Journal of Hydraulic Engineering

An experimental study of turbulence and vertical mixing was carried out in the lower east Pearl River (Mississippi and Louisiana) during a high-discharge period in February and March of 2010. Direct estimates were made of the time-varying profile of turbulent Reynolds stresses (${\tau }_x/\rho =-\overline{u^{\prime }{w}^{\prime }}$) through the water column. From this, both the vertical eddy viscosity ($K_m$) and mean vertical mixing time were calculated, and the water column was shown to be vertically well mixed at all times. Segmenting the stresses into stresses due to river discharge and stresses due to tidal flows shows the coequal nature of each component and highlights the ways they vary in importance during different flow regimes. Two simple turbulence models are examined and compared to the measured mixing parameters.

Among the various applications of the shallow water equations (SWEs) is the simulation of gravity currents (GCs). The SWEs are used as an alternative to track GC motion without explicitly dealing with turbulent processes, and constitute an intermediate solution between simpler integral models and more comprehensive models based on the Navier-Stokes equations. While the SWE equations have been successfully applied to a number of problems involving the release of dense fluids into deep ambient conditions, a more complex application is the simulation of the lock-exchange problem. In this particular problem, the ambient velocity influences the velocity as well as the shape of the GC, especially in the initial slumping stage. Features resembling discontinuities between the two layers are generated, and one numerical solution strategy has been to explicitly track such discontinuities. This work presents a shock-capturing, two-layer SWE model to simulate lock-exchange flows and its discontinuities. The main contributions are a reformulated mathematical model that incorporates the upper layer effects to the GC flow as well as a more efficient numerical solution for the flow at the GC front. The resulting numerical model was implemented using the finite-volume method (FVM) with an approximate Riemann solver, and the results compare well to existing numerical models as well as experimental data collected during this investigation.

Flash floods in ephemeral rivers are characterized by pronounced runoff dynamics, rapidly rising hydrographs, and infiltration through permeable river beds. In numerous countries, this infiltration—which is commonly referred to as transmission loss—is artificially increased by means of dams to promote groundwater recharge of local aquifers. However, flow dynamics are significantly altered by dam operation, i.e., initially pronounced process dynamics are reduced and standing/receding wave effects may occur in downstream river sections. For the adequate portrayal of such flow processes, this paper develops an integrated modeling system for flow routing in ephemeral rivers with groundwater recharge dams. The proposed system is based on a process-oriented description of flow, infiltration, and reservoir evaporation and allows for a robust application under a limited data situation, as usually encountered in arid and semiarid regions. Particularly, the proposed framework accounts for (1) the considerable loss of mass and momentum from the weakly dynamic flow downstream of a dam, attributable to transmission losses; (2) the transient character of transmission losses, which are nonlinearly depending on time and changing channel flow conditions; and (3) circumvents any numerical inconveniences associated with the modeling of dam release flow over initially dry beds by employing an analytical solution procedure of the governing flow equations. Following a comprehensive sensitivity analysis, relevant process parameters are estimated and the modeling system is applied for Wadi Ma’awil, Northern Oman. The application demonstrates both the system’s accurateness and robustness for flash flood routing under transmission losses along the wadi, where a recharge dam causes strong flow retention. Therefore, the proposed modeling system can aid in deriving potential groundwater recharge rates, which is of high importance for a sound water resources assessment in the study area.

This paper studies soil-erosion problem around cracked sewer pipes, which is usually associated with groundwater infiltration in sewer systems. Four factors (soil-particle size, defect size, water head, and soil height) that are associated with the erosion process were investigated experimentally. Two experimental setups were used: a cylindrical bin with a circular outlet at the bottom center to simulate the erosion process and a rectangular box with the outlet on the side wall to show the internal erosion process. The quantitative influences of the four factors on the erosion-void shape, volume, and length were analyzed. It is found that the geometry of the erosion void is determined by water head and soil height. Relationships were developed for calculating the water-discharge rate, sand-discharge rate, and erosion-void diameter.

