Journal of Hydraulic Engineering

Studies on scour in the wake region of piers in cohesive sediments containing gravel were not reported in literature, to the best of our knowledge. Results from the laboratory experiments based study on the process of wake scour are presented here. The sediment used for the experiment were of two types: (1) fine gravel mixed with clay in proportions varying from 20% to 60% by weight, and (2) fine gravel and fine sand in equal proportion by weight mixed with clay in proportions varying from 20% to 60% by weight. The experiments revealed that the process of scour as well as depth, shape, and geometry of scour hole developed in such cohesive sediment were significantly different from that of cohesionless sediments. Proportion of clay fraction and unconfined compressive strength of cohesive sediment mixture were found to be the most significant variables controlling the depth of scour in the wake of piers. On the basis of dimensional considerations, relationships have been proposed for the estimation of the depth of scour in the wake region of piers in clay-gravel and clay-sand-gravel sediment mixtures.

Twenty-three of the more recent and commonly used equilibrium local scour equations for cohesionless sediments were evaluated using compiled laboratory and field databases. This investigation assembled 569 laboratory and 928 field data. A method for assessing the quality of the data was developed and applied to the data set. This procedure reduced the laboratory and field data to 441 and 791 values, respectively. Because the maturity of the scour hole at the time of measurement for the field data was unknown, they were only used to evaluate underprediction by the equations. A preliminary quality control screening of the equilibrium scour methods/equations reduced the number of equations from the initial 23 to 17. For this screening procedure the equations were used to compute scour depths for a wide, but practical, range of structure, flow, and sediment parameters. Those methods/equations yielding unreasonable (negative or extremely large) scour depths were eliminated from further consideration. The remaining 17 methods/equations were then analyzed using both laboratory and field data. Plots of underprediction error versus total error for the laboratory data and underprediction error for field data versus total error for laboratory data along with error statistics calculations assisted in the ranking of the equations. Equations from previous publications were melded and slightly modified to provide the best performing equation in that it yields the least total error and close to the least under-prediction error of those tested. The new equation is termed the Sheppard/Melville (S/M) equation in this paper.

The sediment erosion rate flume (SERF) device was computationally modeled using CD-adapco’s Star-CCM+ at varying flow rates and sample roughnesses so that wall shear stresses could be evaluated during a piston-style erosion test. Shear stress data were matched between the model and data from previous physical tests. Pressure differential upstream and downstream from an eroding specimen displayed similar behavior during both physical and modeled tests in that as eroding sample roughness increased, pressure differential did not appear to change. A series of complicated bed configurations were added to the computational model to quantify the effects of blocking, chunking, or sample overadvancement during an erosion test. Results appeared to indicate that small deviations in sample geometry may have large effects on localized shear stresses. Another series of models was run to provide an explanation for the beginning of blocking or chunking as a result of shear stress development over a rough sample. Results showed that relatively large localized shear stresses may develop simply because of the introduction of roughness. Finally, results also indicated that the most conservative method for future testing is to keep the bottom edge of an eroding sample flush with a flume bottom.

The drift velocity, at which sediment disperses relative to the motion of water-sediment mixtures, is a key variable in two-phase mixture equations. A constitutive relation for the drift velocity, expressed as a power series in the particle bulk Stokes number, was obtained by solving the momentum equation for sediment with the perturbation approach. It shows that gravity and turbulent diffusion are the primary dispersion effects on sediment, whereas flow inertia, particle-particle interactions, and other forces such as lift are the first-order particle inertial corrections that also play significant roles in sediment suspension. Analysis proves that studies based on turbulent diffusion theory are the zeroth-order approximations to the present formulation with respect to the particle inertia effect. The vertical concentration and velocity distributions of sediment in simple flows were investigated with the two-phase mixture equations closed by the drift velocity acquired in the research reported in this paper. The calculated concentration profiles agree well with measurements when the first-order particle inertial effect is considered. The calculated velocity of sediment coincides with available experiments that sediment lags behind water in open-channel flows as a result of turbulence-induced drag.

In this study, an array of rigid cylindrical rods was used to simulate emergent vegetation stems that were subject to unidirectional open channel flows. The instantaneous drag force experienced by the rods was measured with a load cell. In addition, the particle image velocimetry (PIV) technique was applied to sample the flow information in a horizontal plane, and wave gauges were used to record the fluctuation in the water-surface elevation. The results show that the drag fluctuation normalized by the mean value may reach as high as 133% when the Reynolds number (defined based on the stem diameter) varied in the range from 400 to 1,100. High fluctuations were also observed in the flow velocity and flow depth under similar flow conditions.

Although advanced physically based models have been presented in the last years, there is still much interest in equations for predicting the peak discharge. Indeed, simple relations are very useful for technical purposes, especially when rough calculations are acceptable. This paper presents new equations for assessing the peak discharge of the flood hydrograph caused by the gradual erosion of earth dams. The equations are obtained by a regression analysis carried out on the peak discharges provided by a physically based numerical model. The approach proposed in this paper has the advantage of taking into account some parameters usually neglected by common best-fit equations that have strong influence on the peak discharge. In particular, the equations discussed in this paper take into account the susceptibility to erosion of the dam, the shape of the reservoir, and the ratio of filling of the reservoir for failures caused by piping. The usefulness of the equations is shown through a comparison with the numerical model and with other regression relations by applying them to some real dam-breaching events. Finally, a sensitivity analysis is made in order to assess the influence of parameters that have been fixed in the present analysis, like dam crest width or abutment slopes.

