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.

This paper presents experimental and numerical results of air movement in sewer pipes. Both the airflow in a straight pipe and pipes with changing slopes were studied. The result from the straight-pipe model suggests that the air pressure gradient is built up in the pipe even when it is kept at atmospheric pressure at both ends. The combined effects of water drag and the pressure gradient were analyzed. A general method for estimating the airflow rate was proposed. Further, physical and numerical experiments were performed on air movement induced by water flow with a hydraulic jump in pipes with a changing slope. The ratios of air and water flow rates were found to be substantially higher than published values because of different air transportation mechanisms. The rough surface and air/water bubbly flow in the roller of the hydraulic jump can affect the momentum flux of the air phase, and this additional momentum needs to be incorporated in airflow modeling.

This paper presents results of an experimental investigation into the time scale of local scour around subsea pipelines in current-only, wave-only, and combined wave and current flow conditions. To cover a range of geometric scales, the experiments were conducted in a small and a large recirculating (O-tube) flume with model pipelines ranging from 50 to 196 mm in diameter. Based on the experimental results and existing data available in the literature, it is shown that the time scale of the scour process is mainly dependent on the maximum instantaneous Shields parameter in all three flow conditions. Using this result, a new predictive formula for the time scale is presented, which is applicable universally to current-only, wave-only, and combined wave and current conditions. In addition to the time scale analysis, equilibrium scour depth measurements are presented for all three flow conditions. A new method of predicting the equilibrium scour depth in combined wave and current conditions is proposed based on the current component of equilibrium scour depth and the wave component of equilibrium scour depth. Collectively, the empirical formulas developed for time scale and equilibrium scour depth may be used to improve predictive engineering models of scour beneath subsea pipelines.

Surges caused by the sudden compression of entrapped air pockets are a source of serious issues in various hydraulic conveyance systems. Such surges may appear during rapid filling of storm-water sewers and storage tunnels, priming of water pipelines, and other applications when air is quickly displaced by inflows. When ventilation is inadequate during rapid filling conditions, sudden air compression can occur leading to pressures that may exceed design limits and cause structural damage. Past studies have provided various insights in air-water interactions in closed pipes during rapid filling conditions, but the focus of the majority of these studies was to support the development and improvement of numerical models. Fewer experimental studies focused on the effect of geometric characteristics such as pipe diameter and length/diameter ratios, as well as systematic investigation of various air pocket volumes and filling flow rates. The present work addresses this knowledge gap, measuring pressure surge peaks and the damping of pressure surges following the first peak. In this investigation, kinetic energy and linear momentum dimensionless groups were analyzed as potential alternatives to represent peak surges as as well surge damping. Findings in this study provide further insight into parameters governing air-water surge magnitudes, and on the peak surge attenuation following sudden compression.

Unstructured block ramps are fish-friendly structures to stabilize a riverbed, characterized by large macroroughness elements randomly placed on the riverbed. Physical model tests were conducted to investigate their stability and their behavior in case of overload. Different parameters describing the block and the base material were tested to find an optimal combination in terms of ramp stability with a maximum ramp slope typically in the range of 1–3%. A model for the determination of the ramp stability is presented, where the equilibrium slope of the ramp is related to a dimensionless specific discharge including information on the block size, the block placement density, and a characteristic grain size of the base material.

This paper presents a hybrid Euler-Lagrange model for bed-material load considering transition between the bed load and suspended load. The hydrodynamic model and suspended sediment transport model are solved in the Eulerian grid. A Lagrangian model integrating the near-bed grain trajectory and the momentum equations is employed to predict the motion of the bed load. Sediment exchange among stationary bed, bed load, and suspended load are modeled by considering the deterministic nature of the motion of individual bed load particles and the stochastic nature of the behavior of particle groups. The numerical model was applied to two types of laboratory experiments: (1) local scour upstream of a slit weir and (2) sediment release from a dam gate where an open channel flow and a closed conduit flow appear. All validation cases show that the model is able to reproduce temporal variation of the three-dimensional bed geometry around a weir-type structure with sufficient accuracy. Numerical results imply that computing the transition process from the bed load motion into suspension has a key role in three-dimensional simulation of nonequilibrium transport of bed load and suspended load upstream of a weir-type structure.

A new method for the approximation of the explicit probability of entrainment for individual coarse particles is presented. The method is based on the derivation of inertial acceleration measurements, space-state approximation of the dynamics close to entrainment, and the probabilistic approximation of the threshold inertial acceleration that causes incipient motion. Results from flume experiments with a custom-made inertial measurement unit enclosed in an idealized spherical enclosure, under varied flow conditions (achieved through slope change) and two different arrangements (saddle and grain-top positions) are presented to demonstrate the application of the method. The analysis supports the modification of the existing flow velocity–based entrainment criteria so they respect the particle-frame realization of forces during incipient motion.

