Journal of Hydraulic Engineering

Vertical profile of longitudinal velocity in vegetated channels reflects complex mechanics of flow-vegetation interactions and determines the bulk flow velocity and flow rate. Most available models of velocity profiles in vegetated channels are based on a single physical concept that underpins theoretical considerations and data interpretation. However, measured velocity profiles suggest that the use of a single concept is not sufficient to cover all possible scenarios of flow-vegetation interactions. As a result, a number of models in which different concepts are applied to different flow regions have been recently developed. Within this framework, the overall velocity profile is represented with a set of linked segments. Although such segment-based models have improved velocity profile description, there is a need for more robust approaches and better analytical formulations. This paper proposes a new approach where a vertical velocity profile in vegetated channels is modelled as a linear superposition of four concepts: (1) uniform velocity distribution, (2) mixing layer analogy and a hyperbolic tangent profile, (3) boundary layer concept and a logarithmic profile, and (4) wake function concept. In contrast to the segment-based models, the proposed analytical expression combines these concepts simultaneously over the whole flow depth allowing significant overlaps of the momentum transport and turbulence production mechanisms. The model is tested using extensive laboratory experiments.

Partial blockages commonly exist in pressurized pipeline systems, and the rapid remediation of such faults is required to reduce the wastage of energy as well as to maintain the serviceability of the pipe network. Numerous transient test–based techniques (TTBT) have been developed for detecting pipe defects, with each technique providing different advantages. Two previously developed techniques—pressure signal analysis (PSA) and frequency response analysis (FRA)—are experimentally tested in this study on systems of different pipe material and characteristics. Each method is validated using the experimental data, and the results show that PSA is most accurate for locating the blockage while FRA is most accurate for determining the radial constriction and length of the blockage section. To take advantage of the different strengths of the techniques, a coupling of the two methods is proposed. Experimental application results reveal that both detection accuracy and calculation efficiency are improved when the coupled method is used instead of the original techniques applied individually. Moreover, the results indicate that current TTBTs for partial blockage detection are more reliable in elastic (metallic) pipelines compared to viscoelastic (plastic) pipes.

The phenomenon of white mist in a rapidly filling pipeline containing an entrapped air pocket is numerically and experimentally investigated. The air-water flow patterns, pressure, and temperature histories are synchronously recorded to illustrate their interrelations. The white mist phenomenon is particularly observed during fast transients, especially during the first compression of the air pocket. Comparisons between calculations and experiments indicate that the white mist primarily reflects a condensation process. More specifically, the air temperature increases because of rapid compression of an entrapped air pocket, and the high temperature could cause water to adhere to vapor at the pipe surface. For fast transients, the first compression causes a near-adiabatic air compression, but heat exchange effects become more significant in the subsequent compression and expansion cycles. As the initial air length decreases, the maximum pressure first increases and then declines, with the most dangerous air length occurring when about 3.4% is initially occupied by air. The ratio of the maximum pressure to the driving pressure increases approximately linearly with respect to the upstream pressure. A local-interpolation elastic-water model is developed by considering air-temperature change and its validity is confirmed by comparing the model and experimental results.

This paper presents a new, tractable analytical expression to describe the intrusion of fluids into buried pipes under steady-state conditions. The expression is validated with results from novel experiments. The derivation is based on the combination of the relevant existing models of flows through porous media and the losses through an orifice, with the resulting expression relating the intrusion flow rate to an applied driving pressure. The expression is shown to yield results directly equivalent to those generated from a full three-dimensional (3D) computational fluid dynamics (CFD) model of the intrusion process. Results from the experiments, quantifying volumetric intrusion from a realistic 3D porous media, presented here, compare favorably with calculated values, validating the expression. Although the experimental and analytical results show a high level of agreement, it was found that the analytical expression tends to slightly underestimate the intrusion rate seen experimentally. The absolute difference in the values is low and is thought to be attributed to preferential flow path at the porous media and pipe interface that the analytical expression and CFD model do not include. It is shown mathematically and verified experimentally that the viscous and inertial resistance to flow in the porous media reduces the intrusion (or leakage) flow over that predicted by the standard orifice equation and places additional dependencies of the flow on the size of the intrusion orifice. The values obtained from the expression should be considered as a lower bound to intrusion (and leakage) rates, with upper bounds being provided by the standard orifice equation. Although developed to aid in the quantification of intrusion risk, such as that associated with water distribution systems, the expression is also validated for leakage for the limited case that the external porous media is considered to be fully compacted, consolidated, and immobile.

Streamflow measurement techniques of the Mississippi River at St. Louis have changed through time (1866–present). In addition to different methods used for discrete streamflow measurements, the density and range of discrete measurements used to define the rating curve (stage versus streamflow) have also changed. Several authors have utilized published water surface elevation (stage) and streamflow data to assess changes in the rating curve, which may be attributed to be caused by flood control and/or navigation structures. The purpose of this paper is to provide a thorough review of the available flow measurement data and techniques and to assess how a strict awareness of the limitations of the data may affect previous analyses. It is concluded that the pre-1930s discrete streamflow measurement data are not of sufficient accuracy to be compared with modern streamflow values in establishing long-term trends of river behavior.

