Journal of Hydraulic Engineering

River incision can play a role in societally relevant or human timescale, which requires careful engineering planning and design. Physically based incision models can help us perform the engineering computations to prevent erosion damage to engineering structures such as bridges and channel-bank protection works. This paper presents a field application of our previously published work, in which we developed a mechanistically based model of bedrock incision by plucking and macroabrasion (a process of fracturing of the bedrock into pluckable sizes mediated by particle impacts from bedload). The plucking–macroabrasion incision model is tested quantitatively with Unnamed Drainage #1 in Southern Indiana, which showed obvious channel incision over the past 19 years, destroying a highway bridge foundation. The field data of incision rate, hydraulic conditions, rainfall, grain size characteristics, and morphological conditions were collected in the field or obtained from previous studies and the National Oceanic and Atmospheric Administration (NOAA). The plucking–macroabrasion incision model successfully creates a concave-upward profile with knickpoints (sudden change in slope) similar to the field condition and quantitatively simulates and explains the recorded distance of knickpoint-migration and incision rates in the past 19 years. The model is input with sensible parameters estimated with some degree of certainty from the field data. Sensitivity analyses were also implemented for some parameters. Appropriate amended versions of the model are anticipated to be able to simulate quantitatively the evolution of other bedrock streams under various input or boundary conditions over human timescale.

River and open-channel flows present complex three-dimensional (3D) structures in the water column and on the surface as a result of the existence of secondary currents driven by either centrifugal force or turbulence. This paper experimentally investigates secondary currents in a straight, turbulent, rough-bed open-channel flow; the effect of different channel width to flow depth ratios (12.25, 15, and 20) on secondary currents; and surface boils associated with secondary currents at moderate Reynolds numbers. Nearly instantaneous profiles of three components of flow velocity and turbulence characteristics in the water column were measured by using an acoustic Doppler velocity profiler (ADVP). Simultaneously, large-scale particle image velocimetry was used to measure the water surface velocities and turbulence structures. Mean velocity patterns in the water undulate across the channel, indicating the presence of secondary currents in the long-term average flow structure. Secondary currents affect the distribution of bed shear stress, Reynolds stress, and turbulence intensities across the channel. The aspect ratio determines the number of secondary cells in the water column. A mean multiband undulating surface velocity pattern in the transverse direction correlates with the secondary cell distribution in the water column below. The instantaneous position of the upwelling and downwelling regions on the surface may deviate from their long-term mean position, indicating a meandering of the surface pattern. Vortex structures are detected from instantaneous surface velocity maps, and vortex boil lines are identified. Boil vortices mainly occur in upwelling areas with high vorticity. The simultaneous detailed velocity measurements in the water column and on the free surface have shown a good agreement between the secondary cell patterns obtained by the two methods.

This paper describes high-resolution in situ three-dimensional (3D) bed topography and 3D flow field measurements in cohesive, straight channel reaches and provides analysis of reciprocal relationships between flow structure and bed topography. Existence of secondary flows and their structures are shown to be associated with variety of bedforms. Classification of bed microtopography and identification of the major governing flow mechanisms are described for three different types of bed morphology: single trough, single ridge, and sequences of troughs and ridges. The maximum mainstream velocity component of secondary cells attains values as high as approximately 20% of the average mainstream velocity, and secondary cell width can reach $1/2$ bedform width. Analysis of lateral versus vertical fluctuations was performed, revealing a good correspondence ($r^2=0.67$) between cell form and temporal fluctuations of velocity attributes; with diminished size of the secondary flow cell, the number of flow reversals increases. For a channel with given cross-sectional geometry, discharge and mean velocity, the downflows are almost twice as large and have higher turbulent intensities than upflows. The three-dimensionality has vital implications for understanding the flow structure over channel beds, emphasizing the significance of the effect that secondary circulation in addition to lateral and vertical flow perturbations may have on the general flow field and associated bed topography.

This paper presents the results of a large-eddy simulation (LES) of turbulent flow over a channel bed artificially roughened by hemispheres. The Reynolds number of the flow based on the channel depth is 13,680 at a relatively low submergence of 3.42. First- and second-order statistics are compared with corresponding laboratory experiments to validate the LES. The effect of roughness heterogeneity on higher-order statistics is quantified and discussed. The contribution of the dominating turbulent events (i.e., sweeps, ejections) to the Reynolds stress and the anisotropy of turbulence are quantified. Visualizations of the complex three-dimensional turbulence structures reveal the occurrence of a number of different vortex types in the flow. The contribution of turbulence structures to the turbulent kinetic energy and their scaling is assessed through proper orthogonal decomposition.