Scouring downstream of submerged weirs is a common problem resulting from the interaction of the three-dimensional turbulent flow field around the structures and the mobile channel bed. This paper presents the distributions of flow patterns, bed shear stresses, and turbulence structures in the approach flow and the scour hole downstream of a submerged weir. The experiments were conducted under the clear-water scour condition for an equilibrium scour hole. The experimental results show that the flow structures are considerably changed by the presence of the structure. A large recirculation zone and a flow reattachment region are formed downstream of the submerged weir. Strongly paired cellular secondary flows are observed in the scour hole. The turbulence structures ahead of the recirculation zone govern the dimensions of the scour hole.

An experimental study of the supercritical dividing flow in a right-angle, four-branch junction with different channel widths is presented. Lateral-to-mainstream width ratios of $1:1$, $2:3$, and $1:2$ were used to establish the influence of the channel width on the flow distribution. A brief description of the flow characteristics is given. Types I and II flow patterns are confirmed by the experiments in the range of the studied variables. Existing flow distribution models were adapted for width ratios other than $1:1$ and tested for Type II, subregime 3 flow pattern, finding poor results. Other existing flow distribution models for Type II, subregimes 1 and 2 flow pattern and a width ratio of $1:1$ give a very good fit with the experimental data. A new flow distribution model for Type II, subregimes 1 and 2 flow pattern and width ratios of $2:3$ and $1:2$ was proposed, which is a function of the Froude number and the width ratio. A former model for predicting the flow distribution in equal-width ($1:1$ width ratio) junctions was extended for width ratios of $2:3$ and $1:2$. This model expresses the flow distribution as a function of the inflow power ratio and has the advantage over existing models that, for its application, the flow structure (angles of the oblique jumps) in the junction does not need to be determined beforehand. Although it is not necessary to apply the proposed flow distribution model, the flow structure (flow type and angle of the oblique jump or jumps, if presented) can also be predicted through the inflow power ratio, the inflow ratio, and the channel slope combination in the case of width ratios of $2:3$ and $1:2$.

Water distribution network (WDN) models account for customer-demands as water withdrawals concentrated in nodes. Customer-demands can be assumed to be constant or varying with nodal head/pressure entailing demand-driven or pressure-driven simulation, respectively. In both cases, the direct connection of customer properties to the hydraulic system is implicitly assumed. Nonetheless, in many technical situations, the service pipe fills a local private storage (e.g., a roof tank or a basement tank) from which the water is actually delivered to customers by gravity or pumping systems. In such contexts, the service pipe fills the local tank by means of a top orifice. Consequently, what is really connected to the hydraulic system is a tank, which is subject to a filling/emptying process while supplying water to customers. Therefore, since modeling this technical situation in WDN analyses is necessary, the paper develops a formulation for nodal water withdrawals in WDN models accounting for the filling/emptying process of inline tanks between the hydraulic network and customers. The formulation is also introduced in a widely used method for steady-state WDN modeling, the global gradient algorithm, and its effectiveness to increase the hydraulic accuracy of results is discussed using a simple case study and a small network.

Stormwater infiltration and inflow are major contributors to sewer flow, and thus they can be significant triggers for combined and sanitary sewer overflows, both of which introduce contaminants to surface waters. There are few estimates of private residential rainfall-dependent inflow and infiltration (RDII), and this paper proposes that a rapid and cost-effective means to locate points of significant stormwater entry into the sewer system would be advantageous for combined and sanitary sewer management. The authors studied the collection system in the Barthman-Parsons area of Columbus, Ohio, performing detailed drainage and connectivity investigations on 116 private houses located in areas served by separated sanitary sewers. Sources of inflow and infiltration (I/I) were identified for private residential properties, which could include sump pumps, foundation drains, downspouts, cleanouts, yard drains, and defective service laterals. The authors then developed estimates of I/I contributions from residential properties in different neighborhood clusters, and these estimates were extrapolated to the entire Barthman-Parsons area. Field results found that 68% of tested properties contribute to I/I. Of the tested sample, 25% had at least one downspout that tested positive for I/I, and 59% had a lateral that tested positive. The results also showed that downspouts and laterals contributed approximately 98% of the total I/I volume generated by the tested properties in response to testing. These residential I/I sources were estimated to make up approximately 35% of total I/I for short, intense storms with dry antecedent conditions, and approximately 7% of total I/I under low-intensity, long-duration storms with wet antecedent conditions. An internally validated logistic model developed for the downspout data set was fairly accurate at predicting whether a property would test positive. It correctly classifies the downspouts of 17 of 27 houses as contributing I/I, and incorrectly predicts 14 as contributing. Had this model been available prior to testing, a majority of the houses’ downspouts contributing to I/I could be eliminated without testing the entire population, offering significant savings in assessment and testing costs, and leading to an overall faster turnaround in system improvement.

Entrainment of sediment from hillslope and channel usually has a significant influence on resistance of debris flow. This effect is studied by flume experiments of debris flow over rigid and erodible beds. Fifty contrast flume tests show that the flow resistance over the erodible bed is significantly larger than that over the rigid bed because of higher energy consumption resulting from sediment transport. A new formula of debris-flow resistance involving the entrainment effect is obtained from the erodible-bed data by nondimensional multiple regression analysis.