Journal of Hydraulic Engineering

Forum papers are thought-provoking opinion pieces or essays founded in fact, sometimes containing speculation, on a civil engineering topic of general interest and relevance to the readership of the journal. The views expressed in this Forum article do not necessarily reflect the views of ASCE or the Editorial Board of the journal.

The present study consists of a new mathematical-numerical modeling formulation to simulate the spatial and temporal scour development downstream of a rigid bed for both a noncohesive sediment bed and a cohesive sediment mixture with relatively small percentage of cohesive material. Laboratory tests were conducted in a rectangular tilting flume having a recessed box filled with the selected bed sediments and placed downstream of a rigid rough bed. The scour pattern was accurately acquired with a three-dimensional laser scanner at various time instants. The numerical code was calibrated through the scour profile data obtained under steady-state flow condition and then validated on the basis of scour patterns acquired under both steady and unsteady flow conditions (symmetric and asymmetric hydrographs). Contrary to most previous studies conducted with an issuing jet, the present study’s experiments were performed under non–strictly uniform flow conditions. The numerical model utilized information concerning the erosive turbulent flow characteristics as well as physical and mechanical properties of the movable-bed materials. The mathematical framework of the numerical model was a second-order partial differential parabolic equation in which the form of the scoured bed was assumed as an unknown parameter. Instantaneous near-bed velocity data, acquired with an acoustic Doppler velocimeter, was used to modify the Einstein sediment transport method, leading to a significant improvement of the simulated longitudinal scour profiles. Simulated temporal evolution of the scour profiles was in good agreement with the measured ones for both cohesive and noncohesive bed materials under steady and unsteady flow conditions.

This paper presents experimental results on partial-breach dam-break flows. Three-dimensional (3D) flood waves resulting from an instantaneous breach of a dam are experimentally investigated by combining different measurement techniques. The dam-break flow is initiated by quickly lifting a gate separating a reservoir and a downstream dry basin. Three different initial upstream water depths are considered. The study utilizes the digital particle tracking velocimetry (DPTV) technique for measuring the free-surface velocity and water depth. Stereoscopic configuration of two high definition video cameras recording the movement of hollow plastic balls enabled measurement of the 3D velocity field and the water-surface elevation in the near field. The pressure head is measured by pressure transducers at various locations. Ultrasonic velocity profilers (UVP) are used to obtain velocity profiles in the emptying reservoir. A simple nondimensional relationship for the variation of water-surface elevation in the upstream reservoir is presented. The following observations are made: (1) the maximum value for the nondimensional speed of the front wave is close to 1.0; (2) the ratio of surface velocity front wave propagation speed in the lateral direction to that in the downstream direction is almost 1:3; (3) the water depth is greater than the pressure head near the gate and less than the pressure head downstream of the gate due to water surface curvature effect; and (4) normalized velocity profiles in the upstream reservoir under different initial heads show a satisfactory collapse.

In parallel to their most frequent application as an auxiliary dam spillway, piano key weirs (PKWs) are also installed on low-head barrages in combination with run-off-the-river power plants. Their efficient rating curve, advantageous behavior under driftwood blockage, and the absence of mechanical elements are optimal for the hydraulic conditions associated with these run-off-the-river projects. However, in the absence of technical protection measures, scouring of the riverbed can occur at the downstream foundation. The unhindered scour formation and corresponding ridge generation caused by a PKW are investigated in this paper. Physical model tests were conducted under various conditions, providing a general estimation of the equilibrium scour and ridge dimensions. The maximum scour-hole depth dominates the shape of the hole and of the ridge. It was further shown that the maximum scour slope corresponds to the angle of repose of the sediments on the riverbed. Finally, a comparison of the present PKW scour data with predictions provided in the literature revealed that PKW-generated scour might be considered as jet-induced scour.

In the paper, the stability of a plunge pool bottom under the impact of an impinging jet is theoretically analyzed, with reference to the mean characteristics of the flow field: pressure and velocity. In cases when the mean components are relevant in respect to the fluctuating ones, this analysis is exhausting. In other cases, a separate evaluation of the fluctuation effects in lining design is treated by means of experimental evidence. The mean dynamic pressure at the bottom depends strongly on the impingement angle, which assumes a relevant role in the design of floor protections. In plunge pools, which are confined upstream by the presence of the drop structure, the impingement angle is theoretically determined by mass balance and momentum conservation, resulting independent of the jet entrance angle at the plunge pool water surface. The theoretical results are compared with literature, experimental evidence, and numerical simulations. This highlights the capability of the proposed theoretical framework to correctly interpret the physical phenomena and to produce suitable engineering approaches in plunge pool bottom lining design.

