Journal of Hydraulic Engineering

We present a model for predicting the behavior of a buoyant discharge in a cross-flow. The model is developed through a hybrid of the length scale and integral approaches. We compare the predictions of the hybrid model with a range of experimental data. In addition we compare the predictions of our model with those from two existing models: the expert system CORMIX1 and the Lagrangian integral model JETLAG. Through these comparisons we are able to show that the hybrid model predicts, with reasonable accuracy, the behavior of effluent discharged at a variety of angles to the ambient current. The hybrid model retains the flexibility of a length-scale model enabling it to be developed more readily to deal with complex flow configurations, such as the staged merging of small groups of discharges. In addition because the model utilizes integral solutions it requires less empirical information to develop than an equivalent length-scale model, which relies solely on dimensional analysis. Through this study we also are able to identify several areas of single-discharge behavior, which are not well understood and require further investigation.

One of the limiting factors in determining design discharge for water wells is the entrance velocity into the well screen. Most of the well designs assume uniform flow along the well screen and do not consider the complex flow either along the screens or in the well proper. Several investigators have shown that the flow into well screens may be highly nonuniform, and designs based on uniform flow may be improper. Moreover, the traditional well design locates the pump intake above the well screen, which may maximize the nonuniformity of screen entrance velocity. This investigation proposes a finite-difference model that determines the fluid flows into, around, and in the well screen. Results show that locating the pump intake between 30 and 45% down from the screen top minimizes the differential entrance velocity and total drawdown.

A numerical flow model is developed for evaluating different fish bypass systems at the Wanapum Dam on the Columbia River, Wash. The model solves the three-dimensional, Reynolds-averaged Navier-Stokes equations combined with the k-ε turbulence equations. The governing equations are formulated in generalized curvilinear coordinates and the bed roughness is introduced through wall functions. The paper discusses the calibration of the numerical model and its application in the design of a bypass system at Wanapum Dam on the Columbia River, Wash. The numerical model is validated with data taken in a 1:16 scale hydraulic model of a powerhouse intake bay. The model is found to reproduce accurately all of the flow characteristics that are believed to be important for the evaluation of a fish diversion system.

The Hong Kong government has embarked on a project titled Demolition and Realignment of Agricultural Weirs in which polluted weirs located in major streams are being replaced by inflatable dams. The dams are capable of being deflated automatically prior to a flood. This characteristic is advantageous since Hong Kong is subject to flooding during severe rainstorms. The ease of inflation and deflation allows the dams to be routinely cleaned of the deposited silt and accumulated debris. Thus, they are of great benefit in the abatement of pollution.

Wind setup and wave run-up heights are important considerations in determining the total height of a river levee system. In this study, two generalized reliability models for a levee system are developed considering the inherent uncertainties associated with wind as well as hydrologic and hydraulic processes. The generalized models can be simplified to reliability models when only hydrologic and hydraulic uncertainties are incorporated. By using the concept of a wind-flood damage ratio, reliability values of a levee system can be obtained and discussed on the same economic basis. A higher ratio of wind-flood damage requires a higher total levee height to obtain the same reliability value as for a lower wind-flood damage ratio. The total reliability decreases with an increase in service period of the hydraulic structure and/or the wind-flood damage ratio if the total levee height remains the same. The models enable the performance of a levee system to be predicted in terms of reliability. When the total levee height is specified, different options exist for the flood levee height to protect flood overtopping and for the freeboard allowance to withstand wind-generated wave overtopping. The best option can be chosen on the basis of a maximum reliability value. An example application of the models shows how the flood levee height and freeboard allowance may be determined to yield maximum reliability.

A modified logarithmic law is derived to describe the velocity distribution in open-channel flow with an upward seepage. The derived function agrees well with the experimental measurements conducted in a laboratory flume. The roughness function included in the modified logarithmic law is found to be dependent on the boundary Reynolds number and the ratio of the seepage velocity to the shear velocity. With the integration of the modified logarithmic law, the bed-shear stress subjected to upward seepage can be computed based on the depth-averaged velocity, water depth, boundary roughness, and seepage velocity. Using this method, variations of the bed-shear stress over the seepage zone are studied. The results show that the bed-shear stress decreases rapidly from the leading section of the seepage zone and then increases gradually toward the downstream end of the seepage zone.

An understanding of the effect of rainfall on the characteristics of shallow flow is needed for the proper modeling of the physical processes related to erosion and sedimentation. In this study, the turbulence characteristics of shallow rain-impacted flow over a smooth bed were investigated using a fiber-optic laser Doppler velocimeter and a specially designed artificial raindrop simulator. An empirical formula was derived to predict the variation of momentum correction factor β with the parameters related to the flow and the rainfall characteristics. Statistical correlations were developed successfully to predict the distributions of both the longitudinal and the vertical turbulence intensities for rainfalls with different intensities and raindrop diameters. The statistical characteristics of the Reynolds stress under different rainfall conditions also were analyzed. Analysis of the energy spectrum of the longitudinal velocity component indicated that the energy for the large-scale eddies (low frequencies) was increased significantly because of the mixing effect of the raindrop impact for the region near the water surface.

The M₀ and M₄ tidal currents are important for tide-induced net transport of sediment. To develop a physical understanding of the processes responsible for generating M₀ and M₄, a second-order analytical solution for a damped cooscillating tide in a closed-end canal with a horizontal bottom is presented. In deriving the analytical solution, it is assumed that the governing shallow water equations are weakly nonlinear, allowing the use of a perturbation technique. For this, the friction term in the momentum equation is linearized by requiring the tidally and spatially averaged energy dissipation by M₂ to be the same for the nonlinear and linearized friction. To determine the accuracy, for selected canals, the analytical solution is compared with numerical solutions. For M₀ and M₄, numerical and analytical values are within 20%. The analytical solution is used to demonstrate the internal generation of M₀ and M₄ through the nonlinear terms in the equations.

Experiments have been done in order to determine the effects of particle size and bulk density on the erosion of quartz particles. Average particle sizes ranged from 5 to 1,350 μm, while bulk densities ranged from approximately 1.65 to 1.95 g/cm³. For the larger particles, the sediments behaved in a noncohesive manner, i.e., they consolidated rapidly and the surface eroded particle by particle. For the smaller particles, the sediments behaved in a cohesive manner, i.e., they consolidated relatively slowly and the surface eroded in chunks. Erosion rates were a very strong decreasing function of density for the finer particles and were essentially independent of density for the larger particles. For a particular bulk density and shear stress and as the particle size increased, the erosion rate increased rapidly for the smaller particles, reached a maximum, and then decreased rapidly for the larger particles. Critical stresses for erosion were strongly dependent on particle size and, for the smaller particles, were also strongly dependent on bulk density.