Granular Media Filtration Research Articles

BIOFILM FORMATION IN GRANULAR‐BED FILTERS

ABSTRACTA high microbiological quality of drinking water must be ensured to protect public health. The filtration techniques that are used in treating drinking water play an important role; however, a biofilm can form on granular‐media filters and the accumulated bacteria can slough off and enter the filtered water.The aim of this research was to examine (a) the potential for biofilm formation and detachment from filter sand, and (b) the effect of different backwash regimes on biofilm removal. During the operation of the filter, bacteria became attached to the sand media, particularly in the top 30 mm of the filter bed. A water‐only backwash at 20% and 40% bed expansion demonstrated poor removal of biofilm throughout the depth of the bed. Collapse‐pulsing is a more efficient method and results in a reduction in the number of bacteria in the filtered water.

Polyurethane foam filter/PAC Adsorber

Stage 1 of the Disinfection/Disinfection‐Byproducts Rule and the Interim Enhanced Surface Water Treatment Rule require better removal of turbidity and total organic carbon (TOC) from drinking water. Activated carbon adsorption is one method that can be used to remove natural organic matter (NOM) from drinking water in order to reduce the production of disinfection byproducts, but using granular activated carbon (GAC) is expensive. This article proposes using a polyester polyurethane foam filter (whose high porosity enables significantly higher filtration rates than those of conventional granular media filters) at a velocity similar to that of fuzzy filters. The filter could be used as a prefilter when clarification does not provide the desired level of turbidity removal, and because of its low head loss, it may fit within an existing treatment plant's hydraulic profile. In addition, the filter could be loaded with powdered activated carbon (PAC) and used as an adsorber. Because PAC is considerably less expensive than GAC, this process could reduce the cost of the adsorption process if the PAC usage approximates that of GAC. The proposed filter/adsorber exhibits a low head loss and requires a shorter contact time than a conventional fixed‐bed GAC adsorber; it should therefore also reduce energy costs and space requirements. To demonstrate the concept, the filter is used to remove turbidity with three different foams at varying degrees of compression with the filter's clean‐bed head loss also monitored. The filter/adsorber concept is tested using PAC to remove NOM (as measured using ultraviolet absorbance at 254 nm (UV254), demonstrating that a fixed‐bed break‐through curve results.

Improving Filtrate Quality Using Agrobased Materials as Coagulant Aid

Abstract In the present study, an evaluation of agrobased materials (ABM) as a coagulant aid in conjunction with alum has been conducted to determine their efficacy in water treatment. The agrobased materials evaluated are Surjana seed (Moringa oleifera), Nirmali seed (Strychnos potatorum) and maize (Zeemays). Experiments have been conducted simulating a conventional water treatment train consisting of coagulation-flocculation-settling and granular media filtration. Emphasis has been given to the filtration aspect of the treatment train using synthetic turbid water. The filter performance was defined by water quality and head loss development across the filter bed. When Nirmali seed or maize was used as a coagulant aid, the alum dose required was 25 and 15 mg/L, respectively, and the filtrate turbidity achieved was less than 0.2 NTU, whereas alum alone with a dose of 45 mg/L achieved filtrate turbidity levels higher than 1 NTU. Thus, the use of ABM improved the filtrate quality. Head loss in filter with Surjana seed and Maize as coagulant aids was comparable to that of alum alone, whereas it was higher when Nirmali seed was used as a coagulant aid.

New options for achieving regulatory compliance with low-pressure membranes

Protection against chlorine resistant pathogens, and Cryptosporidium in particular, has become a critical issue for water purveyors. For plant retrofits, ozone and more recently ultraviolet (UV) light technologies, are considered cost-effective solutions to provide an additional barrier to pathogens. The membrane filtration technology is, however, rapidly evolving and both costs and system recoveries are consistently improving. By physically removing microorganisms membrane filtration avoids the formation of disinfection by-products (DBPs) and the need to address changing target pathogens. This paper will present data collected using a novel pilot which evaluates the custom design approach and allows multiple types of membrane modules to be evaluated within a single pilot unit. Pilot testing was performed at the Marden Water Treatment Plant, in Sheboygan, North Shore, and Green Bay, Wisconsin, during 2000 and 2001. An analysis, based on data generated during these pilot tests, will be presented that shows that granular media filter boxes can be rehabilitated with membrane filtration and increase filtration capacity by at least 15%. This engineering analysis includes evaluation of specific aspects of the process including filtration mode, flux, backwashing, the layout of the existing facilities, the potential to reuse existing equipment, and cost comparisons with proprietary membrane systems.

