Journal of Environmental Engineering

Water-treatment processes can be regarded as a series of black boxes, with dirty water in and cleaner water out. This approach to coagulation, flocculation, clarification, filtration, and flotation processes may result in satisfactory design, but more rational optimization can be achieved with insight into the internal workings of these process units. Since 1957, University College London researchers have been peering into the “black boxes” of these processes in order to learn the inside story of their detailed workings. The earliest of these used radioactively labeled unicellular algae, so that when filtered through sand, their detailed deposition pattern within the filter could be detected without interfering with the process. Later, studies on floc blanket clarification in a full-size hopper-shaped tank (the normal U.K. design at that time) measured the three-dimensional flow pattern with a lithium tracer. This demonstrated the advantage of flat-bottomed clarifiers, which is now a common design development. Recently, the use of video-endoscopy has allowed direct observation and recording of deposition within filter pores as it occurs dynamically, without the need to extract samples. The same technique, but using high-speed videorecording, has been applied to filter washing, showing the advantage of air-water backwashing, employing the “collapse-pulsing” mode of enhanced cleaning. These latest techniques are being applied to bubble formation in dissolved-air flotation to examine the details of release of air and floc attachment at the depressurization point.

The dynamics of particle capture and head-loss development were studied in laboratory filters over extended periods. Particle-size distributions and pressures were measured above the sand media and at three depths. Conditions included filtration velocities of 1.8 and 5.5 mm/s and mean media sizes of 0.39, 0.78, and 1.85 mm. The suspension was sedimentation effluent from a softening water-treatment plant. Ripening, the improved removal efficiency with time, was extensive for small particles, but breakthrough, the decreased removal with time, was more significant for intermediate and larger particles. Detachment of previously retained particles was demonstrated to contribute substantially to breakoff; apparently, flocs are formed on the media surface by the sequential capture of particles and break off together. Theory suggests that higher velocities and larger media sizes can be offset quantitatively, with respect to particle removal, by proportionately greater depths; data from this research support these ideas. Comparison with previous experiments with glass spheres as the media but otherwise identical conditions showed that, for sand, initial removal was better, ripening was less extensive (perhaps because of initially better removal), and detachment was more extensive.

Considerable research on the process of filter backwashing with fluidization has been done during the last two decades, but there is little fundamental understanding of the mechanisms of particle detachment from filter media. The overall goal of the research reported in this study is to gain a fundamental understanding of the mechanisms by which colloidal-sized particles detach from larger grains. The theoretical part of this study has focused on calculating the interaction forces between a single particle and a collector surface coated with a deposited layer. The interaction forces that would control particle detachment include van der Waals attractive forces, electrical double-layer interactive forces, Born repulsive forces, and structural forces. The experimental studies were performed on laboratory scale model systems with well-defined spherical particle suspensions. Latex particles (5 μm) were filtered in a packed bed column of glass beads. The detachment of particles during filter backwashing under fluidization conditions was studied in relation to the chemical characteristics of the backwash water. The experimental results indicated that higher pH and lower ionic strength enhance detachment during backwashing. Theoretical force calculations based on Derjaguin, Landau, Verwey, and Overbeek (DLVO) theory alone could not explain detachment behavior. Inclusion of short-range forces in the calculations qualitatively explains the experimental results.

Clean-bed filtration models are frequently used to make engineering calculations on aquasol removal in filters. They consist of a filtration equation, derived from a mass balance around a collector, a single-collector collision efficiency, and an empirically derived sticking coefficient. In order to be correctly applied and referenced, the two most commonly used models, the Yao model and the Rajagopalan and Tien (RT) model, require additional clarification. The filtration equation used in the Yao model has a different form than that of an aerosol counterpart due to assumptions made in choosing a reference velocity in the system. The RT model, as originally published, had variables hidden within constants and contained typographic errors. In this paper we identify the differences in the Yao and aerosol filtration model derivations, correctly present the RT model, and compare model predictions with particle removals reported in laboratory columns.

Models describing permeate flux, rejection, and cost are coupled to evaluate the performance and cost of ultrafiltration as a function of raw-water quality. The model for permeate flux extends a previous model for colloidal fouling based on shear-induced diffusivity to include Brownian diffusion. Contaminant removal is modeled as mechanical sieving and molecular diameter is regressed against weight to describe removal of natural organic matter (NOM). Time-dependent permeate flux is considered in estimating operating times required to achieve a specified recovery. Costs are calculated as a function of particle-size distribution in the raw water. Particles with diameters on the order of 10 ⁻¹ μm display minimum diffusivities, which leads to maximum system costs with respect to particle size. Fine materials ( <0.5 μm), with high cake resistance, demonstrate pressure-independent permeate flux for conditions typical of hollow fiber ultrafiltration. In some cases, a minimum in system costs as a function of recovery is observed due to a trade-off between operating time and time-averaged permeate flux. Simulations for four scenarios of variable particle and NOM concentrations suggest that irrespective of adsorptive fouling, permeate flux may be limited by reversible accumulations of NOM on ultrafiltration membranes.

