Reading 1

1

Reading Materials-1

on

Coagulation, Flocculation and Sedimentation

1. Reading Materials-1 is a compilation/extraction of information from various sources. All the copyrights remain with the original sources of information regardless of whether they are indicated.

2. It is only for the EN3502/CV4551 class and cannot be circulated/used for any commercial purposes.

2

Chapter 1Water Quality and Treatment Processes 1.1. Quality of water from different sources 1.2 Effect of storage on water quality 1.3 Common tests for water quality 1.4 Drinking water standards and guidelines 1.5 Selection of water treatment processes 1.6 Effect of common impurities on water treatment 1.7 Latest developments in drinking water quality 1.1 Quality of water from different sources

Raw water from various sources may contain suspended and dissolved inorganic and organic substances and microorganisms. Typical concentrations of common water quality parameters from different sources are given in Table 1.1.

Table 1.1 Typical concentrations of water quality.

3

These impurities may come from natural sources or through contaminants from industrial and municipal wastes. For natural sources, inorganic compounds originate from weathering and leaching of rocks and soils e.g. calcium, magnesium and sodium salts of bicarbonate, chloride, sulphate, nitrate, and phosphate. Trace elements such as lead, arsenic, iron, and manganese may also present. Organic compounds mainly originate from decaying plants and animal matter and from agricultural runoffs. Examples of organic compounds are natural humic materials and synthetic organics from detergents, pesticides, and solvents. 1.2 Effect of storage on water quality Surface runoff from storms carries a large quantity of soil and debris into the water course. When reaching reservoirs or lakes, heavier particles can settle to the bottom and the water becomes clearer with time. Natural purification including oxygen transfer from the atmosphere can improve colour, odour, hardness, and dissolved oxygen of water. The main improvements gained by raw water storage are the reduction in suspended solids and microorganism content (Table 1.2). The main disadvantage of storage is problems associated with excessive algal blooms that could result in taste and odour problems. Other effects of impurities in water are given in Table 1.3

Table 1.2 Effect of storage on water quality (same units as in Table 1.1).

Parameter River Thames (UK)*

Raw water Stored water Colour 19 9 Turbidity 14 3.2 Chloride 30 29 NOx-N 5 0.26 Total hardness 300 259 Phosphate 0.8 0.33 Total dissolved solids 415 360 Alkalinity 207 172 Total coliform (MPN per 100 mL) 20,000 100 * Fair, Geyer and Okun.

Table 1.3 Effects of impurities in water (consumer complaints)

Impurities Parameters Effects Suspended solids Bacteria, protozoans, viruses Some cause disease

Algae Odour, colour, turbidity Silt Murkiness and turbidity Clay Colour, murkiness Colloids Turbidity

Dissolved Cations

Calcium, magnesium Hardness Iron and manganese Hardness and colour Copper, zinc Undesirable taste

Dissolved Anions

Bicarbonate and carbonate Alkalinity Sulphate Detectable taste, laxative Chloride Salty taste Fluoride Tooth mottling

4

Organics Colour, taste, odour, toxicity Dissolved gases Oxygen Corrosion, oxidation agent

Carbon dioxide Acid Hydrogen sulphide Taste and odour Nitrogen None Ammonia Caustic

1.3 Common tests for water quality Common tests used to assess the specific and gross characteristics of water for public water supplies are given in Table 1.4

Table 1.4 Common analyses (Water Quality, TD 365 T252)

Test Definition Purpose of test Dissolved cations Calcium Magnesium Potassium Sodium

Ca2+

Mg2+

K+

Na+

To determine the ionic chemical composition of water and to assess the suitability of water for most alternative uses Dissolved anions

Bicarbonate Carbonate Chloride Hydroxide Nitrate Sulphate

HCO3

-

CO32-

Cl-

OH- NO3

-

SO42-

pH =-log [H+] To measure the acidity or basicity of an aqueous solution

Alkalinity Σ(HCO3-+ CO3

2- + OH-)

To measure the capacity of the water to neutralize acids

Acidity To measure the amount of a basic substance required to neutralise the water

Carbon dioxide CO2 To assess the corrosiveness of water and the dosage requirements where chemical treatment is to be used

Hardness Σ(Multivalent cations)

To measure the soap-consuming capacity and scale-forming tendency of water

Conductivity μS/cm To estimate the total dissolved solids or check on the results of a complete water analysis

Radioactivity C To estimate the presence of radioactive substances

Total organic carbon

TOC To assess the presence of organic matter

Specific organic compounds

To determine the presence of pesticides, solvents, and other organic compounds

Total coliforms MPN/100 mL To determine the number of total and fecal coliform bacteria in water Fecal coliforms

5

1.4 Drinking water standards and guidelines The standards of public water supply generally vary between countries in concentrations of permitted contaminants or parameters and in the number of tests which must be conducted. The US Environmental Protection Agency (EPA) drinking water standards is one of the most stringent while the standards of many countries generally follow that of the World Health Organisation (WHO) drinking water guidelines.

Table 1.5 Drinking water standards and guidelines.

6

1.5 Selection of water treatment processes The evaluation of the raw water quality is important in the selection of a treatment system. Numerous unit processes are available for water purification. The sequence and type of treatment processes used, including pre-treatment, will depend on the raw water source and quality of treated water desired. Some simple methods for treatment based on the complexity of raw water parameters can be classified as follows:

Table 1.6 Treatment systems for different water classes.

7

Class A: No treatment is needed e.g. well water Class B: Disinfection only e.g. well water and upland water Class C: Standard water treatment is required e.g. raw water from lowland rivers and reservoirs. Class D: Special water treatment is required e.g. rural supplies (to remove Fe and Mn), colour removal, trace element removal, industrial water, electronics industry requirement, algae control, organics removal etc. When the level of contaminants is low e.g. in ground water, a simple and more economical treatment can be used. If the concentrations are high, a more complex and expensive process may be needed. For sources with a wide variation of contaminants, both in type and time, the treatment plant must be flexible to treat the extreme conditions as economical as possible. The main objective of a water treatment plant is to provide safe drinking water at a reasonable price. As all water sources contain a broad spectrum of inorganic and organic constituents, they must be removed by appropriate physical and chemical means. A variety of processes are available to separate or remove undesirable constituents from raw water.