An algorithm is developed for real-time estimation of the cross-sectional average velocity of a channel flow by using an upward-looking pulsed wave acoustic Doppler velocity meters (ADVM). The Velocity Contour Weighting Method (VCWM) is applicable to gradually varied flows in prismatic channels and requires little to no calibration. VCWM estimates the average velocity as a weighted average of ADVM bin velocities. Weights are based on the velocity distribution sampled by the ADVM. Collectively, the VCWM is able to adapt to a wide range of channel geometry and roughness features. Expressions for the velocity weights are developed by first applying a validated 3D computation fluid dynamics (CFD) channel flow model to a wide range of flow scenarios including differing channel geometries, discharge rates, depths, and boundary roughness. CFD simulation data are then reduced empirically with the aid of dimensional analysis to obtain the velocity weight equation. Special attention is given to the first weight accounting for near-wall velocity where the ADVM does not measure. Application of the method to a large rectangular flume shows that the VCWM predicts the average velocity with an uncertainty less than $\pm 5\%$ and that this uncertainty can be reduced by minimizing the buffer distance between the channel bottom and the first velocity measurement. In a companion paper, the performance of the VCWM is examined in irrigation canals with trapezoidal cross sections.

The Velocity Contour Weighting Method (VCWM) was developed in Part I to accurately estimate the cross-sectional average velocity of a prismatic channel flow using acoustic Doppler velocity meter (ADVM) measurements of centerline velocity. Here, the VCWM is validated by its successful application to 25 different concrete-lined trapezoidal channels used for irrigation water delivery. At each site, the cross-sectional distribution of velocity is measured by an acoustic Doppler velocimeter (ADV), which is moved horizontally and vertically through a sampling grid. Multiple tests at some sites led to a total of 51 sets of cross-sectional measurements. ADVM measurements are simulated by interpolating ADV measurements along a set of vertically aligned centerline coordinates typical of ADVM deployments. Subsequent application of the VCWM gives an estimate of the cross-sectional average velocity. Secondly, the velocity-area method is applied to the ADV data to directly measure the cross-sectional average velocity for comparison with the VCWM estimate. Based on this comparison, relative percent errors in the VCWM for all 51 tests were within $\pm 6.3\%$ using a probable surface roughness ($k_s$) for the finished concrete of 0.0006 m without calibration. A sensitivity analysis shows that a range of realistic roughness values for finished concrete can be used without degrading the accuracy of the cross-sectional average velocity predictions by more than an additional $\pm 1\%$. Hence, the method is relatively insensitive to poorly characterized roughness values.

Dealing with large deformation and fragmentation of geometries and interfaces (e.g., free surfaces), the regular mesh-based Eulerian methods, such as finite-element and finite-difference methods, have difficulties in fluid-flow modeling. Recently, studies have focused on a new generation of numerical methods called mesh-free particle (Lagrangian) methods. In this study, a mesh-free particle method based on the moving-particle semi-implicit (MPS) particle-interaction model has been developed for simulation of open-channel flow. The model is able to simulate viscous fluid flow with large deformation and fragmentation of free surface in practical fields. Moreover, the model is capable of modeling open-channel problems with both inflow and outflow and inconstant numbers of particles. The model has been validated and applied to some common sample problems. The results show the reasonable accuracy of the model. The final model is capable of modeling free-surface deformation and fragmentation as well as accurate calculation of velocities in open channels.

The overall dam-break phenomenon is currently well understood for reservoirs filled with water. However, many dams face serious problems of reservoir sedimentation. Thus, regarding risk management, the influence of a large amount of sediment in the reservoir on dam-break wave propagation is of great importance. Experimental investigations were performed to investigate the influence of the silting degree of reservoirs and grain size distribution on wave heights and velocities. Dam-break waves in silted-up reservoirs are the result of a two-phase flow formed by a layer of water over a layer of saturated sediments in which complex phenomena occur simultaneously. Sediments flow such as a viscous fluid from the dam breach until complete stoppage, and clear water flows over the sediments creating bed load transport. The final sediment deposition cone, the flow of the positive front, and maximum wave heights were studied. An empirical formula is proposed to describe the celerity of the positive wave front in the case of silted-up reservoirs.

