Control Structures in Stratified Flows

Abstract

During the past decade the Tennessee Valley Authority has designed and built several structures for the purpose of withdrawing cold bottom water from thermally stratified reservoirs. The cold water is used to supply condenser water for steam-generated power plants. During the summer months the primary flows in the Tennessee River system are controlled by releases from upstream storage dams through low level turbine intakes. The cold water, discharged by the turbines, forms a density underflow in the downstream river and reservoirs which may be from 10 to 150F colder than the overlying surface water (Ref. 1, 2, 3). The intake structures in the form of submerged sluice gates are known as "skimmer walls". The water in the condenser water channel downstream of the gate is homogeneous and has a specific gravity equal to the lower, colder water in the intake channel upstream of the gate. The colder water is caused to flow through a vertical opening at the bottom of the gate by virtue of a head differential across the wall. The problem is to determine the maximum discharge of the colder water through the gate without inducing a steady state withdrawal from the warmer layer upstream of the gate. A basic experimental and analytical investigation of this problem was conducted in the Hydrodynamics Laboratory of the Department of Civil Engineering at the request of TVA in the spring and summer of 1954 (Ref. 4) as part of a continuing program of research in stratified flow (Ref. 5, 6). The flow configuration is shown schematically in Figure 1. While the information obtained from this study has proved to be a valid basis for design of skimmer walls of the type shown in Figure 1, questions have been raised in regard to the relative efficiency of other possible geometrical configurations. The proposed Bull Run steam power plant of the TVA is to be located at mile 48 on the Clinch River in the backwater of the Melton Hill Dam (under construction). After completion of Melton Hill, the normal depth of water in the Clinch River at this point will be approximately 20 feet. The Bull Run condenser water intake will be approximately 32 miles downstream from Norris Dam which is the source of cold water during the summer period of thermal stratification in Melton Hill reservoir. In the absence of the Bull Run plant the normal "plunge point" for the cold water in the reservoir would probably be in the vicinity of the Bull Run site. It is estimated that condenser water requirements will cause diversion of most of the river flow for sizeable periods of time. The heated water is to be returned to the river approximately one mile below the intake structure. This addition of heat will result in a reinforcement of the reservoir stratification and will probably move the cold water "plunge point" upstream. Due to the topography, the maximum length of an intake structure is approximately 400 feet. In the horizontal intake skimmer wall the lip of the skimmer wall is essentially at the elevation of the river bed. This configuration requires the excavation of a bottom step of height (b) in order for the fluid to pass through the horizontal opening (a) and flow under the gate into the condenser water channel. In order to have an accurate comparison for the two skimmer wall configurations, the experiments on the two types of walls were conducted using the same quantitative basis for the determination of the discharge at incipient drawdown. The drawdown discharge criterion for the 1954 tests was essentially a visual one, hence.,the tests on the type I structure were repeated. In addition, it was desired to obtain quantitative information on the magnitude of the energy loss across the skimmer wall. The experiments were conducted in the M. I. T. Hydrodynamics Laboratory using salt water and fresh water to simulate the prototype density differences due to thermal effects. For laboratory purposes and reproducibility of results a sharp interface between the two layers is obtained. It is recognized that in the field such a sharp interface is not possible; however, equivalent interfacial heights may be determined by using the depth at which the vertical density gradient is a maximum.