Table 1.7

8

1.6 Effect of common impurities on water treatment Turbidity and Colour Turbidity and colour are the most common physical parameters that must be removed during the treatment process. They play a significant influence on the treatment to be applied e.g. the necessary coagulant dose is directly related to the amount of turbidity or colour to be removed. When turbidity and colour in raw water is high, large amount of coagulant is required to ensure coagulation and floc formation. Hence, the plant must have the capacity to store, mix, and dose the coagulant. For highly coloured waters ranging from 200 to 300 colour units (true colour), many plants must apply in excess of 100 mg/L of coagulant to remove organic colour effectively. The floc formed, mainly from organic colour with little turbidity, is light, fluffy and fragile. Hence, it requires a well-controlled flocculation system to avoid excessive mixing and the transport of flocculated water to the sedimentation basins must be careful to avoid breaking the delicate floc. If not removed in the settling basins, the floc can clog the sand filters quickly as it is very gelatinous. On the other hand, raw water that is low in colour but high in turbidity (300 to 500 NTU) will produce heavy turbidity-laden floc. The flocculation system must provide adequate energy to handle the heavy floc or else they may settle in the flocculation basin and cause operational problems. The transport of flocculated water must avoid settling of heavy sludge yet prevent floc break-up. Due to the large amount of sludge produced, the sedimentation basins must be specially designed to store, remove, and dispose the sludge. Biological Parameters The common biological parameters of relevance are bacteria, viruses, and algae. Raw water sources where large populations and industrialized areas are present would have large quantities of bacteria and viruses. The quantity of bacteria and viruses in water are reduced during the treatment process, in proportional to turbidity removal. If 95% of turbidity is removed, it may be assumed that bacterial and viral loadings are similarly reduced. Algae are more important for treatment plant operations. They can cause serious problems by accumulating on the walls of basins, clogging of filters, and causing taste and odour problems. Algae are present in all surface waters especially in still water like reservoirs and lakes. Measures should be taken to eliminate algae before they reach the treatment plant. If that is not possible, pre-chlorination will help to keep the plant free of algae. However, this may cause a more serious problem if the water contains the necessary organic precursors that may react with free chlorine to form trihalomethane. One way to overcome this is to apply high shock doses of chlorine during times of algal bloom. In the treatment plant, algae is removed by coagulation, flocculation and settling. If sufficient numbers reach the filter system, they may cause serious clogging problems.

9

pH pH indicates the degree of alkalinity or acidity of water and reflects the characteristics of the watershed or ground strata which the raw water passed through. High alkalinity and hardness, hence high pH, is associated with raw water passing through limestone. For surface water, where alkalinity is lower, hence a softer water, pH is lower. pH has a significant influence on the reaction of coagulants with raw water. Figure 1.1 shows the coagulants needed for removing 50% of the clay turbidity in a 50 mg/L kaolin suspension.

Fig. 1.1 Coagulant dose required to remove 50% of clay turbidity from a water sample containing 50 mg/L kaolin.

For maximum effectiveness, alum has a narrow pH range compared to ferric sulphate. In treatment plants using alum as coagulant, the optimum pH is important; or else the coagulant is not effective. During treatment, pH is reduced because of the reaction between the coagulant and alkalinity of the raw water. To avoid corrosion in the distribution system, the pH must be adjusted upwards with lime. As the effectiveness of chlorine is related to pH, (a lower pH is preferred for the formation of hypochlorous acid HOCl than hypochlorite ions OCl), the chlorine is wasted if pH is high. Therefore, chlorine must be applied before the adjustment of pH. Alkalinity Alkalinity must be present in raw water for coagulation to proceed, and for a satisfactory amount of floc formation. Addition of alkalinity may be needed if natural alkalinity in the raw water is low (as in Singapore). The most common coagulant used in treatment is

10

aluminium sulphate or alum Al2(SO4)3. The amount of alkalinity used in reaction with 1 mg/L of Al2(SO4)3 is:

• 0.5 mg/L of alkalinity as CaCO3 or • 0.33 mg/L of quicklime as CaO or • 0.39 mg/L of hydrated lime as Ca(OH)2 or • 0.54 mg/L of soda ash as Na2CO3

Some plants use ferric chloride (FeCl3) as their primary coagulant. The amount of alkalinity used in reaction with 1 mg/L of FeCl3 is:

• 0.92 mg/L of alkalinity as CaCO3 or • 0.72 mg/L of 95% hydrated lime as Ca(OH)2

Laboratory Analysis The standard guide for water analysis is Standard Methods for the Examination of Water and Wastewater (21st edition) published jointly by the American Public Health Association, American Water Works Association and Water Pollution Control Federation (2002). Turbidity – by nephelometric turbidimeter for raw water, settled water, and filtered water every shift. Colour – by photometer. Samples same as for turbidity. pH – by pH meter; for raw water, settled water, filtered water and finished water every 2 hr. Alkalinity – by titration. Iron and manganese – by atomic absorption spectrophotometer. 1.7 Latest developments in drinking water quality (i) Arsenic The US Safe Drinking Water Act requires EPA to revise the existing 0.05 mg/L standard for arsenic in drinking water. On January 22, 2001 EPA adopted a new standard, and public water systems must comply with 0.01 mg/L standard beginning January 23, 2006. (ii) Disinfectants/Disinfection By-Products Rule Disinfection by-products are formed when disinfectants used in water treatment plants react with bromide and/or natural organic matter (i.e. decaying vegetation) present in the raw water. Different disinfectants produce different types or amounts of disinfection by-products. Disinfection by-products for which regulations have been established have been identified in drinking water, including trihalomethanes, haloacetic acids, bromate, and chlorite.