Sediment transport in a turbulent channel flow over the sediment bed with a ripple structure is numerically simulated by means of a large eddy simulation. The filtered Navier-Stokes equations for the channel flow and the filtered advection-diffusion equation with a settling term for the suspended sediment are numerically solved, in which the unresolved subgrid-scale processes are modeled by the dynamic subgrid-scale model of Germano et al. The migration and deformation of the interface between the sediment bed and the fluid flow is captured by the level-set method. The sediment erosion is approached by means of three different pickup relations postulated by van Rijn, Einstein, and Yalin, respectively, partly modified by the authors. Generally, the sediment is entrained into the flow from locations where the shear stress exceeds a critical value—on the upstream slopes of ripple crests—and is advected downstream in suspension by the flow, until it settles again when the local flow condition cannot further transport it, e.g., on the lee sides of ripples. A global effect of these local processes is the migration of ripples. The numerical results on the fluid flow field and the sediment concentration distribution are discussed. The computed migration speed of the ripples, which is only a fraction of the free stream velocity, is compared with known experimental data and a good agreement is demonstrated.

Laboratory experiments have been conducted to understand the behavior of negatively buoyant cylindrical particles of density $\rho$, length $L$, diameter $d$, with and without curvature, freely falling in a fluid of density ${\rho }_w$ at Reynolds numbers based on $d$ of 200–6,000. The paper proposes a parameter based on the cylinder density ratio $S=\rho /{\rho }_w$ and aspect ratio $E=L/d$ that is able to predict the onset of different modes of secondary motion ranging from oscillations to tumbling. The same parameter can also predict the maximum amplitude of the oscillations of the cylinder, on the basis of comparing the magnitudes of the oscillation velocity with the fall velocity. Contrary to previous work that has treated the oscillations and drag coefficient dependence as independent phenomena, this paper argues that the secondary motion reduces the time-averaged projected surface area of the cylinder during its descent, leading to a lower observed drag coefficient $C_D$ computed using the nominal projected area of length times diameter, $Ld$. Curved cylinders adopt an average inclination to the horizontal and an oscillation pattern that depends on the curved particle’s arc angle and its specific gravity. The inclined particle has a smaller projected area than $Ld$, which, similarly to a straight cylindrical particle, leads to a reduced drag coefficient.

Near- and far-field oceanographic measurements of the plume from the Boston outfall were made over two days in April 2001. The results were consistent with the original physical model studies on which the diffuser design was based. New laboratory experiments were run to simulate the actual conditions at the time of the field tests using three-dimensional laser-induced florescence (3DLIF) to measure dilution. The 3DLIF results were very close to the field observations and provided considerable additional insight into the near-field mixing processes: the plumes from each port of the multiport risers merge, first with other ports on the same riser, and then with similar plumes from adjacent risers; mixing transverse to the current direction then gradually erases any vestiges of lateral concentration variations. The field results were consistent with predictions of the mathematical model NRFIELD. Measurements were also made in the far field on the second sampling day. For the observed travel times of up to 24 h, the increase in dilution was small and was due to horizontal diffusion only; vertical diffusion was negligible due to the stable oceanic density stratification. The results, and those of tests conducted in July 2001, indicate that the main features of the near-field processes of ocean wastewater outfalls can be closely predicted by small-scale physical model experiments and the empirical mathematical models derived from them. The results confirm that diffusers with widely spaced risers can effectively dilute and disperse the effluent.

Transverse mixing in an unregulated northern river in Canada was studied in different discharge conditions with and without ice cover. The distribution of cumulative discharge was constructed on the basis of river cross-sectional shapes, and a modified streamtube method was proposed to describe transverse mixing. The modified method only uses raw field data to calibrate the modeled concentration profiles and thus, can produce a reliable mixing coefficient even with relatively low-quality field data. The effects of river discharge and ice cover on the transverse mixing coefficient were examined in a fixed study reach. It was found that the transverse mixing coefficient increased approximately linearly with river discharge (from 84 to $960{\mathrm{m}}^3/\mathrm{s}$). The dimensionless transverse mixing coefficient, scaled with river shear velocity and river depth, appeared unaffected (within 14%) by the river discharge in both open-water and ice-covered conditions, and its value was 21% smaller in the ice-covered condition than in the open-water condition.

Experimental investigations to understand the mechanics of motion of large sandbags and to simulate their settling behaviors are reported in this paper along with the details of experimental setup and procedures. Two high-definition charge-coupled device (CCD) cameras recorded the motion of the sandbags from the side of a glass water tank and a digital particle tracking velocimetry (DPTV) technique was used to determine the settling velocity of the bags. The drag and added-mass coefficients are estimated for the sandbags and the values of the added-mass coefficient are compared to those for solid spheres. The unit weight adjustment and the drag coefficient adjustment approaches are investigated for simulating the settling behavior of permeable sandbags. The results show that the unit weight adjustment approach gives better results than the drag coefficient adjustment approach, if the reduction in the drag coefficient because of porosity is taken into consideration and the effect of permeability of the sandbag on the drag coefficient is small and neglected. An exponential equation to estimate the percentage increase of the unit weight for the permeable sandbags is presented.