Testing of Diatomaceous Earth Filtration for Removal of Cryptosporidium Oocysts

Because previous studies of diatomaceous earth (DE) filtration for Cryptosporidium oocyst removal have been conducted on a relatively small scale, this study was designed to determine the performance capability of DE at the pilot‐scale level and at a flow rate greater than that permitted by many state regulatory agencies. This study used DE of moderate permeability that is commonly used by smaller water systems to provide economical water treatment. It found that Cryptosporidium oocyst removals at a flow rate of 1 gpm/sq ft (2.5 m/h) averaged 6.25 logs and removals at 2 gpm/sp ft (5 m/h) averaged 6.31 logs. In addition, average turbidity levels and particle concentrations were lower at 2 gpm/sp ft (5 m/h). These findings of equal or better performance at 2 gpm/sp ft (5 m/h) are significant because many regulatory agencies ‐ and consequently municipal water suppliers ‐ have historically restricted operation of DE filters to 1 gpm/sp ft (2.5 m/h). The findings are also important because they clearly show that DE filtration is capable of providing far greater reduction in Cryptosporidium oocyst concentrations than is provided by conventional or direct granular media filtration. The data summarized in this article support consideration of DE as a more broadly applicable alternative for Cryptosporidium control; they also support including DE filtration as a serious alternative for a much broader range of applications that include larger plant sizes, higher filtration rates, and backwash treatment applications.

Costs of conventional vs. membrane treatment for karstic spring water

Water production technology for the treatment of karstic springs hasto cope with high fluctuations of various water quality parameters, in particular with high turbidity peaks. In addition, natural and manmade dissolved contaminants such as NOM, pesticides and chlorinated hydrocarbons require complex multistage treatment schemes applying conventional process trains like flocculation, granular media filtration, ozonation, granular activated carbon filtration and final disinfection. Several alternatives using membrane processes are proposed in order to simplify the plant schemes and to reduce treatment costs. A cost assessment of existing conventional treatment plants and following membrane systems of the same size is presented: micro-, ultrafiltration with and without powdered activated carbon (PAC) and the integrated system ultrafiltration—nanofitration (MF, UF, MF/PAC, UF/PAC, UF/NF). Based on full- and pilot-scale experience with karstic spring water, capital as well as operating and maintenance costs (O&M) for conventional and membrane plants are evaluated. Detailed cost analysis reveals that compared to conventional systems, membrane treatment leads to a substantial relative increase of O&M fraction. Especially, the dosing of PAC to MF or UF increases energy consumption and costs for PAC renewal. A comparison of the overall specific production costs shows that for the studied plant capacity in the range of 30–300 m 3/h, the costs for simple membrane treatment with MF or UF are 5% less than for conventional process schemes whereas the costs in combination with PAC are 20% higher. Integrated systems using UF/NF combinations may not compete with conventional and UF/PAC plants.

Suspension concentration profiles during rapid gravity filter backwashing

Rapid gravity, granular media filters are widely used in the water and wastewater treatment industries. Regular backwashing to clean the filters is a vital part of their efficient operation. Experimental data on the development of suspension concentration profiles through laboratory scale filter beds during the backwash process were obtained. Previous attempts to obtain and record backwash profiles of this type have been unsuccessful due to the limited range of existing turbidimeters. The results have been used to validate a new model developed by the authors.