11

EPA has published the Disinfectants/Disinfection By-products Rule to regulate total trihalomethanes (TTHM) at a maximum allowable annual average level of 0.08 mg/L. The trihalomethanes are chloroform, bromodichloromethane, dibromochloromethane, and bromoform. Haloacetic Acids (HAA5) are a group of chemicals that are formed along with other disinfection by-products when chlorine or other disinfectants used to control microbial contaminants in drinking water react with naturally occurring organic and inorganic matter in water. The regulated haloacetic acids, known as HAA5, are: monochloroacetic acid, dichloroacetic acid, trichloroacetic acid, monobromoacetic acid, and dibromoacetic acid. EPA has regulated HAA5 at 0.06 mg/L annual average. Bromate is a chemical that is formed when ozone used to disinfect drinking water reacts with naturally occurring bromide found in source water. EPA has regulated bromate at annual average of 0.01 mg/L in drinking water. Chlorite is a byproduct formed when chlorine dioxide is used to disinfect water. EPA has regulated chlorite at a monthly average level of 1 mg/L in drinking water. (iii) Total coliform The coliforms are a broad class of bacteria which live in the digestive tracts of humans and many animals. The presence of coliform bacteria in tap water suggests that the treatment system is not working properly or that there is a problem in the pipes. Among the health problems that contamination can cause are diarrhea, cramps, nausea and vomiting. EPA set the health goal for total coliforms at zero. Since there have been waterborne disease outbreaks in which researchers have found very low levels of coliforms, any level indicates some health risk. EPA also set a legal limit on total coliforms. Systems must not find coliforms in more than five percent of the samples they take each month to meet EPA's standards. If more than five percent of the samples contain coliforms, water system operators must report this violation to the state and the public. The number of coliform samples a system must take depends on the number of customers that it serves. Systems which serve fewer than 1000 people may test once a month or less frequently, while systems with 50,000 customers test 60 times per month and those with 2.5 million customers test at least 420 times per month. These are minimum schedules, and many systems test more frequently.

12

Chapter 2 Particulates in Water 2.1 Size characteristics and settling velocity 2.2 Particle shape, specific surface, and surface charge 2.3 Stability of colloids Particulates of concern in water treatment are those solids larger than molecules covering a size range from 1 nm to 100 μm. The principal natural sources of particulates in natural water are soil erosion processes and biological activities. Clays are the main component of particulates produced by weathering. Algae, bacteria, and other microorganisms are the predominant types of particles produced biologically. Another source of particulates is from industrial activities e.g. food processing, paints, and minerals processing. These activities also produce undesirable particulate residues which may enter the surface waters through either direct discharge or atmospheric deposition. Because of their fine size, particulates exhibit a large surface area which serves as a potential adsorption sink for accumulation of toxic substances, such as heavy metals and chlorinated hydrocarbons. Ingestion of the particulates may cause acute or chronic toxic effects. The large particulate surface area also causes strong scattering of incident light resulting in high turbidity. Turbidity particles range from about 0.01 to 100 μm in size. The larger fraction is relatively easy to settle or filter but the smaller colloidal fraction from 0.01 to 5 μm is difficult to remove. Their settling times are very slow and they easily escape filtration. Since fine particles smaller than 10 μm settle slowly, efficient removal by sedimentation requires a long time or a large size facility. For effective disinfection of water, these particulates have to be first removed as they could protect pathogens. Particulates present in natural waters are diverse and exhibit a wide range of sizes, shapes, densities, and surface chemical properties. 2.1 Size characteristics and settling velocity The particulate suspension or hydrosol consists of colloids and suspended solids. Colloids are defined as substances which cover a size range from roughly 0.001 to 1 μm. They include organic macro-molecules present in water, biologically produced debris, viruses, bacteria, clays, and inorganic precipitants. Suspended solids are defined as the material, which can be collected on a 0.45 μm membrane filter. They cover the size range from 0.5 to 500 μm. Particulate physical characteristics of interest include particle length, surface area, volume, and mass. In water treatment, turbidity is used to describe particulate suspension, which can be related to known mass, surface area, or concentration of particulates. However, it is possible to have suspensions with similar turbidities but widely divergent size or mass characteristics. Another collective parameter to describe particulate concentration in water is total suspended solids (TSS). It is useful in the size range above 0.5 μm and is greater than 10 mg solids/L in concentration. Below these levels, measurements are not accurate and require large-volume sampling. The methods of size analysis and particulate removal systems are indicated in Fig. 5.1.

13

Figure 2.1 Characteristics of particulates in water.

Typical settling time of particles Table 2.1 lists a number of materials and organisms with their average size and a rough indication of time needed for these particles to settle vertically though water under the effect of gravity alone.

Table 2.1 Effect of decreasing size of spheres on settling time.

Particle diameter Order of size Total surface area* Time required to settle**

10 mm Gravel 0.000314 m2 0.3 sec 1 mm Coarse sand 0.003141 m2 3 sec

0.1 mm Fine sand 0.031419 m2 38 sec 10 µm Silt 0.31419 m2 33 min 1 µm Bacteria 3.1419 m2 55 hr

0.1 µm Colloidal particles 31.419 m2 230 days 0.01 µm Colloidal particles 314.19 m2 6.3 years 0.001 µm Colloidal particles 3141.9 m2 63 years

*area for particles of indicated size produced from a particle 10 mm in diameter with a specific gravity of 2.65. ** based on a sphere with a specific gravity of 2.65 to settle 300 mm.

14

As shown, colloidal particles which make up a large part of the pollution and are specific cause of turbidity, settle very slow under natural settling conditions. Any treatment requires a change in conditions such that the particles can coalesce between themselves to form large agglomerates which are easy to remove. This coalescence does not take place naturally, since colloidal suspensions are characterized by specific forces which hold the matter in the dispersed state with a high degree of stability over time. 2.2 Particle shape, specific surface, and surface charge Particle shape: Particulates present in suspension exhibit shapes ranging from nearly spherical to fibrous elongated shape as shown in Fig. 2.2. These terms provide only qualitative description of the various shapes. The particle shape factor ψ is defined as the ratio of surface area of a sphere having the same volume as the particle to its actual area. The shape factors of a variety of different shaped materials are given in Table 2.2.