Long spillways often include more than one chute aerator to assure an appropriate cavitation protection. The first aerator guarantees sufficient bottom air concentration along a limited streamwise distance attributed to deaeration. Further downstream, the air concentration may be insufficient to protect the chute from cavitation damage. There, a second aerator is required whose operation is affected by preaerated approach flow. The present investigation systematically model-tested typical chute aerators with various approach flow features including preaeration. The resulting air entrainment coefficient as the ratio of air discharge entrained into the flow by the water discharge remains practically unaffected. The jet length is also not affected. Downstream of these aerators, the streamwise development of the average and the bottom air concentrations was affected by preaeration: the average value increased, whereas the bottom value marginally decreased. The effect of preaeration was found to be only relevant for average approach flow concentrations exceeding some 20% of the corresponding uniform flow concentration.

This paper describes a laboratory testing program developed to assess the hydraulic properties of coarse rockfill by using a custom-built large-scale permeameter. Tests were performed by using samples 1.5 m in length and $0.7{\mathrm{m}}^3$ in volume. Knowing the hydraulic characteristics of coarse rockfill is important for assessing the safety of the structures under anticipated flow conditions in flow-through rockfill embankments. Flow in rockfill structures often departs from the laminar flow regime at typical operating flow conditions because of the characteristics of the rockfill materials. These characteristics include porosity, particle shape, particle size, roughness, and the tortuosity of the voids within the structure, which result in high velocities in large interconnected void spaces. For this reason, flow-through rockfill structures cannot be predicted by using Darcy’s law. The design and construction of a large-scale permeameter built at the University of Manitoba Hydraulic Research and Testing Facility is presented in detail. The experimental program, which used the large-scale permeameter, studied the nonlinear hydraulic characteristics of coarse rockfill materials for the coefficient in the power-law relationship between hydraulic gradient and bulk velocity. The results demonstrate that the large-scale permeameter successfully characterized the flow-through conditions of a variety of rockfill sizes and gradations under typical flow conditions. Results also allowed the determination of the coefficients to design flow-through rockfill dams for local rockfills.

This study proposes a method to alter secondary flow pattern in a rotating annulus flume and obtain a resulting uniform bottom shear stress distribution along the radial direction by changing the shape of the channel cross section. Two or three wall angles, except for the bottom wall angle, are simultaneously changed up to 25° for different cross-sectional shapes. The flow characteristics in the channel are analyzed using the computational fluid dynamics technique to find the most effective configuration of the walls, including the rotating top ring. The results show that the bottom shear stress decreases and becomes uniform as the outer wall and the top ring angles increase. In the case of the inner wall, the wall angle has little effect on the bottom shear stress. The stress distribution is most uniform when the top ring, the outer wall, and the inner wall angles are 25, 25, and 5°, respectively.

This paper develops the concept of a nonreflective (or semireflective) boundary condition using the combination of a remote sensor and a control system to modulate a relief valve. The essential idea is to sense the pressure change at a remote location and then to use the measured data to adjust the opening of an active control valve at the end of the line to eliminate or attenuate the wave reflections at the valve, thus controlling system transient pressures. This novel idea is shown here through numerical simulation to have considerable potential for transient protection. Using this model, wave reflections and resonance can be effectively eliminated for frictionless pipelines or initial no-flow conditions and can be better controlled in more realistic pipelines for a range of transient disturbances. In addition, the features of even-order harmonics and nonreflective boundary conditions during steady oscillation, obtained through time domain transient analysis, are verified by hydraulic impedance analysis in the frequency domain.

This work presents a new numerical-model approach to simulate abnormal operational conditions in water mains, including the transition between free-surface and pressurized flow-regimes and subatmospheric unsteady flows leading to cavitation. The method is based on the two-component pressure approach (TPA), which is modified to incorporate a variable acoustic wave-speed calculation that depends on the pipe elasticity properties and on estimated free-gas content. This modification represents an improvement over the original TPA method for water-main applications in that the celerity, originally treated as a simulation parameter, now varies with space and time as would be anticipated in actual conditions. The proposed method was tested in three different conditions that involved unsteady flows and vaporous cavitation, and its results were compared with the original TPA model, the method of characteristics discrete free-gas cavity model, and experimental data. The comparative analysis indicates that the modified TPA model be an interesting and useful alternative to simulate unsteady flows in water mains in cases in which flow-regime transition is anticipated.