Improving Filtration of Cryptosporidium

Surface coating of hydrous iron aluminum oxide on Ottawa sand was investigated as a means of improving the removal of Cryptosporidium oocysts from water by filtration. Coating the sand increased the zeta potential from −40 to +45 mV at pH 7.0, increasing the potential for interaction with the negatively charged (−25 mV at pH 7.0) Cryptosporidium oocysts. Water seeded with Cryptosporidium oocysts was passed through parallel columns of uncoated and coated sand at superficial velocities of 200–800 m/d (3.5–14 gpm/sq ft) and column lengths of 100–400 mm (4–16 in.). In every trial, coated sand removed significantly more oocysts than uncoated sand. Filter coefficients of coated sand were 2.9 times as high as those of uncoated sand, indicating that coating can significantly improve the removal of Cryptosporidium by granular media filters.

High‐Rate Filtration with a Synthetic Compressible Media

An innovative filter that uses a synthetic, porous filter media was tested for more than 2 years at the University of California at Davis. The filter is unusual in a number of ways: (I) the synthetic filter media is highly porous (89%), (2) filter media and bed properties can be modified because the media is compressible, (3) the fluid to be filtered flows both around and through the media instead of only flowing around the filtering media (as in granular media filters), (4) the fluid that is filtered is used to backwash the filter, (5) to backwash the filter, filter bed volume is increased mechanically, and (6) the filter operates at high filtration rates (e.g., 410 to 1230 L/m2·min [10 to 30 gal/min/sq ft]). Performance of the filter, with respect to removal of turbidity and total suspended solids, is similar to the performance of other more conventional filters with the exception that filtration rate is more than 3 to 6 times the rate of other filters. Also, percent backwash water required is significantly less than that used in conventional filtration technologies (typically 2 to 3% versus 6 to 15%).

Expression for critical suspended particle size in granular media filtration processes

The critical suspended particle size is defined as the suspended particle size with a minimum collector efficiency in a packed bed filter. Identification of the critical size particles is important when a granular filtration process is designed for removal of a mixture of suspended particles with different particle sizes. This note intends to derive a simple expression from the filtration model of Rajagopalan and Tien [Rajagopalan, R. and Tien, C. (1976) Trajectory analysis of deep-bed filtration with the sphere-in-a-cell porous media model. AIChE J. 2(3), 523–533.] for the critical particle size and to investigate the effects of temperature, suspended particle density, collector diameter and approach velocity on the critical particle size and the corresponding collector efficiency.

Media properties and their effect on filter performance and backwashing

Granular filter media used in water treatment have a range of physical properties. Media are usually selected on the basis of size, but bed behaviour is also affected by density and voidage of the media, particularly during backwashing. The fluidisation and backwashing bed behaviour of a number of filter media have been investigated. It has been found the fluidisation is affected by the bed packing arrangement which is determined by previous treatment. Filter media may be subject to attrition during backwashing which will affect the grain size and voidage. Combined air and water flow rates for collapse-pulsing are also affected by the media characteristics. When collapse-pulsing is used to clean triple media filters, complete mixing of the layers occurs, but restratification is possible using a high rate fluidising water wash. Backwashing is an important part of the filtration process so understanding the effect of media properties on the process is essential for optimisation.

Media properties and their effect on filter performance and backwashing

Granular filter media used in water treatment have a range of physical properties. Media are usually selected on the basis of size, but bed behaviour is also affected by density and voidage of the media, particularly during backwashing. The fluidisation and backwashing bed behaviour of a number of filter media have been investigated. It has been found the fluidisation is affected by the bed packing arrangement which is determined by previous treatment. Filter media may be subject to attrition during backwashing which will affect the grain size and voidage. Combined air and water flow rates for collapse-pulsing are also affected by the media characteristics. When collapse-pulsing is used to clean triple media filters, complete mixing of the layers occurs, but restratification is possible using a high rate fluidising water wash. Backwashing is an important part of the filtration process so understanding the effect of media properties on the process is essential for optimisation.