The authors develop a transient semianalytical solution for evaluating the hydraulic head distribution in a leaky aquifer under constant surface-ponding conditions while draining through a fully penetrating multisection screen of an injection well. The authors use integral transform techniques such as Laplace and finite Fourier cosine for solving the said problem analytically. The solution in the Laplace domain is inverted to the time domain numerically. The authors test the performance of the derived semianalytical solution by comparing its results with field data. Subsequently, they use the solution to evaluate the hydraulic head distribution within the flow medium for a drainage event. Analyses show that the hydraulic head distribution within the radius of partial screening influence ($R_{ps}$) in the aquifer is controlled by both the radial distance from the well and the monitoring depth but only by the radial distance beyond $R_{ps}$. Further, the $R_{ps}$ does not depend on the screen sections’ lengths and positions within the fully penetrating aquifer. In addition, the results of using a multisection screen are not comparable to the single-section screen (irrespective of its position within the aquifer), having the length equal to the sum of screened depths of the multisection screen. Finally, the authors propose to use the solution derived in the present study for computing hydraulic head distribution in the flow medium specifically within $R_{ps}$ (i.e., two-times the aquifer thickness).

A three-state continuous-time Markov chain is proposed to model the movement of sediment particles across three different states, i.e., bed material, bed load, and suspended load layer. The proposed model incorporates an exchange process between the bed load and the suspended load layer, in addition to the interface between the bed-load layer and the bed material, as normally considered in most of the existing sediment transport studies. In this study the mathematical treatment of the continuous-time Markov chain is clarified to stress the difference in definitions of the continuous-time and discrete-time Markov chains. With the employment of the continuous-time Markov model, the bed-load transport capacity in the long run and the time-varying transport rates can be derived. On the other hand, the three-state model can assist in distinguishing whether the flow is subject to significant suspended load or not. Furthermore, the distribution of particles in each of the three layers can be quantified. The proposed model is validated against natural river data. The comparison has shown a reasonably good agreement and thus the validity of the proposed model is confirmed.

We derive a simple theoretical model for estimating the change in the magnitude of the particle slip velocity and the deflection in the trajectory of a particle induced by the lift force acting on a freely rotating particle settling under gravity in a shear flow. The lift force on the particle in small, intermediate, and large particle Reynolds number, is studied. We have found that the magnitude of the slip velocity is always reduced by the lift force, implying reduced particle settling velocities. The reduction in the slip velocity and the angle of deflection are the greatest when the shear is the greatest and the particle Reynolds number is the smallest. However, these effects decrease as the particle Reynolds number becomes larger. The derived equations are applied to evaluate the effect of the lift force on particle transport within turbidity currents, on the entrainment of small particles embedded in the lower boundary of a shear flow, and on sediment-laden open-channel flow.

The two-dimensional (2D) depth-averaged river model Finite-Element Surface-Water Modeling System (FESWMS) was used to simulate the hydraulic conditions at a contracted bridge site. The site studied was the James River bridges near Mitchell, South Dakota. The parallel bridges are located in a crossing between the two bends of a meander. The floodplain alignment relative to the channel and skewed bridges produces complex 2D flow patterns that cannot be predicted accurately by using a one-dimensional (1D) river model. The 2D model was validated by using flow measurements collected by the USGS during three high-flow events with return periods ranging from 25 to 100 years. The validated model was used to examine the site characteristics that influence the concentrated flow on the right side of the main channel and the exchange of flow between the main channel and floodplains. The rating curves derived from the 2D model and the results of soil erosion tests were used to evaluate live-bed and clear-water contraction scour at the bridge site. The scour analysis was conducted by using the equations in Hydraulic Engineering Circular No. 18 (HEC-18) and a method that accounts for the soil erodibility by using a curve of measured erosion rate versus shear stress. The study found that channel meandering, the no-flow boundary condition imposed by the walls of the river valley and skewed roadway embankment, and the dense trees along the left bank are the three main factors that create the unique hydraulic conditions at the bridge site. It is shown that using a 2D flow model could improve the estimation of contraction scour by providing more accurate information on the hydraulic parameters. The predicted scour depth was very sensitive to the critical shear stress and slope of the curve of erosion rate versus shear stress. Therefore, design should incorporate uncertainty in soil properties. It is also shown that an unsteady-flow approach to scour would produce a more realistic curve of predicted scour depth versus time. However, the cumulative effects of multiple flood events must be evaluated if time-dependent scour is used in design.

