Analysis Of River Flood Control Safety And Levee Stability

The project of Bailuwanlianshui Bridge subordinate to Yiyang-Loudi expressway project that is the highway of economy in Hunan province. The bridge across Lianshui tributary, so the river and slope on both banks could encounter various unsafe factors with the changes of watercourse.We knew the project impact on security of flood control by calculating the pier falls into river effect on cross-section of river, the flood discharge section before and after the project, changes in river flow rate, etc. The project played less effect on the river’s flood discharge capacity after the completion of the bridge, both sides of the river dike, the flow of the water could erode the levee, so that there should be relevant repaired and maintained. The bridge located in a river where the watercourse is narrow and river channel is shallow. The water discharge capacity of the main river channel is limited, banked-up water level of bridge upstream is 0.100 m, scope of influence is 434 m, When the flood occured, it could impact on production and life of residents who living in the upstream of the river.Then, this thesis applied the Grey correlation analysis method to analyze the influence factors of levee stability. There were many factors of levee stability, including the topography, physical and mechanical properties of geotechnical, the action of water, earthquakes, and human factors and so on. When we analyzed the factors, we could conclude them as slope height, slope ratio, average unit weight, internal friction angle, cohesive force, water head, dike distance and load. The Grey relation sequence from high to low was: internal friction angle, cohesion, water head, dike distance, slope ratio, unit weight, slope height, load.Finally, we used the results of grey relational analysis, we considered the factors that ranked within the top six as input factors. We established the model of Artificial Neural Network, and trained the network by collecting the instances of project We used the trained network model to predict the stability of the levees to make sure that the results of prediction were consistent with the actual state. Thus, it could prove that the BP neural network that base on gray relational analysis was convenient, fast and accurate to predict levee slope stability. It avoided complex calculation process in the traditional calculation method, it improved the efficiency of appraising the levee stability, which has very high application value.