FLOOD06499 - Laserfiche WebLink

Home Browse Search

reservoir volume. The bottom profile and elevations were detennined from sediment investiaation cross sections. Con- ditions that existed in the reservoir during the last week of March, 1966 were used to provide the boundary conditions for the simulation. Water temperature profiles, water surface elevations, and flows into and out of the test reach were obtained from the TVA (1969) data. The reservoir was strati- fied and was approximately 108 m (353 ft.) deep'at the dam. Surface heat exchange, wind velocity, and tributary inflows were all assumed to be zero for the purposes of this investi- gation; a steady inflow and outflow of 140 ems (5000 cfs) was used. Di s cus s i on The-fiITIe--and effort necessary to describe the reservoir geometry for both the FDM and FEM models were comparable. To achieve calculated results at comparable locations in space, optional quadrilateral elements were used so that the finite element network (Fig. 10) was almost identical to FDM grid (not shown). It is recognized that this network does not exploit the capability of the FEM model to allow increased geometric resolution where desired, such as near the reservoir outflow point, but this simplification was useful for com- parison of results. The convergence of the FEM solution was noted to be somewhat more sensitive to the magnitude of the turbulent exchange coefficients than the FDM model. The ranges of values of the coefficients over which convergent solutions can be obtained for the two models have not yet been firmly estab- lished. Additional sensitivity investigations shall be under- taken at a later time. Ariathurai, et al (1977) examined similar equations and found that stability and convergence of the solution could be related not only to spatial and temporal step sizes but also to the Peclet number which is the ratio of convective transport to diffusive transport. The flow fields calculated with the FDM and FEM models are shown in Figs. 11 and 12 respectively. The vertical scale of Figs. 10-12 is exaggerated by a factor of 100. Coefficients used (refer to equations 4-7) were: Exx=24, Exz=4.8xlO-3, Ez~=240, Dx=23, Dz=9.3xlO-7 m2jsec (260,0.05, 2600, 250, 10-s ft jsec). Although the models have numel'OUS detailed differ- ences, particularly in the description of boundary conditions, the calculated flow fields are similar and reasonable. For the. test application, the reservoir was thermally stratified, with the incoming fluid cooler and more dense than the fluid in the surface layers. The stable density gradient in the region of the thermocline tends to inhibit vertical momentum and material transport, yet circulation appears in the upper layers. The circulation in the surface layers is driven by internal hori- zontal shearing between the cool water flowing toward the outlet and the warmer water above. A similar flow pattern is also observed in the bottom region below the main flow in the "