At present, bridges are frequently built with geometries that consist of a column founded on a pile cap and supported by an array of piles, referred to as common complex piers or most often (as also in this paper), briefly, as complex piers. This paper addresses prediction of local scour depth at those complex piers. An experimental campaign of 48 long-duration tests was performed with seven complex pier models under steady clear-water flow conditions to quantify the influence of the relative column width, $D_c/D_{\mathrm{pc}}$ ($D_c$ = column width, $D_{\mathrm{pc}}$ = pile cap width), the relative pile cap thickness, $T/h$ ($T$ = pile cap thickness, $h$ = approach flow depth), and the pile-group configuration on the equilibrium scour depth, $d_{\mathrm{se}}$. Several tests were carried out for the possible range of variations in the pile cap position relatively to the initial bed level. The values of $d_{\mathrm{se}}$ were estimated by extrapolation of the experimental data series. The performance of three methods, i.e., Auckland, HEC-18, and Florida Department of Transportation (FDOT), to predict $d_{\mathrm{se}}$ is analyzed and discussed. Based on the experimental results of the present study and on the conceptual approaches of Auckland and FDOT methods, an alternative formulation for a predictor of $d_{\mathrm{se}}$ at (common) complex piers is suggested and validated.

Current hydraulic design/analysis literature for labyrinth weirs is limited to headwater ratio (total upstream head/height of weir) values less than 1.0. This study utilized physical and numerical models to expand the current hydraulic performance theory of headwater ratios up to 2.0 for labyrinth weirs. The physical and numerical data provide practical design guidance for high headwater ratios and indicate that the empirical discharge coefficient design curves developed by Crookston may be acceptable for headwater ratios up to 2.0 or more. This was accomplished by evaluating the effects of higher upstream heads for a labyrinth weir with a 15° sidewall angle. Physical modeling was conducted in a rectangular flume under controlled laboratory conditions using two geometrically similar labyrinth models. A total of 120 data points were collected over a range of upstream heads. In addition, a numerical model of one of the physical model geometries was conducted using commercially available computational fluid dynamic (CFD) software. Seven upstream heads were tested and compared with the experimental results; excellent agreement was found, particularly at higher heads. In addition, several turbulence closure methods were evaluated, showing that computed discharge was relatively independent of the turbulence model.

Diurnal dynamics such as stratification/destratification affects a shallow lake more than a deeper one. Surrounding topographic features such as forests and mountains affect a small lake more than a large one. This paper present a study of a small, shallow, sheltered, and weakly stratified lake in response to meteorological (MET) conditions and indirect effects of surrounding topographic features. Wind measurements at the lake shores indicated that spatial variations of wind speed and direction were substantial in nearshore areas with dense forest stands. Field measurements of velocity profiles and temperature spatial distributions were measured during the summer of 2004. A three-dimensional (3D) ocean model modified by a generic length-scale turbulence model was implemented to address the effects of spatially nonuniform wind caused by trees and weak thermal stratification on lake dynamics. Simulated vertical and horizontal distributions of mean velocities suggested that wind sheltering altered current patterns, particularly in sheltered areas, and weak stratification produced subsurface maximum velocities. Simulated water temperature profiles reproduced an observed diurnal thermal pattern. A modified stratification indicator is shown to yield reliable measure of thermal stratification without internal thermal structure; this indicator had a strong relationship with the 3D model output.

Theoretical methods have been developed to estimate the longitudinal dispersion coefficients ($k$) in natural rivers. The triple integral expression for longitudinal dispersion caused by the transverse velocity gradient and the solution for the depth-averaged streamwise velocity distribution in the transverse direction in a rectangular flume are used in this study. The longitudinal dispersion coefficient was calculated after changing the nonintegral formula for the velocity distribution into a trigonometric function series by Fourier transformation and then by substituting this series into the triple integral expression. A dimensionless formula for longitudinal dispersion in the flume was then obtained by regression analysis and was consistent with the experimental results obtained in previous studies and this study. By analyzing the measured dispersion coefficients of the natural rivers and the corresponding values obtained from the formula for a flume with the same hydraulic parameters, a formula was derived for the longitudinal dispersion coefficient of natural rivers. The formula provides a suitable method for estimating pollutant transport in natural rivers.