Fouling experiments of thin film composite and cellulose acetate reverse osmosis membranes by aluminum oxide colloids are reported. Membrane fouling was investigated at various solution chemistries under fixed hydrodynamic conditions. Results show that the fouling rate increases with an increase in the ionic strength of the solution. Fouling was significant at high ionic strengths, including in the presence of background dissolved organic matter, resulting in a gradual decrease in product water flux. Under the chemical conditions tested, colloidal fouling was found to be reversible, thus indicating that pore blockage is not an important factor in colloidal fouling of reverse osmosis membranes. The role of chemical-colloidal interactions in colloidal fouling of reverse osmosis membranes is elucidated. Fouling is controlled by particle-membrane and particle-retained particle interactions, which, in turn, are determined by solution chemistry, chemical properties of colloids and membranes, and the magnitude of permeation drag.

The effect of low water temperature on flocculation kinetics was examined using ferric nitrate to coagulate kaolin clay in water under tightly controlled treatment conditions. Both the particle size distribution data obtained from an automatic image analysis system and the on-line measurement of the degree of turbidity fluctuation in a flowing suspension by a photometric dispersion analyzer were used to measure flocculation kinetics. Results show that cold water temperature had a pronounced detrimental effect on flocculation kinetics. To improve flocculation kinetics at low water temperature, maintaining constant pOH to adjust water chemistry for temperature changes was found to be partially effective, but only in the more acidic pH range studied (≤pH 6.8). However, the improved performance did not reach the performance at room temperature.

Although the jar test has long been used in coagulation and flocculation, there is little available information on vessel hydrodynamics. Vessel-average parameters such as the root-mean-square velocity gradient, $\overline{G}$ , have traditionally been used for hydraulic simultude between the jar test and the full scale. With the recognition that vessel-average parameters may not truly represent hydraulic conditions within a vessel, there is a need to characterize jar test hydrodynamics. Analysis of turbulent flow in a standard jar test apparatus was completed using a two-dimensional laser doppler anemometer. Three dimensional velocity measurements were used to determine the general flow field, dissipation rates, and turbulent length scales within the jar. Substantial variations in turbulent characteristics were found to occur within the vessel. Scaling relations were developed for various turbulent parameters. Values for scaling constants were found to be different than those reported for the Rushton turbine, which is by far the most studied mixing impeller. For example, “scaling constants” for local turbulent dissipation rates in the impeller zone were found to be in the order of five times less than the Rushton turbine. This result is significant because many researchers have accounted for inhomogeneties in mixing characteristics within the jar test apparatus by using findings from available data on the Rushton turbine.

The objective of this study is to evaluate the significance of four structural design features (inlet configuration, weir location, intermediate baffle existence, and sludge draw-off location) and their interactions on the control of density currents in activated-sludge secondary clarifiers. A 1:25-scale model of a prototype rectangular clarifier was used, and experiments were conducted in the field with activated-sludge mixed liquor. A factorial design procedure was used to estimate quantitatively the significance of each structural design feature and their interrelationships. The effect of the structural design features on the performance of an activated sludge secondary clarifier depended on the predominant flow pattern (bottom or surface density current) through the clarifier. Because of the strong interactions among the structural design features, the effectiveness of each structural-design-feature level (i.e., the particular configuration of a structural design feature) on the performance of the clarifier was strongly related to the condition of the other structural features. A regression equation for the effluent suspended-solids concentration was developed from the experimental data.

Heavy-density currents occur when the inflow is denser than the ambient fluids in a basin. These currents exist in most secondary sedimentation basins. This paper investigates the internal flow regimes associated with the density currents due to cold influent entering circular clarifiers. The experimental studies included flow with unsteady influent temperature under different conditions of geometry, momentum, and negative buoyancy. Velocity and temperature of different flow regimes were measured throughout the basin. The flow phenomena progressed through a sequence of density flow regimes: the denser wall jet, internal hydraulic jump, rebound at the gate, and stratification. The data were analyzed to determine the characteristic depths, velocities, and isotherms for the radial density currents in a basin. This study shows that circular tanks subject to variable influent temperatures (or concentrations) experience internal surges in the form of moving hydraulic jumps. Depending on the tank geometry, this phenomenon could result in degraded performance of clarifiers during transient density events.