Journal of Hydraulic Engineering

This paper presents a relative standard uncertainty (RSU) model for stationary acoustic Doppler current profiler (ADCP) open-channel discharge measurement. The model is a combination of the Type A and Type B evaluation of uncertainty and complies with the latest methodology for evaluating uncertainty in measurements implemented by the International Organization for Standardization. The Type A uncertainty component in the model accounts for the site-specific measurement conditions such as ambient turbulence; ADCP pitch, roll, and heading variations/errors; and the system noise of the ADCP used at the site. It is determined by statistical analysis of ADCP subsection ensemble discharge data. The model provides an in situ tool for quality control of a stationary ADCP discharge measurement.

Permeable friction course (PFC) is a porous asphalt pavement placed on top of a regular impermeable roadway. Under small rainfall intensities, drainage is contained within the PFC layer; but under higher rainfall intensities, drainage occurs both within and on top of the porous pavement. A computer model—the permeable friction course drainage code (Perfcode)—is developed to study this two-dimensional unsteady drainage process. Given a hyetograph, geometric information regarding the roadway layout, and hydraulic properties of the PFC media, the model predicts the variation of water depth within and on top of the PFC layer through time. The porous layer is treated as an unconfined aquifer using Darcy’s law and the Dupuit-Forchheimer assumptions. Surface flow is modeled using the diffusion wave approximation to the Saint-Venant equations. A mass balance approach is used to couple surface and subsurface phases. Straight and curved roadway geometries are accommodated via a curvilinear grid. The model is validated using steady-state solutions that were obtained independently. Perfcode was applied to a field monitoring site near Austin, Texas, and hydrographs predicted by the model were consistent with field measurements. For a sample storm studied in detail, PFC reduced the duration of sheet flow conditions by 80%. In a second sample storm, PFC prevented sheet flow conditions completely. The model may be used to improve the drainage design of PFC roadways.

Pluvial flooding in urban areas drained by storm sewer networks is characterized by surcharge-induced inundation. Urban inundation models need to reproduce the complex interaction between the sewer flow and the surcharge-induced inundation to make reasonable predictions of the likely flood damage in urban areas. In the framework of the present work, the storm sewer model SWMM5 and a newly developed two-dimensional (2D) noninertia overland-flow model have been coupled to simulate the interaction between the sewer system and the urban floodplain. The solution of the 2D model is on the basis of an alternating direction implicit scheme that solves the 2D noninertia free-surface shallow-water equations. For accuracy reasons, the time step is limited and controlled by the use of iteration to home-in on an accurate solution at each sweep. The dynamic interaction between the two models is bidirectional, and the interacting discharges are calculated according to the water level differences between the flows in the sewer system network and aboveground flows. The paper presents details of the newly developed 2D model and describes the way in which the model was coupled with a one-dimensional (1D) sewer network model (SWMM5). The models were tested on one hypothetical case study and one real-life case study.

A new monitoring approach for detecting, locating, and quantifying structurally weak reaches of steel-lined pressure tunnels and shafts is presented. These reaches arise from local deterioration of the backfill concrete and the rock mass surrounding the liner. The change of wave speed generated by the weakening of the radial-liner supports creates reflection boundaries for the incident pressure waves. The monitoring approach is based on the generation of transient pressure with a steep wave front and the analysis of the reflected pressure signals using the fast Fourier transform and wavelet decomposition methods. Laboratory experiments have been carried out to validate the monitoring technique. The multilayer system (steel-concrete-rock) of the pressurized shafts and tunnels is modeled by a one-layer system of the test pipe. This latter was divided into several reaches having different wall stiffnesses. Different longitudinal placements of a steel, aluminum, and PVC pipe reach were tested to validate the identification method of the weak section.

A stable and well-balanced finite volume method was proposed to solve the shallow water equations coupled with a sediment transport equation. An approximate Riemann solver with a monotone upstream-centered schemes for conservation laws (MUSCL) reconstruction was used for the computation of the flux terms. A special but relatively simple treatment of source terms made it possible to ensure numerically and physically balanced computational results on very steep sloped beds. Various numerical simulations conducted on rigid beds have demonstrated that the proposed numerical scheme could provide stable computational results in the cases of flat and nonflat bottoms. From the comparisons of experiments on erodible beds in one- and two-dimensional channels, good agreement was obtained when proper parameters for turbulent flow and sediment transport were provided. Lastly, the effects of the parameters in relation to hydraulic characteristics included in the flow and the sediment transport models were investigated and briefly discussed.

