Particle-laden jets are an important type of multiphase flow which can be found in various natural and technical processes. This study focuses on the flow mechanisms in a horizontally discharging sediment-laden jet that is of particular interest in environmental science and engineering. Experimental techniques and mathematical models are developed to investigate horizontal sediment-laden jets, both for the buoyant and non-buoyant jet discharge cases. In the laboratory, the separation of images of the fluid and the particulate phases is achieved by harnessing light signals of visualization at different wavelengths. Whole field measurements of velocities of the two phases are made by the adoption of particle image velocimetry (PIV) algorithms. Numerical models are developed in two approaches with regard to the treatment of the particulate phase. In the Lagrangian approach, individual sediment particles are tracked while the flow field of the fluid phase is computed with large-eddy simulation (LES). This simulation successfully captures the transient nature of the particle-laden flow. In the Eulerian approach, a two-phase model is used to obtain steady flow simulations in a much shorter computation time. The experimental and numerical results for the horizontal momentum jets show that, at low initial particle concentrations, the sediment particles generally follow the jet flow but with some levels of deficit velocities. In the upper layer of the jet the particles do not follow the fluid flow as well as in its lower layer. More particles are observed in the lower layer than in the upper one. For the momentum-dominated zone of a horizontal buoyant jet, the flow exhibits similar behaviors as the horizontal particle-laden momentum jet, except that there are some slight modifications from the effects of buoyancy. In the bending zone of the buoyant jet, the effects of buoyancy become significant. Notably, the locations of maximum velocity magnitude and those of maximum turbulence intensity are well separated in this zone. A strong correlation of particle abundance and high turbulence intensity is observed in the lower outer jet layer in this bending zone. Significant modifications to the global behaviors of horizontal sediment jets are observed as the particle concentration increases to relatively high levels. The jet trajectories are brought downwards by the particle loads and the jet widths are also increased. For the flow regime being investigated, turbulence intensity in the fluid flow is found to be increased by the presence of sediment particles. The results suggest that turbulence helps suspend sediment particles in horizontally discharging jets. A Stokes number is proposed to represent the ability of particles to follow the fluid flow. It is defined as St=W_s/U_j , where ws is the particle settling velocity in still fluid and Uj is the jet exit velocity, which indirectly governs the turbulence characteristics of the jet flow. The advecting large eddies in a turbulent jet are found to play the role of organizing particles in patches. Interaction and coalescence between particle-concentrated eddies may result in the sudden drop of a group of particles, which contributes to sediments falling from a horizontal jet in the form of particle-rich “fingers”. / published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Sediment-laden turbulent buoyant jets are commonly encountered in the natural and man-made environments. Examples of sediment-laden buoyant jets include volcanic eruptions, deep ocean hydrothermal vents (“black smokers”), ocean dumping of dredged spoils and sludge, and submarine discharge of wastewater effluent. It is important to understand the fluid mechanics of sediment jets for environmental impact assessment, and yet there is currently no general model for predicting the mixing of sediment-laden jets. This study reports a theoretical and experimental investigation the sediment mixing, fall-out and deposition from sediment-laden buoyant jets. It is well known that turbulence generates fluctuations to the particle motion, modulating the particle settling velocity. A general three-dimensional (3D) stochastic particle tracking model is developed to predict the particle settling out and deposition from a sediment-laden jet. Particle velocity fluctuations are modelled by a Lagrangian velocity autocorrelation function that accounts for the loitering and trapping of sediment particles in turbulent eddies which results in the reduction of settling velocity. The model is validated against results of independent experimental studies. Consistent with basic experiments using grid-generated turbulence, the model predicts that the apparent settling velocity can be reduced by as much as 30% of the stillwater settling velocity. The mixing and deposition of sediment-laden horizontal momentum jets are studied using laboratory experiments and 3D computational fluid dynamics (CFD) modelling. It is shown that there is a significant settling velocity reduction up to about 25-35%, dependent on jet turbulent fluctuations and particle properties. The CFD approach necessitates an ad hoc adjustment/reduction on settling velocity and lacks generality. Using classical solutions of mean velocity, and turbulent fluctuation and dissipation rate profiles derived from CFD solutions, 3D particle tracking model predictions of sediment deposition and concentration profiles are in excellent agreement with measured data over a wide range of jet flow and particle properties. Unlike CFD calculations, the present method does not require any a priori adjustment of particle settling velocity. A general particle tracking model for predicting sediment fall-out and deposition from an arbitrarily inclined buoyant jets in stagnant ambient is successfully developed. The model incorporates the three flow regimes affecting the sediment dynamics in a buoyant jet, namely turbulent jet flow, jet entrainment-induced external flow and surface spreading current. The jet mean flow velocity is determined using a well-validated jet integral model. The external jet-induced irrotational flow field is computed by a distribution of point sinks along the jet trajectory. The surface spreading current is predicted using an integral model accounting for the interfacial shear. The model is validated against experimental data of sediment deposition from vertical and horizontal sediment-laden buoyant jets. / published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Debris flows in at least 529 Grand Canyon tributaries transport poorly-sorted clayto boulder-sized sediment into the Colorado River, and are initiated by failures in weathered bedrock, the "fire-hose effect," and classic soil-slips often following periods of intense rainfall coincident with multi-day storms. Recent debris flows had peak-discharges from about 100-300 m3/s. Twentieth-century debris flows occurred from once every 10-15 years in eastern tributaries, to once in over a century in western drainage areas. Systemwide, debris flows likely recur about every 30-50 years, and the largest recent flows were initiated during Pacific-Ocean storms in autumn and winter. Three idealized hydrographs are inferred for recent debris flows based on deposits and flow evidence: Type I, has a single debris-flow peak followed by a decayed recessional streamflow; Type II, has multiple, decreasing debris-flow peaks with intervening flow transformations between debris flow and non-debris flow phases; and Type III, may have either a simple or complex debris-flow phase (begin as either Type I or II), followed by a larger streamflow peak that reworks or buries debris-flow deposits under streamflow gravel deposits. From 1987 through 1995, at least 25 debris flows constricted the Colorado River, creating 2 rapids and enlarging at least 9 riffles or rapids. In March-April, 1996, reworking effects of a 7-day controlled flood release (peak = 1,300 m³/s) on 18 aggraded debris fans in Grand Canyon were studied. Large changes occurred at the most-recent deposits (1994-1995), but several other older deposits (1987-1993) changed little. On the most-recent fan deposits, distal margins became armored with cobbles and boulders, while river constriction, flow velocity, and streampower were decreased. Partial armoring of fan margins by relatively-low mainstem flows since the debris flows occurred, was an important factor limiting fan reworking because particles became interlocked and imbricated, allowing them to resist transport during the flood. Similar future floods will accomplish variable degrees of fan reworking, depending on the extent that matrix-supported sediments are winnowed by preceding mainstem flows.

