The risk assessment of collapse of tunnel involves many fuzzy factors and the evaluation method is complex, which makes the evaluation system constitute a gray system and the applicability of the evaluation method is not high. It is difficult to accurately assess the risk level of tunnel landslides. Using the basic idea of gray system and vector projection principle, a set of tunnel collapse risk evaluation model based on gray vector projection method is proposed. Based on the comprehensive analysis of the influencing factors of tunnel collapse under complex stratigraphic conditions, the group analytic hierarchy process (GAHP), which takes into account the degree of difference in the views of decision makers, is used to reduce the dimensionality of the pre-selected evaluation index system. Then, according to the comprehensive decision weight coefficient, 13 preferred evaluation indicators that can fully reflect the different, representative and highly differentiated evaluation indicators of the collapse mechanism are selected to establish a tunnel collapse risk optimal evaluation index system. In addition, grey group clustering (GGC) and anti-entropy weight (AEW) are introduced to determine subjective and objective single-level sorting weights, respectively. The improved combination weighting method of game theory (ICWGT) is used to obtain the optimal comprehensive weight distribution coefficient to determine the total ranking weight of the indicator. Finally, according to this model, the risk of collapse of five groups of tunnels was evaluated by the model. The results show that the risk assessment level of tunnel collapse is highly consistent with the actual excavation. The reliability and accuracy of the model are verified.

In view of the differences between national, continental and global scale precipitation data sets, there are many studies, but the reliability of data sets on river basin scale is rarely considered. To this end, the maximum/minimum difference percentage method and TFPM-MK trend analysis method were used to study and evaluate the differences between 12 widely used precipitation data sets in precipitation estimation and change trends in 6 292 basins around the world. The results show that there are large differences between precipitation data sets. There are 2 469 basins with an annual average precipitation difference of more than 500 mm/a, covering an area of 4 419.6×10⁴km², mainly distributed in Greenland, Africa, Oceania and West Asia. The existing data products cannot reach an agreement on the precipitation trend of all basins, and it is difficult to provide a reasonable scientific basis for the formulation and implementation of effective policies to ensure water conservancy and power generation, domestic water supply, flood control and early warning in the basin.

In order to improve the safety management level of sluice project construction and the traditional manual management mode, this paper proposed the automatic identification process of large-scale sluice project construction hazards based on the database and BIM technology. For solving the low efficiency of construction safety management, database and BIM technology were used to store the large-scale sluice project construction hazards sources. A identification system of large-scale sluice project construction hazard was developed by using computer Technology, which improves the efficiency of sluice project construction safety management and provides reference for information management of construction hazard sources.

To analyze the potential influence of the construction and operation of the cascade junction on the hydrologic regime of four major Chinese carps spawning sites in the downstream of the Xinjiang River, 21 parameters were selected to establish the hydrological index system of fish spawning sites. The one-dimensional hydrodynamic model was adopted to analyze the hydrological regime of Daxidu spawning site before and after the construction of the cascade junction. The range of variability approach was used to analyze the hydrological changes of spawning grounds during the breeding period before and after the construction of cascade junction. The results show that the flow during the breeding period decreases after the operation of the cascading junction, which made the variation range of water flow limited in a small range. Furthermore, the number of high flow pulse decreased, which reduced the stimulation signal to spawning of four large fishes. This study quantified the impact of the construction of cascade junction in the downstream of the Xinjiang River on the spawning grounds during the breeding period, and provided scientific guidance for the design of the local hub operation scheme.

During the landfall of Typhoon Billis in July 2006, Wujiang River, a tributary of the Beijiang River basin, was hit by an unusual rainstorm in history, resulting in the largest flood ever recorded in the basin. Based on precipitation record and Doppler weather radar data, it is found that the "train effect" induced by a convective storm, which originated in southern Hunan and crossed the divide, is the main reason for the extreme precipitation in Wujiang River basin. The study of the convective storm activity by using GIS technology shows that the complex topography of Wujiang River basin changed the direction of the convective storm several times, which greatly prolongs the duration of the storm in Lechang city, resulting in local rainstorm and flash flood. The study also found that the Huangcen Reservoir in Qitianling Mountain area was affected by typhoon, which probably enhanced local convections. The study reveals the influence of mesoscale and miso-scale topography on typhoon rainstorm, which contribute to the prevention of typhoon-induced disasters.

