direction, and the highest pressure value was 82 464.15 Pa when the flow rate reached 124.92 m³/s. The calculation results can provide a reference basis for the modification and optimization of the bulb tubular turbine.