The US Army Corps of Engineers has constructed numerous river engineering structures in and along the Middle Mississippi River. River training and bank stabilization measures include pile dikes, stone dikes, bendway weirs, chevrons, and revetments. Concerns have long been voiced about the effects of these structures on flood stages. Recent debate concerning the scientific basis for these concerns has been invigorated by specific gauge analysis that appears to show a rising trend in flood stages over the past 150 years. This paper attempts to advance the debate by providing an objective review of the specific gauge analysis technique that explains how the method should be performed and the results interpreted; identifies strengths and limitations; examines the uncertainties associated with application to the Middle Mississippi River given the available data; and reassesses the conclusions that can and cannot reasonably be drawn regarding the impacts of dikes and levees on flood stages, based on specific gauge analysis of the Middle Mississippi River. The application of specific gauge analysis to records from the St. Louis gauge, by using methods that carefully follow the guidelines set out in this paper, provides no evidence that dike construction has raised flood stages, but indicates that rises in flood stages are more probably related to levee building.

This research provides a new numerical assessment of the diffusion wave model regarding the effect of selecting different state variables (water depth versus discharge) on the accuracy of the results. It also demonstrates the detailed implementation of a rapid and efficient method (differential quadrature) for this model. The diffusion wave model is still an active area of research in hydraulic and hydrologic engineering because of its wide range of applicability, low computational cost, and high efficiency. The diffusion wave equation can be cast in five different forms, depending upon the state variable selected, whereas Q-based (discharge) and h-based (water depth) can be considered as the main alternative formulations. A rigorous mathematical assessment of these different forms is lacking in existing literature. Numerical experiments were carried out to assess the numerical accuracy of the h-based and Q-based formulations. The fully dynamic wave results were used as benchmarks. Several parameters such as bed resistance, upstream hydrograph shape, and channel bed slope were changed to compare the accuracy of different formulations of the diffusion wave model. The results show that two formulations can give accurate results for steep slopes. However, h-based results were subjected to more errors compared with Q-based methods in very mild slopes. Conversion of the downstream boundary condition from kinematic to diffusion reduces the discrepancy dramatically. DQM can be considered as a powerful alternative numerical scheme to explore such avenues in diffusive flood routing on the basis of its efficiency and accuracy.

Groups of piles are frequently used as bridge foundations. Different group configurations, characterized by different pile spacing, skew-angle, number, and arrangement of pile group columns interact differently with the flow field and lead to different scour patterns and equilibrium scour depth. There have been a number of past studies on the characterization of scouring at pile groups, but most of them report short duration scour experiments. A priori, such short durations may be postulated to inherently carry important uncertainties into existing scour predictors. In this study, 75 long-duration laboratory tests were run under steady, clear-water flow close to the threshold for initiation of sediment motion, to address the effect of time, pile spacing, skew-angle and number of pile group columns on the equilibrium scour depth. Pile groups consisted of matrical arrangements of one, two, or three columns of four rows, with spacings of 1, 2, 3, 4.5, and 6 pile diameters; the tested skew-angles were 0, 15, 30, 45, and 90°. Important contributions were achieved on (1) the impact of the duration of tests on the shape of the scour hole as well as on the precision of coefficients involved in current predictors, (2) the most unfavorable skew angle, (3) the behavior of collapsed pile groups, (4) the maximum scour depth at pile groups composed of a single alignment, and (5) the performance of two current predictors of scour depth at pile groups. Two formulations of a predictor for the calculation of an aggregated pile group factor are suggested.

This study examines the probability for sediment entrainment to bed load and the probability for the threshold condition of sediment to be in suspension. The theoretical analysis is based on a simple one-sided exponential distribution of probability function. The probability distributions are derived from a truncated universal Gram-Charlier series expansion based on the exponential or Laplace-type distributions for turbulent velocity fluctuations. The key criterion of sediment entrainment is the hydrodynamic lift acting on a solitary particle to exceed submerged weight of the particle. In this way, a simple probability function for sediment entrainment to bed load in terms of Shields parameter containing the lift coefficient is obtained. It was found that the value of lift coefficient as 0.15 satisfactorily fitted the probability function versus Shields parameter curve with the experimental data. On the other hand, the key criterion of the threshold of sediment suspension is the fluctuations of the vertical velocity component to exceed terminal fall velocity of the particle. The probability function for the threshold of a sediment particle to be in suspension is obtained in terms of Shields parameter as a function of shear Reynolds number. Curves for different values of probabilities are drawn in respect to a Shields diagram. For the value of probability 0.05, the threshold of sediment suspension is indicated. The prediction curves for the threshold of sediment suspension are proposed in terms of Rouse number versus Shields parameter and also Shields parameter versus shear Reynolds number.

Bed-load sediment transport is important yet difficult to measure in large, sand-bed rivers. Prior work established in theory and validated in a flume study a method known as integrated section surface difference over time version 2 (ISSDOTv2), which computes bed-load transport using sequential three dimensional (3D) bathymetric profiles. The same work identified a source of systemic error leading to an underprediction of computed transport rates. This paper demonstrates how the systemic error can be removed from ISSDOTv2 calculations to produce a more accurate dune transport value. This is demonstrated by analytic and geometric examples, and with field data from the Missouri River at Kansas City. The results of these analyses indicate that field data do exhibit the systematic error and that it is possible to extrapolate a corrected dune bed-load transport rate from field data obtained at different measurement intervals. Additionally, error bounds on this corrected value can be set. Independent validation of the resulting transport rate was not attempted due to the unreliability of physical measurements for bed-load transport in large sand-bed rivers.