The weighted usable area (WUA) is one of the popular metrics to quantify the stream physical habitat and the effects of instream structures. However, most habitat studies only considered existing or proposed channel geometries in their WUA computations and not how the channel will evolve over time. This study reports a numerical investigation for the effects of spur dikes on WUA along the Juanchiao levee, Touchien River, Taiwan. Seven virtual dikes were placed along the outer bend in the study reach and the Rhinogobius candidianus was selected for target species. The horizontal two-dimensional (2D) flow fields at various flows and the corresponding WUAs for R. candidianus were computed under four channel bathymetry configurations, including: (1) existing; (2) immediately after dikes are implemented; (3) after 10-year channel migration without dikes; and (4) 10 years after dikes are implemented. The WUA obtained from the first two bathymetry configurations aimed to investigate the short-term spur dikes effect, whereas the last two examined the long-term effect. The channel geometry after 10 years channel migration under the conditions of both with and without spur dikes were simulated through a 2D mobile bed model. The analyses of short-term effect showed the WUA of R. candidianus, shortly after the virtual dikes were placed, increases due to increased water depth near the dikes, especially in low-flow conditions. As for long-term effects, however, the WUA in the case with spur dikes are lower than the case without spur dikes, especially in relatively high-flow conditions. The adverse effects of spur dikes on WUA appeared in the long-term investigation and resulted from the ever-increasing water depth and flow velocity around the tip of spur dikes and the main channel. This study showed that the habitat quality and availability might be overestimated if the channel bed changes over time. When using the WUA approach to evaluate the effects of instream structure on physical habitat prior to actually placing them in the river, quantitative predictions of the long-term variation of channel geometry and the associated physical habitat conditions are important.

Trajectory and centerline mean dilution data obtained from negatively buoyant jets, using a laser-induced fluorescence system, are presented at locations beyond the point at which the discharged fluid returns to its release height. Although this return point has been a focus of recent studies, site specific variations in source height and boundary angle can result in the flow impacting the boundary at locations that are distinct from this reference point. The new data are presented for source inclinations of 30, 45, and 60° above the horizontal. Previously developed semianalytical solutions are extended to include the aforementioned site-specific geometric parameters. The ability of these extended solutions and those of the CorJet model to predict the location and mean centerline dilution at various impact points are assessed for nominal boundary angles that range from 0–20° below the horizontal for trajectory and 0–15° for dilution. Comparisons with the experimental data indicate that the extended semianalytical solutions provide superior predictions under the prescribed conditions.

This paper reports a study of the turbulent wake of two-dimensional rectangular cylinders in an open channel. Three cylinders with different aspect ratios were mounted vertically at the midspan of an open channel and oriented through four angles of incidence. A particle image velocimetry technique was used to conduct detailed velocity measurements at various streamwise-spanwise planes around the cylinders. The results show that the head loss coefficients increase with aspect ratio and angle of incidence. Self-similarity in the mean velocities was achieved with the proper choice of velocity and length scales, irrespective of the angle of incidence and cylinder aspect ratio.

Affinity laws relate to the characteristics of pumps operating at different speeds, and in a water distribution context, are usually used to predict the pump curve of variable speed pumps (VSPs). VSPs can adjust the pump curve to meet the network requirements more efficiently with resultant savings of energy. The estimation of the effectiveness of a VSP is based on hydraulic simulations, in which the behavior of VSPs is described using the affinity laws. The affinity laws, however, contain approximations because they do not take into account factors that do not scale with velocity. In particular, the approximation inherent in the affinity law that computes power and efficiency can produce a misleading result, especially for small-size pumps. The research reported in this paper estimates the error in efficiency for a wide range of pump sizes and tests the use of a previously proposed formula as an alternative to the affinity law. Results show that a better estimation can be achieved for the efficiency of small- and medium-size pumps. Moreover the formula can be easily implemented in hydraulic solvers.

Experimental tests have been performed to observe and determine, in stationary flow conditions, the local head loss at the transition from a free-surface channel to a conduit. These investigations considered a wide range of discharge and varied dimensions and positions of a rectangular cross-sectional conduit connected to the downstream extremity of a rectangular free-surface channel. From the head-loss evaluation results, simple analytical expressions to predict the local head-loss coefficient value at the rectangular transition from a channel to a conduit are proposed and validated.