DE filtration to remove Cryptosporidium

DE filtration is far more effective than conventional or direct granular media filtration in reducing concentrations of Cryptosporidium oocysts.Laboratory‐scale testing investigated the degree of Cryptosporidium oocyst reduction provided by diatomaceous earth (DE) filtration. Three grades of DE from two manufacturers were used in the tests. Reduction was measured using seeded river water applied to Walton test filters. Tests were run for filtration rates of 1 and 2 gpm/sq ft (2.4 and 4.9 m/h). Each run was replicated three times, and quality control was rigorous. Approximately 6 logs of reduction in the concentration of Cryptosporidium oocysts can be expected under routine operating conditions using DE grades having permeability of <1.2 Darcy. Log reductions varied predictably according to the permeability of the DE grade and the filtration rate; they were significantly higher at the higher filtration rate. Agreement was excellent between runs, enabling researchers to distinguish with consistent statistical significance the differences in performance.

Flow and filtration through granular media—the effect of grain and particle size dispersion

A mathematical model of granular media filtration has been constructed based on an adaptation of the Carman-Kozeny equations for flow in porous media to the cells of a non-homogeneous media, and a consideration of the capture probability when a particle passes close to a filter grain. The model was based on a matrix of five size fractions and five pore voidages. The suspended solids also comprised five size fractions. The shear stress on the deposit was calculated at each time increment in each cell for each size fraction, and if this exceeded given limits either deposition was inhibited or the solids were flushed out. All coefficients used have a physical significance. The resultant calculations produce a striking simulation of the behaviour found in real filters such as the formation of “wormholes” in the clogged layers, breakthrough when the flow is increased on used filters, a linear increase in headloss with time, etc. A new form of maturation resulting from the redistribution of flow in the matrix has also been identified. The work was undertaken with the aim of producing a general model that might be used by designers to predict all aspects of filter behaviour. While there may be much more that could be done some progress would appear to have been made.

Assessment of tertiary treatment technology for water reclamation in San Francisco, California

Tertiary treatment processes were tested to demonstrate their ability to produce effluents that meet the most stringent standards in the California Wastewater Reclamation Criteria. The processes studied were in-line, granular media filtration and microfiltration for particulate removal, and chlorination and UV light for disinfection. In-line filtration met CWRC turbidity requirements, although filter performance was dependent on feed water turbidity, and alum and polymer doses. Microfiltration produced an effluent virtually free of suspended solids and coliforms, and a 1.9-log removal of MS2 phage. This decreased the level of microorganism inactivation required of the UV system. MS2 phage sensitivity to UV light was not constant, which may question its suitability as an indicator organism to assess virus inactivation. Minimum chlorine Ct value of 340 mg.min/l, and UV dose of 800 J/m2 were identified for in-line filtration effluent, while a dose of 450 J/m2 was sufficient for microfiltration effluent. In-line filtration followed by UV disinfection is less costly for water reclamation than any other combination of treatment units tested, and, therefore, these processes were selected for a 53.000 m3/d water reclamation plant in San Francisco.

Wastewater effluent reuse by in-line flocculation filtration process

Granular media filtration is an important treatment process for various reuse applications. Size density, and attachment strength of filter deposits produced by the interaction of effluents with alum, alum-polymer aids and cationic polymers as primary flocculants and the filter media during contact filtration were calculated and compared using existing filtration models. Secondary effluent filtration without flocculant addition was relatively ineffective. Effective alum dose for contact filtration ranges 10 - 20 mg/L. The 10 mg/L is superior, while the latter creates bulky deposit. An addition of 0.05 - 0.1 mg/L of low anionic high molecular weight polymer strengthens the alum-particle bond. Grain size influence seems to be more pronounced at low approach velocities (5-10 m/hr). High cationic, medium to high molecular weight polymers performed equally or better than alum as primary flocculants. High molecular weight cationic polymer-effluent particle deposit has a higher attachment strength than that of the alum, resulting in a smaller detachment with increase of velocity or increase of filtration run length. Cationic, high molecular weight polymers are effective at doses as low as 0.5 mg/L, at different charge densities. High cationic, medium molecular weight polymers are effective at doses > 5 mg/L. The pressure gradient increase and the filtration efficiency is proportional to the polymer dose in the (1-7 mg/L) range. Compared to the high molecular weight polymer, the pressure build up is milder. High cationic, low molecular weight polymers are not effective at doses up to 7 mg/L. The pressure build up is slow following a slow ripening. In high particle loadings, the differences among the various treatments are less pronounced.

