英文河流动力学水力学Project【文章】.doc

_Relevance between Stream Barb Length and Bank Erosion Analysis Using Delft-3D Flow Model at Sawmill CreekAbstractThe Sawmill Creek is a small creak which is located in Ottawa, Canada. It generates from the south of Lester Road, flowing through several primary residential area and finally drains into Rideau River. There is a wide range of discharge and water level fluctuation taking place in sawmill creek. And the watershed of Sawmill Creek responds quickly to big rainfall events across the year. Several bends in Sawmill have witnessed notable band erosion and aquatic habitats reduction in the recent years due to high flow rate and distinct secondary flow which is towards the outer bank near the water surface and towards the inner bank near the bed at the bends. Several barbs have been constructed in two primary bends in October, 2009 and the result and following measurement show improvements in the band erosion and aquatic habitats problems. The objective of this paper is trying to study the relevance between the length of barbs and band erosion reduction effect to find the proper design plan for stream barbs and obtain the best effect of band erosion reduction and aquatic habitats protection. The three dimensional hydraulic software, Delft-3D, is applied to study the barbs effect in this paper.1. INTRODUCTION1.1. Introduction to Sawmill CreekAs It is shown in Figure1. Sawmill Creek begins in a wetland south of Lester Road, Ottawa, flows towards west and then towards north through South Keys and Figure 1. The Location and Flow Path of Sawmill Creek(Sawmill Creek 2014 Summary Report, 2014)Heron Park and finally flows into the Rideau River near the intersection of Bank Street and Riverside Drive. The total watershed of Sawmill Creek is 27.73km2 and the total length is approximately 11km. The surficial geology condition has a high degree of diversification, consisting of 40% sand, 29% clay 12% gravel, 9% diamicton, 6% organic deposits and 4% Paleozoic bedrock. The surficial geology in the watershed area is mainly marine clay plains with sand and rock ridges (RVCA,2008).The natural watercourse accounts for 41% of the total water course and the rest section is channelized. According to Sawmill Creek 2014 Summary Report, there were 26 aquatic species , including 4 game fish species, observed in 2014 and It is shown in Table 1. Table 1. Fish Species Observed in Sawmill Creek in 2004(Sawmill Creek 2014 Summary Report, 2014)Figure 2. The Fluctuation of Discharge and Water Surface Elevation(WSE) in One Year for 2009 (E. C. Jamieson et al.,2013)Although Sawmill Creek is one of the last free-flowing cool water streams left in the urban core of the City of Ottawa, the lower and middles reaches of the creek are highly urbanized and the creek corridor is degraded and confined by development and transportation infrastructure (RVCA 2012).The land use in Sawmill Creek watershed includes 48% urban/rural, 16%wooded area, 12% transportation area, 11% wetland, 1% agriculture, 1% water body and 11% unclassified area. 54% of the land use was made up of residential, industrial or commercial, infrastructure and recreation. The high degree of the urbanization and quick respond to rainfall in heavy rainfall events directly causes wide range of water level fluctuation in Sawmill Creek watershed. Figure 2 shows the discharge and water surface elevation fluctuation condition across one year in Sawmill Creek.1.2 Bank Erosion in Sawmill CreekBank Erosion is a kind of flow-bank interaction change process due to strongly secondary flow when water flows in the bend. As flow approaches the bend, the flow rate near the outer band is faster than that near the inner bank, which causes the low cohesive material at the surface of the outer bank eroded and gradual sedimentation at the inner bank. Figure 3. Erosion along Sawmill Creek(Sawmill Creek 2014 Summary Report, 2014)Excessive erosion and deposition of sediment within a stream have a detrimental effect on the bank stability, channel change and aquatic habitat. And poor bank stability caused by bank erosion can make the bank become easier to be eroded. With increase in runoff during rain season in Ottawa will adjust to accommodate the additional flow, increasing stream bank erosion. Accelerated stream bank erosion is part of the process as the stream seeks to reestablish a stable size and pattern. Damaging or removing streamside vegetation to the point where it no longer provides for bank stability can cause a dramatic increase in bank erosion(Steam Notes, Volume 1 Number 2). The loss of bank vegetation is also a significant problem due to the interaction between bank failure and bank erosion, resulting in trees falling into the stream and the potential to impact in aquatic habitat and migration path. Figure 3 shows high levels of bank erosion were observed along