Fig. 2.2 Shapes of granular materials: (a) spherical (b) rounded (c) worn (d) sharp (e) angular (f) crushed.

Table 2.2 Particle shape factor of granular materials.

Description Particle shape factor, Ψ Spherical 1.00 Rounded 0.98 Worn 0.94 Sharp 0.81 Angular 0.78 Crushed 0.70

15

Specific surface: For a sphere, the ratio of surface area to volume is equal to:

3m

2m d6

6/3d

2d VS

Volumearea Surface

=π

π== (2.1)

Hence, for a spherical particle of 1 μm in diameter, its S/V ratio is 6 x 106 m2/m3. If the particle has a specific gravity of 1, i.e. a density of 1 x 106 g/m3, then its surface area to mass (S/M) ratio is:

ρ×=

ρ×=

1d61

VS

Massarea Surface (2.2)

gm 6

1011

1016 2

66 =×

××

= −

For a given volume, the sphere has the smallest surface area compared with other non-spherical particles. Hence, the S/V ratio of spheres will also be smaller. The particle shape factor, Ψ, is used to account for the difference in the S/V ratio between non-spherical and spherical particles. Consider an irregular “angular” shaped particle with a Ψ value of 0.78. The S/V and S/M ratios for this angular particle are:

d6.8

d 78.06

d 6

VS

==ψ

=

sd6.8

s d 6

MS

ρ×=

ρψ=

For an equivalent spherical diameter d = 1 μm and specific gravity = 2.5, S/V ratio of the angular particle = 8.6 x 106 m2/m3 and S/M ratio = 3.4 m2/g. Surface charge: As indicated in Table 2.1, the surface area of colloids is very large (relative to its mass). As a result, surface properties of colloids, such as electrostatic charges, become more important. The charge may vary in magnitude, and it may be positive or negative, according to the nature of the colloid. Metallic oxides (e.g. aluminium, iron) are mostly positively charged, while non-metallic oxides, metallic sulfides, clay, organic colour and most proteins are usually negatively charged in water. Particles of like charge repel one another, whereas oppositely charged particles attract one another. Besides surface charge, colloidal particles can be hydrophobic (water-hating) or hydrophilic (water-loving). Colloids usually encountered in water treatment are mostly hydrophobic. Metallic oxide and non-metallic oxide are hydrophobic colloids. Soap, detergents, soluble starch, and soluble proteins are examples of hydrophilic colloids. Stability of these colloids (hydrophobic and hydrophilic) depends upon the hydration and the electric charge on their surface.

16

2.3 Stability of Colloids In the case of hydrophilic colloids, they are stabilized by the formation of adherent thick layers of oriented water molecules around the particle. For hydrophobic colloids, they are stabilised by an electrostatic repulsion between the particles arising from ions that are attracted to the colloid surface from the bulk solution. Colloids found in natural waters are typically negative charged on their surfaces. When a colloidal particle is immersed in water, electrical charges develop at the particle-water interface. For electro-neutrality to occur, a layer of oppositely charged ions from the bulk solution, is attracted to the particle surface. A portion of the counter-ions remain in a compact “Stern” layer on the colloid surface. The remainder of the counter-ions extends into the bulk of the solution as the diffuse layer.

Figure 2.3 Diffuse double layer.

These two layers of counter-ions surrounding the particle are called the electrical double layer. The magnitude of the charge on a colloid cannot be measured directly, but the value of the potential at some distance from the colloid can be computed. This potential is know as zeta potential. The magnitude of the zeta potential is an approximate measure of colloidal particle stability. Low zeta potentials indicate relatively unstable systems (particles tend to coagulate) while a high zeta potential represents strong repulsive force and a stable system (particles tend to suspend). A suspension of particles held apart by electrostatic repulsion is said to be stable: the particles will not combine to form large aggregates even if brought into contact by vigorous mixing. We can reduce the repulsive force or destabilising the particle suspension by changing the solution chemistry.

17

Instability of Colloids: Instability of colloidal particles is caused by Brownian motion and van der Waals forces of attraction. Due to Brownian motion, particles come close to each other, so that the influence of interacting forces may become effective. The van der Waals force is significant only when particles are very close to one another; otherwise it is small compared to the repulsive force. Thus, whether particles will repel or attract one another depends on the resultant of the electrostatic repelling and the van der Waals forces. If the electrostatic repelling force is substantial, it gives rise to an energy barrier that must be overcome before the particle can approach closely enough to adhere to one another (see Figure 5.4). The zeta potential can be lowered by coagulation through the addition of ions of opposite charge to the particle surface.

Figure 2.4 Forces acting on a colloid (Qasim et al.)