Due to interparticle forces between clay particles in fine-grained sediments, erodibility and transport mechanisms for fine sediments are different from those for coarse sediments, and the capability to predict the erosion resistance of fine sediments is still in question. In this study, sediment specimens with different kaolinite contents were prepared by mixing ground silica and Georgia kaolin with tap water. Geotechnical tests were carried out to obtain the sediment physical properties. The erosional and yield strengths of the specimens were determined through hydraulic flume experiments and rheometer tests, respectively. Relationships among the erosional and yield strengths and the sediment physical properties were developed. In dimensionless form, erosional and yield strengths of fine sediments are predicted by sediment properties including clay-size fraction and water content. In addition, a relationship is obtained to predict Shields’ parameter as a function of the corresponding dimensionless yield stress. This procedure can be used potentially in engineering applications such as estimating river bank stability and evaluating erosion risk of river beds in proximity to flow obstructions for fine sediments.

Many studies have aimed to characterize pressurized transient hydraulics. However, it remains difficult to assess the importance of dynamic effects in a robust manner, and modeling is further complicated by the tension between computational efficiency and physical accuracy. To address such challenges for incompressible flows, this article presents an adaptive modeling approach that combines a novel hybrid formulation, termed the hybrid global gradient algorithm (HGGA), with a variable time step (VTS). The HGGA combines the generalized and rigid water column global gradient algorithms, so it can adapt to inertially-dominated flows and those without such effects. Computational efficiency and physical accuracy are balanced by adjusting the formulation according to the simulated hydraulics. Three physically-based indicators are then introduced to characterize unsteady flow: these actively inform the HGGA of how to model a system. Two pipe networks are used to demonstrate the current work. The first illustrates the utility of the inertial indicators, and the second comprises an extended period simulation with the VTS scheme. Although more computationally intensive than conventional modeling, the methodology is shown to provide a better representation of dynamic hydraulics.

Strong free-surface vortices are employed extensively in the hydraulic engineering industry in areas such as flow regulation, energy dissipation, and energy generation. Despite their long history of use, the literature on strong free-surface vortices appears to lack detailed experimental investigations, particularly with regard to subcritical approach flows. This paper reports a comprehensive experimental program that was implemented on 12 scaled vortex chamber geometries to identify the key dependent hydraulic parameters. Two-dimensional (2D) laser particle tracking velocimetry (PTV) was employed to determine the field circulation, ${\mathrm{\Gamma }}_{\infty }$. It was found that the field circulation and, hence, the circulation number ($N_{\mathrm{\Gamma }}$) is strongly dependent on the approach flow geometry, which was characterized by a nondimensional approach flow factor, $y_{\alpha }$ comprising the approach flow, depth $h/d$, and geometric factor, $\alpha$. The discharge number ($N_Q$) varied inversely with the circulation number following relationships governed by two further empirical parameters: the constant ($k_{\alpha }$) and exponent ($n_{\alpha }$). Specific to each geometry, empirical models that related these terms to the approach flow geometry are presented. These findings collectively deliver an alternative simple model to determine the depth-discharge relationship in vortex flows. The values of the radial Reynolds (Rr) number and Weber number (W) in the experiments suggested that the model should be scalable according to the criteria of previous studies. This has been supported by a validation using two prototype systems reported in the literature producing errors of less than 15%. Finally, two new flow classes in describing vortex flows have been defined: transitionally subcritical, when $0.7<{\mathsf{F}}_i<1$; and unstably subcritical, when the nondimensional approach flow factor $\alpha <1$ for relatively large approach flow depths.

This experimental and analytical study investigates the double-averaged (DA) turbulent flow characteristics within an array of large gravel obstacles found atop a porous gravel bed. Analysis of the experimental data reveals that the DA streamwise velocity preserves the logarithmic law above the form-induced sublayer, while a linear law and a third-degree polynomial function apply within the form-induced and interfacial sublayers, respectively. The form-induced shear stress is 70% of the DA Reynolds shear stress (RSS) occurring at the virtual bed level. The DA turbulent kinetic energy (TKE) components, streamwise and vertical, attain their peak values at the obstacle crest level, while they diminish sharply below the virtual bed level. The fluxes of the TKE streamwise and vertical components, however, change their signs slightly below the crest level, indicating a changeover of the dominance of the bursting events. For the TKE budget, the TKE production, diffusion, and pressure energy diffusion rate terms attain their peak values at the crest level, while the TKE dissipation rate has its peak value at the virtual bed level. Third-order moments of velocity fluctuations follow the linear relationship, and their signs change slightly below the crest level. The quadrant analysis suggests that the sweep events are the governing mechanism at the near-bed flow region, while the ejection events become predominant with an increase in vertical distance. The quadrant plots of the form-induced velocity components display a pseudo-elliptical scatter within the interfacial sublayer and a small circular cluster above the form-induced sublayer.