Because longitudinal dispersion is becoming increasingly germane to the problem of accurately determining water quality, water-resource managers and engineers seeking to represent solute transport in drinking-water systems must be able to compute relevant dispersion coefficients. Accordingly, the present study was undertaken to develop and experimentally verify a modified advection-dispersion-reaction transport equation as well as the formulas used to calculate the axial dispersion coefficient. The analysis assumes laminar flows, constant mean velocities, and short travel times (dimensionless time, $T<0.01$). With regard to the modified transport equation, the dispersion term was assumed to be direction-dependent. Thus, two distinct dispersion rates (forward and backward) were recognized and quantified as opposed to the single value used in conventional dispersion models. With the dimensionless travel time taken to be the independent variable, the developed dispersion coefficients increased at about one-fourth of the growth rate exhibited by the conventional dispersion formula. The proposed scheme demonstrated a large improvement over the conventional formula when its performance was judged against experimental runs using various combinations of pipe lengths, tracer injections, mean flow velocities, and solute properties. This study’s findings should lead to accurate water-quality predictions and corresponding quantitative risk assessments, especially in pressure zones where low-speed flows prevail.

The randomness of transient events, and the variability of its associated dependencies, ensures that water hammer and surges in a pressurized pipe system are inherently stochastic. To improve reliability-based structural design, a stochastic transient model is developed for water conveyance systems in hydropower plants. The statistical characteristics of key factors in boundary conditions, initial states, and hydraulic system parameters are analyzed on the basis of a large record of observed data from hydro plants in China; the probability distributions of annual maximum water hammer pressures are then simulated by using a Monte Carlo method and verified with an analytical probabilistic model for a simplified pipe system. The key loading characteristics (annual occurrence, sustaining period, and probability distribution) are introduced and discussed. By using an example of penstock structural design, it is shown that the total water hammer pressure should be split into two individual random variable loads—the steady/static pressure and the water hammer pressure rise during transients—and that different partial load factors should be applied to individual loads to reflect specific physical and stochastic features. Particularly, the normative load (usually the unfavorable value at a 95-percentage level) for steady/static hydraulic pressure should be taken from the probability distribution of its maximum values over a pipe’s design life, whereas for the water hammer pressure rise, as the second variable load, the probability distribution of its annual maximum values determines its normative load.

The Nakdong River Estuary Barrage (NREB) prevents salt-water intrusion but causes sedimentation problems in the Lower Nakdong River in South Korea. Its mitigation requires mechanical dredging to maintain the flood conveyance capacity during typhoons. This analysis focuses on the possibility of replacing mechanical dredging with sediment flushing through gate operations changes at NREB. The new approach first defines sediment flushing curves as a function of river stage and discharge. The feasibility of flushing is then assessed from the comparison of the flushing curves with the flow duration curves. The detailed analysis of long-term simulations using a quasi-steady numerical model provides detailed simulation results. The model applications from 1998 to 2003 incorporate tidal effects at 15-min intervals and also include major floods caused by typhoons Rusa in 2002 and Maemi in 2003. Accordingly, about 54% of the mean annual dredging volume could be eliminated by sediment flushing at NREB. The model also quantified the flood stage differences for sediment flushing operations with and without dredging. The resulting stage difference at NREB during floods would be less than 30 cm.

It is proved on the basis of the axiomatic energy equation that the theoretical upper limit for the hydropower gained by a water wheel or turbine per unit width in a rectangular open channel cannot exceed $(2/5{)}^{5/2}\rho g^{3/2}{H}_{\mathrm{eff}}^{5/2}$ by any means in which $H_{\mathrm{eff}}$ = effective water head, which is the sum of specific energy $E_1=h_1+u_1^2/2g$ and the drop of ground level $\Delta z$; $\rho$ = density; and $g$ = gravity body force. As a dimensionless measure, a coefficient of performance or harvesting factor is introduced that is an analogue to the one introduced by Albert Betz for wind turbines.

Although most stepped spillway–design guidelines were developed for uniform step heights, a nonuniform stepped design might be a practical alternative in some cases. A physical study was conducted in a moderate slope-stepped chute (1V:2H) and five stepped configurations were tested for $0.7<d_c/h<1.9$. Detailed air-water flow measurements were performed for each configuration and the results were compared in terms of flow patterns, energy dissipation, and flow resistance. The basic findings showed minor differences between all configurations and indicated that the rate of energy dissipation was about the same for uniform and nonuniform stepped configurations. But the observations suggested that the nonuniform stepped configurations might induce some flow instabilities for smaller flow rates.