18

Chapter 3 Coagulation (of Colloids) 3.1 Coagulation theory 3.2 Chemical agents for coagulation 3.3 Coagulant (Flocculation) aids 3.4 Optimum conditions for coagulation Coagulation is defined as the process that causes a reduction of repulsion forces between particles or the neutralisation of the charges on particles. In the coagulation process, chemicals called coagulants are added to the water that either breaks down the stabilising forces, enhancing the destabilising forces, or both. Traditionally, metal salts such as aluminium sulphate or alum, ferric sulphate, ferric chloride, and ferrous sulphate have been utilised as coagulants. In recent years, polymers have been used in conjunction with, or in lieu of, metal salts to enhance the coagulation process. For hydrophilic colloids, the coagulation process is hindered by the protective layer on the colloids. A higher amount of coagulant is required than for hydrophobic colloids. Raw waters having suspended solids concentrations greater than 30 to 50 mg/L normally required coagulation. In practice, coagulation is used in all water treatment plants that employ rapid sand filtration. 3.1 Coagulation Theory In water treatment, chemical coagulation is usually accomplished by adding trivalent metallic salts such as Al2(SO4)3 (aluminium sulphate or alum) or FeCl3 (ferric chloride). Although the exact method by which coagulation is accomplished cannot be determined, four mechanisms are often thought to occur: Double layer compression: the quantity of ions in the water surrounding a colloid has an effect on the decay function of the electrostatic potential. A high ionic concentration compresses the layers of counter-ions surrounding the colloid. If this layer is sufficiently compressed, then the van der Waals force will be predominant and the energy barrier will reduce. However, the dose of aluminium or ferric salts used in water treatment is not sufficient to increase the ionic concentration of water drastically to cause ionic or double layer compression. (Fig. 3.1 b-c; Fig. 3.2b; Fig. 3.3) Counter-ions adsorption for charge neutralisation: The counter-ions from the coagulant can also be adsorbed onto the surface of the colloidal particles (Fig. 3.1d). The repulsive charges on the particle surface may be fully neutralised and the zeta potential reduced. The destabilised particles can adhere to each other forming floc. However, the net charge on the particle may be reversed by the adsorption of an excess of counter-ions. Enmeshment in precipitated solids: The dosage of metal salts used in coagulation is usually slightly in excess of the amount required for reduction of the zeta potential. The excess metal salts hydrolyse into hydroxides which are extremely insoluble in water. As the hydroxide precipitate forms (Al(OH)3(s), Fe(OH)3(s)) and accumulates, small colloidal particles are entrapped or enmeshed in the sticky flocs and settle. This process by which colloids are swept from suspension is known as sweep-floc coagulation. The process can be enhanced when the colloidal particles themselves serve as nuclei for the formation of the precipitate.

19

Interparticle bridging: This occurs when a large polymer molecule is adsorbed onto the surfaces of separate particles. The resulting structure grows into a single particle several times larger than the individual colloids (Fig. 3.2c). An excess dosage of polymer may cause restabilisation of the destabilised particles.

Figure 3.1 (a) zeta potential (b) – (c) reduction in zeta potential due to

compression of ion layer (d) adsorption and charge neutralisation.

Figure 3.2 Coagulation of colloids (a) stable suspension of particles where repulsive force

greater than attraction force (b) addition of coagulant suppresses double layer charge (c) agglomeration of destabilised particles by coagulant and polymer bridging.

20

Figure 3.3 Forces acting on a colloid after compression of double layer (Qasim et al.).

3.2 Chemical agents for coagulation The coagulants most widely used for removing turbidity, colour, taste, odor, bacteria and surface charge of particles in water treatment are compounds of iron and aluminium, and polymers: Alum (hydrated aluminium sulphate) Ferric (or ferrous) chloride or sulphate Organic polymers Coagulation using Aluminium Sulphate Aluminium sulphate or alum is available in dry or liquid form. The dry form is more common: powdered, granular, or in lump form. It contains a variable amount of water of crystallisation (Al2(SO4)3 .x H2O where x is usually ranged from 12 to 16) with x =14 being most common. When alum is added to water, it hydrolyses and produces sulphuric acid as well as the aluminium hydrate e.g. aluminium hydroxide, as shown in the following hydrolysis reaction:

Al2(SO4) 3 + 6H2O ⇔ 2Al(OH)3 ↓ + 3H2SO4 (3.1) The water must contain sufficient amount of natural or added alkalinity to react with the acid as it forms, to keep the resultant pH within the desired range. When alum is added to water that has sufficient alkalinity, it will hydrolyse into complex hydroxides, and the actual hydroxides formed depend on the chemical composition of the water particularly its pH and the coagulant dose. If the pH is too low, a soluble compound (Al.OH.SO4) will form. If the pH is too high, the aluminium becomes complexed into soluble aluminate ions (AlO2

-).

21

The optimum pH range for alum is from about 4.5 to 8.0, since aluminium hydroxide is relatively insoluble within this range as shown in Fig. 3.4. The square box in the figure is the usual operating region used in water treatment. The typical dosage of 30 mg/L alum as Al2(SO4)3 in coagulation is approximately 10-4 M, and is well within the solid region of Al(OH)3 (s) at pH values of 6 to 8. For Al(OH)3 (s) to occur, alum dosages must be greater than 10-6 M (0.3 mg/L alum) and pH between 6 and 8.

Figure 3.4 Equilibrium solubility domain of aluminium hydroxide in water (Qasim et al.).

The aluminum hydroxide which is insoluble and colloidal, forms the floc. The resulting floc will indicate if the correct amount of coagulant was used: large, feathery flakes generally indicate that the amount of coagulant was too great. Floc particles of the size of a pin head are normally desirable. Fig. 3.5 shows the solubility of aluminium and iron in water. Aluminium is least soluble within a pH range of 5.7 to 6.2. Note that the allowable limit of aluminium in drinking water is 0.2 mg/L (0.3 mg/L for iron). This figure shows the pH range at these allowable limits is 5.5 to 7.5 for aluminium and 4 to 13.5 for iron.

Figure 3.5

22

When aluminium salt is added to water in quantities less than the solubility limit of the hydroxide, the hydrolysis products formed will adsorb on the colloidal particles. Adsorption of the hydrolysis products will cause destabilisation by charge neutralisation. Example: high turbidity water with low alum dosages – adsorption and charge neutralisisation will be the predominant mechanism. The amount of alum added in a conventional water coagulation process usually far exceeded the solubility limit of aluminium hydroxide. The hydrolysis products will then form precipitates. In this case, charge neutralisation and enmeshment in the precipitate both act to destabilise and coagulate the colloids. Example: low turbidity water with high alum dosage – the predominant turbidity-removal mechanism is sweep-floc coagulation. The simplified chemical reaction to produce hydroxide floc by reacting alum with sufficient alkalinity is: Al2(SO4)3.14H2O + 3Ca(HCO3)2 → 2Al(OH)3↓ + 3CaSO4 + 6CO2 + 14H2O (3.2) MW: 594 3 x 100 CaCO3 2 x 78 3 x 136 6 x 44 14 x 18 Equation 3.2 predicts that each mg of alum will consume approximately 0.5 mg (as CaCO3) of alkalinity and produce 0.26 mg of hydroxide sludge and 0.44 mg of carbon dioxide. Note that non-carbonate hardness (calcium sulphate) is also produced in the water. If the alkalinity is not sufficient to react with the alum and buffer the pH, it is necessary to add alkalinity to the water in the form of lime, sodium bicarbonate, soda ash, or other similar chemicals. The following are the stoichiometric reactions involving calcium hydroxide (hydrated lime) and soda ash, respectively:

Al2(SO4)3.14H2O + 3Ca(OH)2 → 2Al(OH)3↓ + 3CaSO4 + 14H2O (3.3) Al2(SO4)3.14H2O + 3Na2CO3 + 3H2O → 2Al(OH)3↓ + 3Na2SO4 + 3CO2 + 14H2O (3.4) Slaked lime (milk of lime), Ca(OH)2 is produced by reacting quicklime CaO with water in lime-slaking equipment. Quicklime is available in the dry form as granules or lumps and usually contains 70 to 96% CaO. Coagulation using Iron Salts Ferrous Sulphate: It also requires alkalinity in the form as hydroxide ion (e.g. hydrated lime) in order to produce a rapid reaction. This reaction is an oxidation-reduction reaction – dissolved oxygen in the water is reduced:

2FeSO4.7H2O + 2Ca(OH)2 + 0.5O2 → 2Fe(OH)3↓ + 2CaSO4 + 13H2O (3.5)

23

For this reaction to occur, the pH must be raised to about 9.5 where the ferrous ions are precipitated as ferric hydroxide. Ferrous sulphate and lime coagulation is usually more expensive than alum but the ferric hydroxide precipitate formed is a dense and quick-settling floc. Ferric Sulphate: It also produces a dense and rapid-settling ferric hydroxide floc. Slaked or hydrated lime may be used if natural alkalinity is insufficient. The optimum pH range for ferric sulphate is from about 4 to 12 since ferric hydroxide is relatively insoluble within this range.

Fe2(SO4)3 + 3Ca(HCO3)2 → 2Fe(OH)3↓ + 3CaSO4 + 6CO2 (3.6) Ferric Chloride: The simplified reaction of ferric chloride with natural bicarbonate alkalinity to form ferric hydroxide is given in Eq. 3.7 and that with slaked lime is given in Eq. 3.8. The optimum pH range for ferric chloride is also from about 4 to 12 and the floc formed is generally dense and rapid-settling.

2FeCl3 + 3Ca(HCO3)2 → 2Fe(OH)3↓ + 3CaCl2 + 6CO2 (3.7)

2FeCl3 + 3Ca(OH)2 → 2Fe(OH)3↓ + 3CaCl2 (3.8)

3.3 Coagulant (Flocculation) aids Coagulant aids do not aid in coagulation but rather in flocculation. Floc formed in many waters with alum is light and fragile and difficult to settle. The use of coagulant aids can produce larger floc; help to reduce the dosage of metal coagulant; and removing organic colour from water. The agents used include: (a) oxidising agents (b) weighting agents (c) polyelectrolytes. They are normally added after applying coagulants, from 5 to 600 seconds after mixing. Oxidising agents are thought to improve the coagulation-flocculation process by oxidation of organic compounds and iron and manganese which otherwise might interfere with normal coagulation process. Oxidation with air and chemical oxidants such as chlorine, ozone, and potassium permanganate can aid coagulation by reducing coagulant requirements. Weighting agents are used in waters of low turbidity. The addition of materials like bentonite clay increases the particle density, the average weight of suspension, and provides a considerable surface for the adsorption of organic compounds. Weighting agents like activated silica, a short chain polymer, can react with alum to form a tougher and more settleable floc at dosage of between 5 and 10 mg/L. Polyelectrolytes, on the other hand, are long chain polymers which can be used as coagulants and as a coagulant aid. The bridging effect can link active sites of a large number of colloids and combining them into a large particle which can be as much as 100 times than that produced by metallic coagulant alone. Dosage of polyelectrolyte can be as low as 0.1 mg/L.

24

3.4 Optimum conditions for coagulation The chemical parameters involved in the coagulation process are: choice of coagulant, dosage, and pH value of water. Physical parameters include reaction time and mixing intensity of the chemicals. The chemical reactions give an idea as to how the reactions proceed, but the equations cannot be used to predict the actual amount of chemicals required owing to the complexity of natural particle surfaces and the complexity of coagulant chemistry. Therefore, selection and optimum dosages of coagulants are determined experimentally by the “Jar test” instead of quantitatively by equations. Jar test is performed with several (usually six) 1-L beakers and a mechanical stirring device. Each of the six jars is filled with 1-L raw water whose colour, turbidity, pH, and alkalinity have been predetermined. One jar is used as a control and the remaining five are dosed with different amounts of coagulant. The water is mixed rapidly for 1 minute to ensure complete dispersion of the chemicals, and then slowly mixed for 15 to 20 min to aid in the formation of flocs. The mixture is allowed to settle and the supernatant is tested for the quality parameters. The optimum dose for maximum removal of colour and turbidity is selected. The pH may also vary and the optimum value for effective coagulation determined by repeating these steps.

Figure 3.6 Typical jar test results using alum.

25

Measure of effectiveness of coagulation : CF/Co = Final turbidity (or SS concentration) /Turbidity (or SS conc.) before treatment A typical result of CF/Co versus coagulant dose is shown in Figure 3.7. Zone I: stable suspension Zone II: destabilised suspension Zone III: destabilised suspension (charge reversal) Zone IV: enmeshment

Figure 3.7 CF/Co versus coagulant dose and corresponding EPM values.