The presence of aquatic vegetation changes the flow characteristics in a channel. The impact of the presence of submerged, flexible vegetation in a channel with an additional parameter such as downward seepage is analyzed in the present work. Two patterns, staggered and uniform, and two different vegetation spacing, 15 and 10 cm center–center, were adopted. Experiments were conducted for the case of no seepage, 10% seepage, and 15% seepage. Velocity profiles show the presence of an inflection point near the top of the vegetation where maximum Reynolds stress is achieved. The velocity profiles measured in the unobstructed region of uniform pattern are higher compared to the velocity measured in line with the vegetation stems. Downward seepage increases the velocity, Reynolds stress, and turbulent intensities in the average range of 10–15% as compared to the no-seepage case. Higher-order turbulent statistics such as moment analysis, quadrant analysis, and turbulent kinetic budget are derived through experimental measurements and approximation. It can be inferred from moment analysis results that the downward seepage increases the flux transport in the downward direction and diffusion in the streamwise direction, which is shown by the governance of sweep event over ejection event from quadrant analysis. Quadrant analysis shows that sweep and ejection are the only two phenomena that contribute most to Reynolds stress. The sweep event is more pronounced for staggered as compared to uniform patterns where the measurement location for the uniform pattern is located in the unobstructed region. Turbulent Kinetic Energy (TKE) budget is also evaluated. Turbulent diffusion or transport is one of the essential components of TKE budget. Turbulent diffusion transports energy near the vegetation edge towards to the free surface as well toward the vegetation zone. The transport towards the vegetation zone increases with the downward seepage while it decreases in the upper free surfaces.

A reliable depth-integrated model that is suited for capturing the complex flow found with an array of submerged boulders over a permeable rough bed can be an attractive tool for river management. This paper presents a new nonhydrostatic quasi-three-dimensional (3D) calculation method called the general bottom velocity computation of the fourth-degree polynomial equation for vertical distribution of velocity coupled with the dynamic wall law (GBVC4-DWL). The computational domain of the GBVC4-DWL method considers the presence of the boulders, a vortex layer on the bed surface, and a roughness layer under the bed surface. The novelty of this study is that it introduces a dynamic rough wall law for the bottom boundary condition that includes the continuity and momentum equations for vortex and roughness layers to evaluate the nonequilibrium flow induced by the boulders. Comparisons of the results of the present method with previous experimental results and results of previous calculation methods indicate that the present method exhibits considerable improvements in calculating the velocity distribution over a rough bed with isolated submerged boulders.

High-resolution measurements of streambank retreat (SBR) rates are important for engineering applications, such as infrastructure planning and stream restoration. Conventional methods of measuring SBR, such as erosion pins or total stations, lack sufficient point density to accurately capture the spatial variability of bank retreat over an entire streambank surface. This paper presents a newly developed terrestrial laser scanner (TLS) that is affordable, robust, and simple. The TLS was employed to profile a $\sim 9\text{-}\mathrm{m}$ wide portion of a meander bend on Clear Creek in Coralville, Iowa. Bank profiles were acquired in July 2012 and July 2013; the resulting subtraction yielded a mean SBR rate of $2.13\mathrm{m}/\mathrm{year}$, with a maximum of 3.7 m at the upper portion of the bank.

This paper advances prediction of the clogging profile using the apparent porosity concept. The clogging profile is the vertical distribution of fine sand particles within a gravel-bed substrate that is overlain by hydraulically rough open channel flow. The author introduces the apparent porosity to represent the conveyance of fluid momentum through porous media, which is therefore a function of both the media and fluid, i.e., an apparent property. The author parameterizes the apparent porosity considering that coherent structures inject fluid momentum and sand into a clean gravel bed and the flow becomes laminar at its limit when the gravel is fully packed with sand. The author couples the apparent porosity with the previously published momentum-impulse model to provide a process-based prediction of the clogging profile, and the results show good agreement with experimental data. Results analyzed in dimensionless form via the dimensionless clogging depth show that the gravel porosity and roughness Reynolds numbers are the primary factors controlling clogging, which agrees with empirical research. In application, researchers can integrate the clogging profile model with sediment continuity models for gravel-bed rivers.

As shown by the examples given in this paper, a commonly used, commercially available water distribution network modeling software may produce unrealistic results for flow control valves (FCVs) with a nonzero loss coefficient. This paper presents and discusses a simple technique to remedy this problem. According to this technique, an FCV with a nonzero loss coefficient $K$ is modeled as a composite link that consists of an FCV of the same setting with a zero loss coefficient and an artificial pipe of negligibly small major loss with a diameter equal to the FCV diameter and a minor loss coefficient equal to $K$. The proposed technique can be implemented using the discussed software without any modification of its hydraulic solver.