The assumption of perfect stratification in an aquifer has been widely used in solute-transport modeling studies. This assumption is especially useful for applied studies where limited site characterization data are available, but geologic well logs indicate significant layering. Chapter 3 investigates the issue of vertical sampling density via a sensitivity analysis of the number of aquifer layers used in a model of tracer transport through a heterogeneous synthetic aquifer. Tracer breakthrough in the synthetic aquifer is predicted by layered models. Given a variance of ln K of 2 and an exponential covariance function, sampling the synthetic aquifer at more than 12 elevations did not produce any significant improvement in the predictions. Even six sampling points, however, produced more accurate predictions of transport compared to a full-aquifer, homogeneous approach employing a local-scale dispersivity. Chapter 4 presents and interprets data from a dual-well, forced-gradient tracer experiment conducted in a confined aquifer underlying a contaminant source zone of a Superfund site. Tracer breakthrough was monitored at an extraction well and at four levels of a centerline monitoring well. A perfectly stratified numerical transport model based on multi-level data successfully predicted tracer breakthrough at the extraction well. Given the added vertical resolution associated with the layered model, it was possible to use dispersivity values more than an order of magnitude lower than the value used in a vertically integrated model. It is expected that the multi-layer model would allow for more robust analyses of solute transport at the site. In Chapter 5, TCE elution during the same dual-well experiment is predicted with a stratified numerical model incorporating rate-limited desorption, rate-limited diffusion, and rate-limited dissolution of nonaqueous phase liquid (NAPL). Based on model results, initial mass calculations, and other indirect lines of evidence, it is concluded that NAPL is the primary cause of rate limitations for TCE transport at the site. NAPL presence is the primary reason a large pump-and-treat system at the site has failed to reduce contaminant concentrations to federal drinking water standards. Alternative remediation technologies are thus necessary for restoring the aquifer, especially in the contaminant source zone.