Note: Surface charge of particles can be measured indirectly by observing the velocity of particle movement in an electric field. The measurement is expressed as EPM and is proportional to surface charge at the shear plane:

Electrophoretic Mobility, EPM = velocity/(electric field strength) = [µm/s]/[Volts/cm]

26

Chapter 4. Mixing Systems for Coagulation and Flocculation 4.1 Reactors 4.2 Hydraulics of mixing 4.3 Velocity gradient 4.4 Rapid mixing systems for coagulation 4.5 Flocculation process and systems 4.1 Reactors Reactors are tanks in which physical, chemical, and biochemical reactions take place. They are classified based on their flow characteristics and mixing conditions. In batch reactors, materials are added into the tank and mixed for a sufficient time to allow reaction to occur and then drained. Although the reactor is well mixed, and the contents are uniform at any instant of time, the composition within the tank changes with time as the reaction progresses. Thus, a batch reaction is unsteady (i.e. it changes with time). In flow reactors, materials flow into, through, and out of the reactor at all times. They are further classified based on mixing conditions as completely mixed flow reactors where mixing conditions are uniform throughout the tank, and plug-flow reactors in which the content passes through the tank in sequence. Reactor Design Equations (i) Completely mixed batch reactor (CMB)

First order reaction: ii kC

dtdC

−= (4.1)

CE = CI e–k t (4.2)

where CE = concentration of effluent

CI = concentration of influent k = reaction rate constant (depending on temperature and pressure)

Time required to achieve a given value of CE : ⎟⎟⎠

⎞⎜⎜⎝

⎛=

E

ICCln

k1t (4.3)

V, Ci

27

(ii) Plug flow reactor (PFR)

First order reaction: CE = CI e– k t (4.4)

Time required to achieve a given value of CE : ⎟⎟⎠

⎞⎜⎜⎝

⎛=

E

ICCln

k1t (4.5)

(iii) Completely mixed flow reactor (CMFR or CSTR)

First order reaction: tk1

CC IE += (4.6)

Time required to achieve a given value of CE : ⎟⎟⎠

⎞⎜⎜⎝

⎛−= 1

CC

k1t

E

I (4.7)

Note: For CMB and PFR: CE = CI e– k t (for 1st order reaction)

% removal = 100 ⎟⎟⎠

⎞⎜⎜⎝

⎛−

I

ECC1 = 100 )e1( kt−− (4.8)

For CMFR or CSTR (completely mixed stirred tank reactor)

tk1

CC IE += (for 1st order reaction)

% removal = 100 ⎟⎟⎠

⎞⎜⎜⎝

⎛−

I

ECC1 = 100 ⎟

⎠⎞

⎜⎝⎛

+−

kt111 (4.9)

4.2 Hydraulics of mixing For time-dependent reactions, the time that a fluid particle remains in the reactor affects the reaction complete time. In ideal reactors, the time in the reactor (known as hydraulic detention time or retention time) is defined as:

QV t = (also HRT, θ, td, tR etc.) (4.10)

where t = theoretical detention time, s V = volume of fluid in basin, m3

Q = flow rate into basin, m3/s

Q CI

Q CE

Δx

28

In actual reactors, they do not behave as ideal reactors because of: density differences due to temperature or density, short circuiting because of uneven inlet or outlet condition, and local turbulence or dead spots in the tank corners. As a result, the detention time in real tanks is generally lower than the theoretical detention time using Eq. 4.10. The actual detention time of water particles in a basin can be determined by dye tracer test. A pulse of dye tracer is introduced in the influent during steady-state flow and the concentration of the dye in the effluent is measured over an extended period of time. The mean residence or detention time can be calculated from:

ii

ii i50 t C

ttCtΔ∑Δ∑

= (4.11)

where t50 = mean residence time (time to centroid) ti = time from injection of dye to collection of effluent sample Ci = concentration of dye in effluent sample collected at ti Note: if samples are taken at equal time intervals, Eqn. 4.11 is simplified to:

CtCt

i

i i50 ∑

∑= (4.12)

Fig. 4.1 Residence time-distribution (RTD) curve (Hammer & Hammer, 2005).

The experimental mean residence time t50 is often shorter than the theoretical or ideal t as a result of flow dispersion with back-mixing and short-circuiting of flow past the water held in stagnant pockets. 4.3 Velocity gradient Coagulation begins by dissolving chemicals in a rapid-mix or flash-mix tank. Flocculation is the slow mixing process in which destabilised colloidal particles are brought into intimate contact in order to promote their agglomeration. Initial collisions between the colloidal

29

particles occur from Brownian motion and contacts brought about due to differences in settling velocities of heaver and lighter particles. The intensity of agitation required for optimum rapid mixing and flocculation is measured by the velocity gradient G value. The G value concept was developed by Camp and Stein (1943) for mechanical or pneumatic agitation. G is related to the power dissipation per volume of water in a mixing unit:

5.0

V PG ⎟⎟

⎠

⎞⎜⎜⎝

⎛μ

= (4.13)

where G = mean velocity gradient, m/s/m or s-1

P = power imparted to the water, N-m/s or W μ = absolute viscosity of water, N.s/m2

V = volume of water, m3 Figure47.2 shows the relationship of G to speed for a stirrer commonly used in jar testing:

Fig. 4.2 Relationship of applied velocity gradient to rotational speed and water temperature (Wagner & Pinheiro)

4.4 Rapid mixing systems for coagulation Coagulant chemicals can be mixed by several methods: mechanical devices in a dedicated basin; in-line blending; air mixing; and hydraulic mixing. Mechanical mixing: propeller- or paddle-type mechanical mixers in a dedicated basin are the most commonly used rapid mix systems in water treatment plants. The water is violently

30

agitated to ensure uniform mixing of the chemicals added. This operation is termed flash- or rapid-mixing. Design parameters for rapid-mix units are power dissipation or head loss incurred for chemical dispersion, detention time in the mixing unit, t, and velocity gradient, G, which is a measure of mixing intensity. Typical design values for most mechanical rapid mix systems provide detention times of 10 to 60 s and G values of 600 to 1000 s-1. A relationship has been developed for the power requirements for turbulent flow conditions where Reynolds number Re > 104. The power number depends on the type of impellers, blade number, and blade width to diameter ratio.