A lock-exchange density current flowing in a 90° bending channel is investigated using dynamic large eddy simulation. In the middle of the current and far from the head and tail, this type of flow shows common characteristics with the previously studied steady-state submarine currents, e.g., superelevation of the flow near the outer bank of the bending section. The secondary flow direction has been a topic of interest among various researchers lately, and the seemingly contradictory results of these studies have demonstrated both senses of rotation near the channel bed at the bend apex. The simulations in this study show a secondary flow pattern toward the outer bank near the channel bed. A part of the current paper is devoted to discussion of this issue. On the other hand, all previous studies have been conducted in a steady-state setting, and so the flow properties in this work are altered relative to the previous ones because of the lock-exchange nature of the current. Secondary flow changes to different patterns near the tail of the current and is nearly absent in the head section of the flow. Also, the current density distribution does not have any directional preferences in the tail of the current and may oscillate between the outer and inner banks of the section due to the thinning of the density and reduction in the size of the downstream velocity. As the current enters the curve, wall turbulence streaky structures demonstrate an amplified turbulent intensity near the inner bank because of the increased magnitude of the tangential velocity. The tangential velocity decreases near the outside bank due to the presence of an inhibiting outer wall, whereas the reverse occurs near the inner bank of the section. The turbulence structures also show separation of flow parcels from the inner bank toward the outside of the channel section, while advecting the turbulence characteristics with themselves. This phenomenon occurs synchronously with the shifting of the maximum streamwise velocity from the inner bank toward the outer bank. Finally, the effect of the channel mean radius on the flow properties has been investigated. The current study is the first high-resolution numerical study of the density currents in bending channels, and it investigates the flow phenomena that have not been previously addressed.

This research compares the spatially distributed turbulent wake characteristics associated with a submerged bed-mounted cube at three flow stages. The experiments were conducted in a large laboratory flat-bed flume at high Reynolds numbers. The velocity gradient under each flow condition was kept near constant, which yielded novel data that isolated the effect of flow depth on turbulent flow structure in the wake. Spatially distributed instantaneous three-dimensional velocities were recorded using multiple acoustic Doppler velocimeters (ADVs). With decreasing relative depth, marked variations in the spatial distributions of mean and turbulence statistics were observed that have important implications for local sediment mobility and fish habitat.

A numerical model was used to better understand ice-related scour processes at a large lock and dam. Large scour holes in front of Melvin Price Lock and Dam, the largest navigation structure on the Mississippi River, present a unique opportunity to explore this effect because the scour has been so severe and well documented. Likely scenarios for river ice accumulation and passage were identified based on historical flows and observations. For all simulations, the ice accumulated upstream of the gates as observed in the field and the ice thicknesses were found to match well with the general shape of the scour surveys. Ice thicknesses as large as 5.9 m and under-ice depth-averaged water velocities as great as $6.6\mathrm{m}/\mathrm{s}$ were observed in the simulations. By using a shear stress criterion, the rock size necessary to prevent the scour exceeded the original rock size of 0.33 m for six of the seven cases. This study demonstrates that the presence of a localized thick ice accumulation increases the potential for bed shear and may cause movement of stone bed protection sized for open water conditions. Ice processes should therefore be considered in the design or rehabilitation of riprap protection at large lock and dam structures.

A modified expression to predict the boundary shear distribution in compound channels is derived and is found to provide significant improved results. The practical method to predict the stage-discharge relationship uses the one-dimensional (1D) approach by taking due care of the momentum transfer. The proposed approach is tested for its validity using available experimental data. Error statistics including standard error and coefficient of determination ( $R^2$ ) are applied to ascertain the effectiveness of the model. The method also gives satisfactory results when applied to two sets of natural river data.

A method is proposed to study water networks pipe friction parameters (PFPs) identification problem of water networks by solving a PFPs identification equation. The PFPs identification equation corresponding to an actual engineering problem usually has non-unique solutions. To express identification results, a method is proposed which considers the Moore-Penrose pseudo-inverse (MPPI) solution of the PFPs identification equation as the PFPs identification results. Because the PFPs identification equation is a nonlinear equation, the MPPI solution of the PFPs identification equation cannot be expressed directly. Then, a progressive linearization iteration methodology is presented for solving the PFPs identification equation. The identification method is applied in a water distribution network case study, and identification results illustrate that the method is efficient, practical, and reasonable.

This note presents the general features of velocity and variation in length scales of the transition expanding jump, free jump, and wall jet in the rectangular horizontal open channel. The experiments were performed for a range of Froude numbers from 2 to 6 and at an abrupt expansion ratio of 2. The results indicate that the postdepth $y_2$ required to form an expanding jump is distinctly smaller than that for the corresponding classical jumps. The expanding jump length was 1.25 times the corresponding free jump length. The velocity profiles for the transition jumps were similar to those for the wall jet. For the transitional expanding jumps, the boundary layer thickness $\delta$ , normalized by the supercritical flow depth $y_1$ , grows more rapidly relative to the free jumps. The normalized boundary layer thickness $\delta$ in terms of the length scale of $b$ equal to the flow depth $y$ , where the velocity $u$ is equal to half of the maximum velocity $u_m$ , against the longitudinal direction $x$ normalized by $y_1$ , which was stabilized at 0.31 for expanding jumps. The length scale $L$ , equal to $x$ where the maximum velocity is equal to half of the initial velocity $u_0$ , of the free jump was approximately 0.75 that of the corresponding transitional expanding jump. The growth rate of the length scale ( $b/y_1$ ) for the expanding jumps was approximately 2.4 times that of the corresponding wall jet and free jumps.