Mature karst topography is well recognized within the hydrology and geology communities to include subterranean fluid pathways that act as turbulent conduits conveying fluid from surface stream sinks called swallets to sources called springs. However, we find that little knowledge has been reported with regards to the transport and fate of terrestrially-derived sediment organic carbon (SOC) within karst watersheds. This study investigated the hypothesis that karst pathways could act as biologically active conveyors of SOC that temporarily store sediment, turnover carbon at higher rates than otherwise considered, and recharge depleted SOC back to the surface stream within the fluvial system. Mixed research methods were applied within a mature karst network. Methods included high resolution measurements of water and sediment characteristics of surface streams, carbon and stable carbon isotope measurements of transported sediment, and numerical modeling of water and sediment pathways. The mixing of sediment during net zero deposition and erosion was investigated in this study using a parameter calibrated to SOC data. Results of this study showed that heterotrophic bacteria in the subsurface conduit oxidized 0.05 tCkm-2y-1 resulting from the temporary storage of terrestrial carbon in the karst conduit. The subsurface conduit transports 0.15 tCkm-2y-1 out of the fluviokarst watershed.

This dissertation examines the dynamics of sediment transport and channel morphology in the lower Mississippi River. The area of research includes the portion of the river where reach-averaged downstream flow velocity responds to the boundary condition imposed by the relatively uniform water-surface elevation of the receiving basin. Observational studies provided data that are used to identify channel-bed sediment composition, and measure bed-material sediment flux and the properties of the fluid-flow field over a variety of water-discharge conditions. The analyses demonstrate that a significant portion of the channel bed of the final 165 kilometers of the Mississippi River consists of exposed and eroding underlying relict sedimentary strata that qualify as surrogate bedrock. The exposed bedrock is confined to the channel thalweg, particularly in river-bend segments, and actively mobile bed-material sediments are positioned on subaqueous bars fixed by river planform. The analyses for sediment flux provides insight to the nature of sediment transport: during low- and moderate-water discharge, bed-material movement occurs primarily as minimal bedform flux, and so bed materials are not transferred between alluvial bars. During high-water discharge, bed-material transport increases one-hundred fold, and sands move as a part of both suspended and bedform transport. Physical models are used to show that skin-friction shear stress increases by a factor of ten for the measured water-discharge range. This change is not possible given conditions of uniform water flow, and therefore non-uniform flow in response to the Mississippi River approaching its outlet has a significant impact on the timing and magnitude of sediment flux through the lower river. In order to estimate the dynamics of bed material movement from the uniform to non-uniform segment of the river (lower 800 km), data for channel morphology are used to construct a model that predicts spatial changes in water-flow velocity and bed-material flux over a range of water-discharge conditions. The model demonstrates that non-uniform flow tends to produce a region of net channel-bed aggradation between 200-700 kilometers above the outlet, and a region of channel-bed degradation for the final 200. The implication for these results for the spatial variability of channel morphology and kinematics is explored. / text

Storm surge, the pileup of seawater occurring as a result of high surface stresses and strong currents generated by extreme storm events such as hurricanes, is known to cause greater loss of life than these storms' associated winds. For example, inland flooding from the storm surge along the Gulf Coast during Hurricane Katrina killed hundreds of people. Previous storms produced even larger death tolls. Simultaneously, dune, barrier island, and channel erosion taking place during a hurricane leads to the removal of major flow controls, which significantly affects inland inundation. Also, excessive sea bed scouring around pilings can compromise the structural integrity of bridges, levees, piers, and buildings. Modeling these processes requires tightly coupling a bed morphology equation to the shallow water equations (SWE). Discontinuous Galerkin finite element methods (DGFEMs) are a natural choice for modeling this coupled system, given the need to solve these problems on large, complicated, unstructured computational meshes, as well as the desire to implement hp-adaptivity for capturing the dynamic features of the solution. Comprehensive modeling of these processes in the coastal zone presents several challenges and open questions. Most existing hydrodynamic models use a fixed-bed approach; the bottom is not allowed to evolve in response to the fluid motion. With respect to movable-bed models, there is no single, generally accepted mathematical model in use. Numerical challenges include coupling models of processes that exhibit disparate time scales during fair weather, but possibly similar time scales during intense storms. The main goals of this dissertation include implementing a robust, efficient, tightly-coupled morphological model using the local discontinuous Galerkin (LDG) method within the existing Advanced Circulation (ADCIRC) modeling framework, performing systematic code and model verification (using test cases with known solutions, proven convergence rates, or well-documented physical behavior), analyzing the stability and accuracy of the implemented numerical scheme by way of a priori error estimates, and ultimately laying some of the necessary groundwork needed to simultaneously model storm surges and bed morphodynamics during extreme storm events. / text

published_or_final_version / abstract / Mechanical Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / toc / Geography / Doctoral / Doctor of Philosophy

No description available.