Bridge-pier scour in alluvial rivers often occurs in live-bed conditions with dune migration. The scour depth attains its deepest value when the dune trough reaches the pier, followed by depth reduction when the crest arrives. Therefore, the instantaneous scour depth may be larger than its time-average counterpart. Consequently, accurate prediction of the instantaneous scour depth is pivotal to the safety of the bridge. Although published methods may be used to estimate the temporal variations of pier-scour depth, only few have focused on the live condition. This study presents a prediction method to simulate the temporal evolution of live-bed pier-scour depth. The erosion rate related to the combined effect of the local vortex system and sediment transport into the scour hole associated with the passage of dunes are examined to evaluate the pier-scouring rate. Temporal variations and computed equilibrium time-average scour depth are compared with published data with satisfactory agreement.

Experimental sediment transport and river morphologic studies in laboratory flumes can use two sediment-supply methods: an imposed feed at the upstream end, or a recirculation of sediment from the downstream end to the upstream end. These methods generally produce similar equilibrium bed morphology, but temporal evolution can differ. The adjustment of natural rivers may be reproduced by both modes. Nevertheless, computer models of river morphodynamics typically use a sediment-feed boundary condition, which can impact the simulated evolution of transient features such as bedforms. The effect of sediment transport boundary conditions on bedform dynamics was analyzed through numerical experiments using a two-dimensional, depth-averaged sediment transport model. Two different boundary conditions were imposed at the inlet (constant sediment feed and sediment recirculated from the outlet) for two bedform scales (dunes and bars). The type of sediment transport boundary condition greatly influenced dune development. The sediment-recirculating condition produced a more dynamic bed morphology with dunes of higher amplitude. The associated zones of higher shear stress had a direct impact on the hydrodynamics and patterns of sediment transport. In the case of the bar bed morphology, the simulated bars had similar mean length and height for both sediment boundary conditions. However, the sediment-recirculating case produced a more dynamic bed, in which the dominant bar length varied over time. Finally, the simulated bed morphology with bars, agreed much better with that observed when using a calibrated sediment transport equation to match sediment discharges instead of the standard empirical sediment transport equations available in literature.

The paper presents a novel contact image sensor (CIS) to monitor the local scour process around a model pile in water flows. The CIS is an optical sensor that tracks the evolution of the interface between sand and water based on the difference in the refelectivity of light from different media. A number of sensitivity tests are reported. It was found that the sensor reading is insensitive to the temperature in the range of 2–30°C and can provide reliable and continuous high accuracy measurements if suspended sediment concentration is lower than $40\mathrm{kg}/{\mathrm{m}}^3$ without needing precalibrations. Furthermore, the sensor was found to successfully detect local scour development around a vertical pile under steady current and tidal current. The CIS sensors have the potential to be applied to field monitoring of local scour around subsea infrastructure.

This paper introduces new formulas for the transportation of bed load that disputes Einstein’s use of his time factor, $t_E$, in calculating transport rate. His $t_E$ accounts for the falling velocity in still water of a given material, yet he applies this same $t_E$ in the context of a transport-rate problem. The authors introduce a new way of analyzing an old problem that everyone assumes needs no further development and as a result, respectfully disagree with Einstein’s use of $t_E$ within a transport-rate problem. This research presents an original formula that introduces a new time factor, $t_Z$, to be included instead. The major difference between $t_E$ and $t_Z$ is moving $v_f$ from the denominator to the numerator. The logic behind this is that the magnitude of a dense particle’s falling velocity is larger and, therefore, $t_E$ should be shorter in time. However, in the context of a transport-rate problem, it should take longer for a dense particle to be transported to its destination; thus, the introduction of $t_Z$ resulting in a longer time. With the new expression for the time factor, $t_Z$, as the settling velocity of bed-load material particles in still clear water divided by the constant of gravitational acceleration, Einstein’s bed-load transportation formula is then modified accordingly as the variables in the formula are redefined. Experimental data used by Einstein are reanalyzed, and the data clearly show that separate formulas for different specific gravity bed-load materials are needed. The modified formula for gravel grains is verified by Smart’s test data in steep channel (region of high intensity of transport) as is seen from the predicted curve running near the data points. When the bed-load transportation formulas in silty water and in differential temperature water are derived, the parameter related to the characteristics of the carrying medium is replaced accordingly. In both cases, the rate of bed-load transportation is found to be inversely proportional to the settling velocity of bed-load particles in the carrying medium. As the bed-load study involves three factors, namely, (1) characteristics of flow; (2) specific gravity of bed-load materials; and (3) characteristics of carrying medium, it can be seen that the formula derived considers these three factors in the formula modification. As Einstein is the only author in the field who brings the settling velocity of bed-load particles into the bed-load transportation study, this paper further addresses the importance of settling velocity and its role.

