This article was contributed by Ying-Chieh Lin, Hervé Capart, and Der-Liang Young of Department of Civil Engineering and Hydrotech Research Institute/National Taiwan University in Taipei, Taiwan, the winner of the 2nd Flow Science 30th Anniversary Simulation Contest.
The typhoons Sinlaku (September 2008) and Morakot (August 2009) in Taiwan exposed significant vulnerabilities of the many bridges crossing Taiwan rivers. Observations of bridge scour failures at various sites indicate a number of features that are special to Taiwanese conditions and outside the range of typical conditions that have been examined in worldwide bridge scour research. Conditions specific to Taiwan include the pulsed nature of rainfall and discharge, rapid rates of erosion, joint bedrock and alluvial controls, and interference between different types of structures built along streams, such as weirs and bridges.
The Houfeng Bridge Failure in Taiwan
The Houfeng Bridge failure in September 2008 can be used to illustrate some specific issues of concern in Taiwan. Due to the tremendous change of the river sediment, the water surface approaches water supply pipeline. To protect the pipeline, the Taiwan Water Corporation built a concrete structure to cover the pipeline, which caused a sudden drop in water surface. In order to understand the detailed process, the present work seeks results by combining 3D numerical simulation and laboratory data. For local scour, three-dimensional modeling of local flow patterns has been done using FLOW-3D from Flow Science. To define realistic scenarios and check modeling outcomes, the data from numerical modeling is then compared with data from small-scale laboratory experiments (scale factor 1:200).
Three Dimensional Local Flow Modeling
The fully three-dimensional computational fluid dynamics model FLOW-3D is used for simulations. A three-dimensional simulation is necessary because of the strong vertical velocity component that exists at the end of water supply pipeline. Large vertical velocity variations make the flow patterns complicated and also enhance the scour before the bridge piers. One of the primary goals of this study is to show the influence of the local sill. To achieve this, a fine mesh (0.25 cm3) is adopted near the water supply pipeline and bridge piers. The total number of grids used in this model is about 700,000.
In the pure water simulations, the results from FLOW-3D show good agreement with small-scale laboratory experimental data. The enhanced water surface in front of the first pier shows slight variation in height as shown in Figure 3. The predicted data is similar to the measured data, and we can observe that the three experiments exhibited some variation in water surface height even if they had the same configuration. The non-bedrock simulation plays an important role in validating inflow and outflow boundary conditions and choosing an appropriate grid resolution for simulations, since this problem is simple and easy to simulate. In addition, the computed results predicted the maximum velocity magnitude after the drop water impacted the bed and this region could be considered a high erosion region.
Testing Numerical Modeling Approaches
The next step was to test numerical modeling approaches, and laboratory experiments with small-scale models will be performed. We envision setting up experimental analogues of bedrock-alluvial transitions and local bridge scour configurations, and acquiring measurements using optical and acoustical imaging techniques. For illustration, we present below some results from preliminary experiments and FLOW-3D modeling designed to simulate the Houfeng Bridge collapse (see Figure 4). Figure 5 shows the distribution of the bedrock and the color contour presented the packed sediment height average rate of change. From these results, the model shows that the present setting of sediment scour model similar with the experimental results, and we could get some information about the sediment erosion due to a sudden change in the height of the river bed.
FLOW-3D Simulation Results
The simulation results clearly show how the local structure (water supply pipeline) would affect the flow of water and the erosion of bedrock. Also, the numerical model predicted the velocities of flow field, water surface elevation and sediment height of change. The model could also be used to simulate different shapes and sizes of the local structures in alluvial river. This information will help us to understand how the variations in local region of river would damage the bridge piers, weirs and river course.
Figure 5. The packed sediment surface and the color contours present the packed sediment height average rate of change. (a)T=10 sec.; (b)T=20 sec.; (c)T=40 sec.; (d)T=80 sec.