many sections of Sawmill Creek downstream of Walkley Road. Figure 4 shows the image of bank erosion happening in one bend of Sawmill Creek.Figure 4. Erosion in One Bend of Sawmill Creek(Sawmill Creek 2014 Summary Report, 2014)1.3. Stream BarbsStream barb a type of groyne which has been used for centuries under a variety of purposes ranging from river training to stream bank protection. It is a kind of linear rock structure connected with the bank and stretching into the channel. It is typically anchored. It is one of the most reliable and economically attractive approaches for stabilizing eroding banks in incised channels. The general configuration of barbs is shown in Figure 5.As for the bend area of the small stream, 3 or 4 stream barbs are generally installed at the outer bank and parallel with each other with a spacing ranging from 1m to 3m, determined by specific condition. Barbs are generally constructed out of large rock riprap, between 500 and 600mm in diameter, with additional smaller riprap(d:50 230mm) along the bank side slope (50% above/below bank full) upstream of each barb to provide additional protection in the area.( E.C. Jamieson et al.,2009) Spurs were spaced at roughly twice the average baseflow channel width (< 7 m) with lengths roughly 40 percent of the average width. Crests were level, 2 m wide, and 1 m above the bed, or about 60 cm above baseflow water surface elevation. Stone size ranged from 0.2 to 450 kg, with 50 to 85 percent of the stones weighing less than 36 kg (F D. Shields et al.,1998).Figure 5. Conceptual Drawing of Spurs (darker rock)(F D. Shields et al.,1998)This configuration redirects the attacking current and the primary secondary flow cell away from the outer bank towards the centre of the channel (Minor et al., 2007, Jamieson et al. accepted). Moreover, unlike other bank protection measures (e.g. riprap, concrete paving and gabion walls), stream barbs require less material and can promote vegetated stream banks, maintain deep pool habitat through the development of scour holes at barb tips and increase aquatic species diversity (Shields et al., 1998; Engelhard et al., 2004). However, Stream barbs do not address bank failure due to soil instability andDrawdown (Technical Note 23: Design of Stream Barbs, U.S, 2005).Therefore, additional work should be done before the barbs installation such as stream clean up, vegetative planting and site monitoring. The plan is to incorporate vegetative planting and other bioengineering practices during the construction phase to address these additional mechanisms of bank failure( E.C. Jamieson et al.,2009). Moreover, vegetation provides additional roughness to dissipate energy along the stream bank and enhances wildlife habitat and water quality( E.C. Jamieson et al.,2009). 1.4 Target Reach InformationThe target reach of this study is a 50-m reach in Sawmill Creek, which generally consists of two small bends and receives approximately 90% of the contributing watershed area. The reach is experiencing bank erosion and mass wasting at two consecutive bends. The linear distance and area of eroding bank in the first (second) bend were 13.4m and 41.7m2 (28.0 m and 46.0 m2)(E. C. Jamieson et al., 2013).A number of large boulders dominated the morphology of the second (downstream) bend, and a riffle had formed close to the apex of the bend. (E. C. Jamieson et al.,2013) target reach was selected not only because the eroding banks indicated the presence of erosion processes but also because of relative ease of site access and lack of proximity to critical infrastructure. (E. C. Jamieson et al.,2013)The site image of the target reach and two bends are presented in figure 2.Figure 6. (A) Aerial photo of the Sawmill Creek study area; red dashed lineindicates right bank and flow is from south to north. (B) Bend 1 lookingupstream, and (C) outside bank of bend 2 looking downstream.( E.C.Jamieson et al., 2009)In September 2009, a series of seven stream barbs were installed to protect the two eroding outer channel banks and to serve as a demonstration project for the use of these structures in a semi-alluvial channel( E.C. Jamieson et al., 2009). Three barbs were placed in the first upstream) bend, and four barbs were placed in the proceeding (downstream) bend (labeled B4B7) (Figures 1 and 2) ( E.C. Jamieson et al., 2009).Figure 7. Stream barbs at Sawmill Creek: (a) bend 1 and (b) bend 2.Barbs are numbered in the downstream direction. Photos taken on 9 November2009, during low flow conditions (discharge, Q 0.3m3/s). (E. C. Jamieson et al.,2013)2. OBJECTIVETable 2.Barb Installation Plan In the Two Bends of Sawmill Creek(E. C. Jamieson et al.,2009)According to the barb design plan in previous study (E. C. Jamieson et al.,2009), the barb installation angle was determined due to laboratory experiments (Matsuura, T et al., 2004 and numerical modeling (Minor, B. et al,. pp1087-1095). In these experiments, which considered 90 and 135º channel bends, optimum bank protection was achieved for a series of three barbs, each with an alignment angle of 30º (E. C. Jamieson et al.,2009).However, the length of stream barbs are determined due to the USDA guidelines recommendation value which says barb lengths should not exceed one-third of the cross section top width at bank full flow but does not specify the optimum value of barb length. The barb length design in the two bends of Sawmill Creek was not taken into account in detail. It can be concluded that there may be a better barb design plan with the optimum length value.The objective of this paper is trying to study the relevance between the length of barbs and band erosion reduction effect to find the proper design plan for stream barbs and obtain the best effect of band erosion reduction and aquatic habitats protection using Delft-3D hydraulic flow model.3. METHODDelft-3D hydraulic flow model (Delft-3D-FLOW) is a unique, fully integrated computer software suite for a multi-disciplinary approach and 3D computations for coastal, river and estuarine areas. It can carry out simulations of flows, sediment transports, waves, water quality, morphological developments and ecology. It has been designed for experts and non-experts alike(Delft3D-FLOW_User_Manual, Deltares System. The Delft-3D suite is composed of several modules, grouped around a mutual interface, while being capable to interact with one another (Delft3D-FLOW_User_Manual, Deltares System. It is a multi-dimensional (2D or 3D) hydrodynamic (and transport) simulation program which calculates non-steady flow and transport phenomena that result from tidal and meteorological forcing on a rectilinear or a curvilinear, boundary fitted grid(Delft3D-FLOW_User_Manual, Deltares System).The general process is to simulate several flow conditions with different value of barb length and analyze the velocity and shear stress profile to see the differences.3.1 Grid Generation The Delft-3D-RGFGRID interface is used to generate splines and grids based on the channel, which is the first step of the flow simulation. ABDCFigure 8. Grid Generation(A: Spline Generation; B: Change Spline into Grid; C: Orthogonalisation; D: Refinement .)The channel ichnography is obtained from the Geographic information System (GIS) which can be used to determine the land boundary in The Delft-3D-RGFGRID interface.After determining the land boundary, several splines can be drawn along the channel and perpendicular to the channel, roughly. The splines should be as smooth as possible and overlap with the channel boundary as much as possible when drawing the boundary splines.After splines are determined, grid can be generated from the splines. In order to make the grid cell uniform and smooth, orthogonalisation should be done several times until getting the optimum grid. The grid was refined two times to improve the resolution in these regions for better flow and sediment transport calculations. The first time of refinement sets the M refinement factor and N refinement factor equal to 3 and 3,and the second time equal to 2 and 2, respectively. Finally, a optimum grid with 74 points in M-direction and 14 points in N-direction is generated successfully. This process is shown in Figure 8.3.2 Bathymetry GenerationThe Delft-3D-QUICKIN interface is used to generate continuous bathymetry image based on the depth data and the gird generated previously, which is the second step of the flow simulation.BAFigure 9. Bathymetry Image(A: Colored view; B: 3D view)The depth data are obtained from field measurement in 36 cross sections in the two bends area and several thalweg lines on the river bed. After the gird and depth data input is done, the size of gird cell can be averaged using Grid Cell Averaging function. Then, the continuous bathymetry image can be displayed through triangular interpolation and internal diffusion. The internal diffusion is used to fill the blank depth data in the channel. This process is shown in Figure 93.3 Virtual Barbs Installation The Delft-3D-FLOW_INPUT interface is used to simulate of flow and record the data in observation points and cross sections based on the bathymetry image and the gird generated previously, which is the third step of the flow simulation.ParameterValue/SettingParameterValue/SettingSimulation Start Time13/04/2016 00:00Water Density1000 kg/m3Simulation Stop Time14/04/2016 01:00Temperature15 Duration15 minRoughness (Manning)0.05Time Step0.01 minHorizontal Eddy Viscosity2 m2/sTime Interval for Map Store120 minHorizontal Eddy Diffusivity10 m2/sHistory Interval for Map Store10 minVertical Eddy Viscosity2 m2/sWater Level1.0 mVertical Eddy Diffusivity10 m2/sSalinity30 pptOzmidov Length Scale11 m2/sUpstream Discharge2.5 m/s3Model for 3D turbulancek-EpsilonGravity9.81