The objective of this paper is to develop a straightforward, robust, stable, and accurate mesh-free numerical technique for modeling the dynamic behavior of free surface, incompressible, multiphase granular flows. This method (henceforth, the multiphase MPS method) is based on the fully Lagrangian method of moving particle semi-implicit (MPS). MPS provides approximations to the strong form of PDEs on the basis of integral interpolants. The fluid is represented with particles, and the motion of each particle is calculated based on the interactions with neighboring particles by means of a kernel function. In the proposed multiphase MPS method, a single set of governing equations is solved on the whole computational domain. Density and viscosity smoothening schemes are applied to the phase interfaces to ensure shear stress continuity and stability of the pressure field. The multiphase MPS method is then applied to model deformable submarine landslide and dam-break over an erodible bed. It is assumed that the granular media is behaving as a non-Newtonian fluid. The simulated water and sediment surface profiles accurately match the available experimental data. This proves the applicability and accuracy of the proposed multiphase MPS method for modeling similar environmental granular flows.

Einstein’s integrals constitute one of the salient developments in theoretical sediment mechanics. An analysis of the accuracy and computational efficiency of proposed methods for the calculation of the Einstein’s integrals is presented. First, the accuracy of those techniques is determined using comparisons against highly accurate numerical results. For an infinite series solution, a study of accuracy versus number of terms in the partial sum is performed. Then, the central processing unit (CPU) times of the procedures are determined and compared over a full set of Rouse numbers and relative bedload-layer thicknesses. Finally, parallel versions of the methods are presented, and their parallel efficiency is assessed. Based on the criteria of accuracy, CPU time, and parallelization efficiency, it is concluded that the method by Guo and Julien, with modifications by Srivastava, is overall more efficient for implementation in sediment-transport codes.

The classic sidewall correction procedure, Vanoni and Brooks’ method, originates from Einstein–Johnson’s work. However, Johnson’s contribution is almost unseen in recent literature because his friction equation for the sidewall flow is difficult to solve. This note then revisits Einstein–Johnson’s sidewall correction procedure and solves Johnson’s friction equation explicitly and exactly. Besides, it extends Einstein–Johnson’s procedure to transitional-, and rough-sidewall turbulent flows. The presented exact procedure facilitates Einstein–Johnson’s sidewall correction for applications in flume experiments and sediment transport analysis in narrow canals, streams, and rivers.

Turbulence measurements over rough beds are used for a variety of purposes, including studies of habitat diversity for aquatic organisms, stream restoration efforts, and assessment of drag induced by vegetation. Turbulent kinetic energy (TKE) has been used to estimate bed shear stress by use of a proportionality constant of 0.19–0.21. The TKE method of bed shear stress estimation is less sensitive to flow misalignment relative to Reynolds stress or mean velocity measurements, but it has not been tested across a broad range of hydraulic conditions. In the work reported here, TKE was compared to bed shear stresses for gravel ($D_{50}=35\mathrm{mm}$) and cobble ($D_{50}=156\mathrm{mm}$) beds with varying discharges and levels of sand fill ($D_{50}=0.3\mathrm{mm}$). The ratio of bed shear stress (${\tau }_0$) with TKE was not strongly affected by mean velocity for either the gravel or cobble beds, but it was affected by bed roughness. The mean of the median values of ${\tau }_0/\mathrm{TKE}$ across three mean velocities was 0.24 for gravel with no sand in the bed and was reduced to 0.21 for sand fills of 62–216%. Sand fills of 0–39% for the cobble bed resulted in ${\tau }_0/\mathrm{TKE}=0.48$; increasing the sand fill to 46–65% produced ${\tau }_0/\mathrm{TKE}=0.40$. These results show that the roughness of the bed can affect the ratio of bed shear stress to TKE, which means that caution is required if TKE is to be used as a surrogate measurement for bed shear stress.

Laboratory experiments on rosette diffusers for dense concentrate disposal from seawater desalination plants are reported. The rosettes each had four nozzles pointing upwards at 60° and arranged 90° apart in planform. Single and multiple risers in two planform orientations in stationary environments were tested for various riser spacing. Three-dimensional laser-induced fluorescence (3DLIF) was used to map tracer concentration fields from which the dilution, length, and thickness at the end of the near field were obtained. The effect of riser spacing depends on the dimensionless parameter $s_r/dF$, where $s_r$ is the riser spacing, $d$ the nozzle diameter, and $F$ the jet densimetric Froude number. For $s_r/dF>\sim 2$, the results were independent of riser spacing but the near-field dilution was approximately 27% lower than it is for a single jet. For $s_r/dF<\sim 2$ the dilution, near-field length decrease, and empirical equations for their dependence on spacing were proposed. Dilution is less dependent on riser spacing than might be expected based on multiport diffuser results. This is because the jets from a rosette diffuser do not readily merge and block penetration of entrained diluting water, so the diffuser can be shorter for the same dilution.

In the present study, a new physically sound approach to experimentally obtain the contribution of each complex pier component to the ensemble scour depth is presented and illustrated. The new approach can be assumed to take into account the interactions of the different flow structures and their impact on the local scour depth. A total of 67 long-duration tests were performed to quantify those contributions according to the new approach, 35 of which were run in the scope of this study. Results showed that the contribution of each component is highly dependent on its position relative to the initial bed level and on the pier geometry. The experimental results obtained through the new approach were compared with the predictions of two widely accepted methods, showing that these tend to implicitly disregard interactions and joint effects of the different complex pier components with or on the flow structure.