Assessment of tertiary treatment technology for water reclamation in San Francisco, California

Tertiary treatment processes were tested to demonstrate their ability to produce effluents that meet the most stringent standards in the California Wastewater Reclamation Criteria. The processes studied were in-line, granular media filtration and microfiltration for particulate removal, and chlorination and UV light for disinfection. In-line filtration met CWRC turbidity requirements, although filter performance was dependent on feed water turbidity, and alum and polymer doses. Microfiltration produced an effluent virtually free of suspended solids and coliforms, and a 1.9-log removal of MS2 phage. This decreased the level of microorganism inactivation required of the UV system. MS2 phage sensitivity to UV light was not constant, which may question its suitability as an indicator organism to assess virus inactivation. Minimum chlorine Ct value of 340 mg.min/l, and UV dose of 800 J/m2 were identified for in-line filtration effluent, while a dose of 450 J/m2 was sufficient for microfiltration effluent. In-line filtration followed by UV disinfection is less costly for water reclamation than any other combination of treatment units tested, and, therefore, these processes were selected for a 53.000 m3/d water reclamation plant in San Francisco.

Wastewater effluent reuse by in-line flocculation filtration process

Granular media filtration is an important treatment process for various reuse applications. Size density, and attachment strength of filter deposits produced by the interaction of effluents with alum, alum-polymer aids and cationic polymers as primary flocculants and the filter media during contact filtration were calculated and compared using existing filtration models. Secondary effluent filtration without flocculant addition was relatively ineffective. Effective alum dose for contact filtration ranges 10-20 mg/L. The 10 mg/L is superior, while the latter creates bulky deposit. An addition of 0.05 - 0.1 mg/L of low anionic high molecular weight polymer strengthens the alum-particle bond. Grain size influence seems to be more pronounced at low approach velocities (5-10 m/hr). High cationic, medium to high molecular weight polymers performed equally or better than alum as primary flocculants. High molecular weight cationic polymer-effluent particle deposit has a higher attachment strength than that of the alum, resulting in a smaller detachment with increase of velocity or increase of filtration run length. Cationic, high molecular weight polymers are effective at doses as low as 0.5 mg/L, at different charge densities. High cationic, medium molecular weight polymers are effective at doses > 5 mg/L. The pressure gradient increase and the filtration efficiency is proportional to the polymer dose in the (1-7 mg/L) range. Compared to the high molecular weight polymer, the pressure build up is milder. High cationic, low molecular weight polymers are not effective at doses up to 7 mg/L. The pressure build up is slow following a slow ripening. In high particle loadings, the differences among the various treatments are less pronounced.

Particle characteristics and headloss increase in granular media filtration

Size density relationships for aggregated particulates in suspension are transferred into a model describing the accumulation of particulate deposits in the pore space of granular media filters. Using data from several shallow filter layer experiments, the deposit density and the actual pore volume occupied by the captured particulates were estimated for solids of different characteristics. Based on extension of existing headloss models, the effects of particulate size, particulate density, filtration rate, and media grain size on headloss development during particle deposition were evaluated.

富栄養化堀水の粒状層ろ過の性能に及ぼす設計操作条件の影響

Characteristics and performance of granular media filtration of eutrophicated moat water were studied in the bench-scale equipment to determine the design and operating parameters (AL/T, medium size, filtration rate) for optimum filter operation.The reasonable aluminum/turbidity ratio (AL/T) was about 0.01 in both terms of achieving sufficient stability of the filter performance and lessening the influence of residual aluminum salt on fish and shellfish in the moat. The 1.2mm size medium had the highest SS removal ratio, but the weight of removed SS was the lowest due to clogging in the short term. The 2.4mm size medium was suited for the removal of the SS which were mostly particles larger than 20 μm and the VSS in influent, making long-term filtration possible. The filtration process of 2.4mm media was expressed in a filtrate equation by Mintz and a conservation equation. Higher filtration rates decreased the SS removal ratio more rapidly and reduced the ultimate deposit (σu). Judging from the calculation result, the filtration rate should be about 100m·d-1 at the high SS concentration of influent in summer and should be about 100-175m·d-1 at the low SS concentration of influent in winter.