To maximize the cost-effectiveness of green infrastructure (GI) implementation, a pilot project in Guangzhou Tianhe Smart City was selected to conduct the study. Setting runoff volume reduction rate, suspended solids load reduction rate, and life cycle cost of GI as optimization objectives, SPEA2 was coupled with SWMM to obtain a series of GI optimal spatial allocation schemes with land utilization as the constraint condition. The results show that with the increase of the rainfall recurrence period, the integrated reduction rate and the range of the GI optimal spatial allocation scheme set gradually reduce. The sunken lawn (SL) is the most cost-effective GI while the green roof (GR) is the least. When the investment is low, the area of SL can be planned to the maximum before planning other GI. If the integrated reduction rate is high with sufficient funds, GR can be considered based on the large allocation area of other GI. The results can provide theoretical support for the implementation of the sponge city program.

In order to verify and optimize the “dragon-drop-tail” layout plan of the No. 1 flood discharge tunnel in the expansion project of the Chentianhe Reservoir, a numerical calculation model for the discharge was established. The numerical feedback of cavity backwater behind the airfoil flip bucket of indoor experiment was made use of to calculate the rationality of the numerical model. The optimal objective function of the shape of the airfoil flip bucket was identified based on the principle of safety and economy. The feedback model was then used to optimize the layout of aeration devices. The research shows that results of the simulation which adopts the FAVOR and Tru-VOF methods of structured rectangular grids, the RNG κ-ε turbulence model and the numerical model of fine grid are in line with indoor experiment results in terms of cavity shape and length. The optimal values of the bucket angle α and slope rate of the bottom slope behind the airfoil flip bucket θ of the No.1 flood discharge tunnel are 10% and 20%, respectively.

Due to the lack of measured temperature on the channel surface, it is difficult to accurately simulate the channel temperature field. Considering the effect of solar thermal radiation on the temperature of the shady slope, sunny slope and bottom slab of the channel, a calculation model of channel surface temperature increase was established. The finite element software ABAQUS was used to simulate the temperature field change process of the channel and analyze the frost heave of the channel lining under the condition of three times of cooling. The results show that under the influence of solar thermal radiation, the freezing depth of the shady slope of the channel is significantly greater than that of the sunny slope and the bottom slab, the calculated freezing depth and the damage law of the lining are also in good agreement with the actual situation. During the cooling process, the concrete lining of the shady slope is more prone to frost damage than that of the sunny slope and the bottom slab, the same rule applies when calculated according to the multi-year maximum freeze depth utilization specification. The research results can provide new ideas for the temperature control and crack prevention design of channel concrete lining.

In order to improve the water environment carrying capacity of Nanchang during the dry period, this paper constructed a water environment model at the lower end of the Ganjiang River and proposed an optimal scheduling scheme. The results show that the optimal scheduling water levels for the Ganjiang main river and its four downstream branches-the main, the north, the middle, and the south-were between the minimum ecological water level and the alert control water level, and that the scheduling time was within 1.11 days when the flow rate was less than 1 020 m³/s. Conversely, when the flow rate was greater than 1 020 m³/s, the alert control water level was the flood control limit water level. The optimal scheduling water levels of the main river and the four branches were between the minimum ecological water level and the flood control limit water level, and the scheduling time was within 4-7 d. It was suggested that each branch barrage hub should be operated close to the alert control water level to ensure that the water quality of the river control section meet the standard, and to maintained the maximum amount of water resources available in Nanchang.

In stone column composite foundations, the Priebe parameter composite method is commonly used to calculate the equivalent shear strength. Most of the existing studies are based on low stresses in conventional embankments, but few studies have been carried out on the applicability of this method under high stresses. This study focuses on the foundation treatment project for a clay core rock fill dam in Rwanda. The stress-strain behavior, micro-failure mechanism and equivalent shear strength of composite specimens with different area replacement ratio were investigated by simulating large-scale triaxle tests with discrete element numerical methods and compared with the widely used parametric composite method. The results show that the stress-strain characteristics of stone column composite foundations at high stresses differ from those at low stresses, and show a strain hardening trend. Composite specimens show a significant non-linear trend in strength at high stresses. The shear strength parameters calculated by the parameter composite method overestimate the strength of composite foundations, and it is recommended that they be discounted in engineering applications.