P = Np ρ n3 d5 for Re > 104 (4.14)

Reynolds number μ

=ρnd Re

2 (4.15)

where P = power imparted to water, N-m/s or W

Np = power number of impeller, a constant based on the type of impeller d = impeller diameter, m n = impeller speed, revolutions per sec, rps ρ = mass density of water, kg/m3 µ = water viscosity of water, kg/m.s

Table 4.1 Power numbers of various impellers (Qasim et al.)

Flow Impeller type Power number, Np

Radial

Straight blade turbine 4 blade (w/d = 0.15) 2.6 Straight blade turbine 4 blade (w/d = 0.2) 3.3 Disc turbine 4 blade (w/d = 0.25) 5.1 Disc turbine 6 blade (w/d = 0.25) 6.2

Axial

Propeller 1:1 pitch 0.3 Propeller 1.5:1 pitch 0.7 45o pitched blade 4 blade (w/d = 0.15) 1.36 45o pitched blade 4 blade (w/d = 0.2) 1.94

31

Fig. 4.3 Radial- and axial-flow impellers (Qasim et al.)

(a) straight-blade turbine impeller (b) disc impeller (c) & (e) pitched-blade axial-flow impeller (d) propeller-type axial-flow impeller

Fig. 4.4 Mechanical rapid-mix systems (Reynolds & Richards).

Table 4.2 Detention times and G of rapid-mixing basins (Reynolds & Richards).

Detention time t (sec) Velocity gradient G (s-1) 20 1000 30 900 40 790

50 or more 700

32

Example A rapid-mixing basin is to be designed for a water treatment plant with a design flow of 16,140 m3/d. The basin is to be square with a depth equal to 1.25 times the width. The velocity gradient is to be 900 s-1 and the detention time is 30 s. Determine: (a) The basin dimensions (b) The input power required (c) The diameter of the propeller if a 4-blade pitched propeller (w/d = 0.2) is employed and

n = 110 rpm (a) V = Q t = 16140 m3/d x 30 sec = 5.6 m3 Dimensions = W x W x 1.25 W = 5.6 → W = 1.65 m Water depth D = 1.25 W = 2.06 m (b) P = µ G2 V = 0.89 x 10-3 x 9002 x 5.6 = 4051 W Select motor power = 4.05/0.8 = 5.1 kW (say, efficiency of motor = 80%) (c) For 4-blade pitched propeller (w/d = 0.2): Np = 1.94 (Table 7.1) n = 110 rpm = 1.83 rps d = (P/Np ρ n3 )1/5 = (4051/(1.94 x 103 x 1.833)1/5 = 0.81 m width of propeller blade w = 0.2 x 0.81 = 0.16 m In-line Blending: like mechanical mixers, it allows instantaneous mixing of chemicals at short detention times. Suggested G values are 3000 to 5000 s-1 and detention time of 0.5 sec.

Fig. 4.5 In-line jet mixing system.

Air mixing: such basins employ aeration devices to provide the necessary agitation. The detention times and velocity gradients are of the same magnitude and range as that in mechanical mixing. Velocity gradient may be varied by varying the air flow rate. By selecting a G value, the power required can be obtained by Eq. 4.16. The air flow rate to impart the desired power to the water can be determined from the following equation:

⎟⎠⎞

⎜⎝⎛ +

=4.10

4.10hlogG3904P 10a (4.16)

33

where P = power imparted to water, W Ga = air flow rate (m3/min)

h = depth of the diffusers, m

Fig. 4.6 Pneumatic rapid mixing (Reynolds & Richards).

Hydraulic mixing: depends on turbulence to furnish the desired velocity gradient.

Fig. 4.7.Baffle basin rapid mixing (Reynolds & Richards).

The velocity gradient for baffle basins is given by: th G

μγ

= (4.17)

where γ = specific weight of water, N/m3

h = head loss due to friction, turbulence etc., m t = detention time, s The typical G values in coagulation, flocculation, activated sludge process and that in the natural environment are given in Table 4.3.

Table 4.3

Condition G (s-1) P/V (W/m3) Flocculation in water treatment 10 – 100 0.1 – 10 Rapid mixing in water treatment 500 – 1000 250 – 1000 Activated sludge process 100 – 250 10 – 60 Natural streams 10 – 100 0.1 – 10

Note: V = volume of water mixed in m3

34

4.5 Flocculation Process and Systems Flocculation can be simply defined as the aggregation of particles into larger elements. The aggregation of colloidal particles takes place in two separate and distinct phases: (1) the repulsion force between particles must be overcome; this requires that the particles be destabilized (i.e. coagulation); and (2) contact between the destabilised particles must be induced so that aggregation can occur (i.e. flocculation). As mentioned earlier, the destabilisation step is achieved by addition of chemicals to modify the electrochemistry properties on the particle surfaces. This coagulation process step is virtually instantaneous, in milliseconds to seconds, following addition of the coagulant in rapid mix tanks. The aggregation step on the other hand, requires more time for the development of large flocs, by gentle stirring in the flocculation tanks. There are three major mechanisms of flocculation: (a) Brownian or Perikinetic flocculation in which aggregation of particles is the result of

random (Brownian) motion caused by the continuous bombardment by the surrounding water molecules. The driving force for this type of particle movement is the thermal energy of water. It has only a minor influence on transport of particles larger than about 1 μm i.e. it is effective only for particles < 1μm in size. Aggregation of particles by Brownian flocculation is a slow process e.g. for a colloidal suspension of clay particles with a concentration of 108/L, it would take about 20 days for this to be halved (to 0.5 x 108 particles).

(b) Orthokinetic flocculation where aggregation of particles is by induced velocity

gradients in the water. The suspended particles follow the streamlines with different velocity and eventually lead to interparticle contacts. Thus, to accelerate aggregation for coarse sized particles > 1 μm, mechanical mixing must be employed. It is the predominant mechanism in water treatment.

(c) Differential settling which is caused by the different settling velocities of particles. Kinetics of Particulate Aggregation A quantitative description of the particle aggregation process will provide a basis for designing coagulation-flocculation facilities, as well as identifying the key process parameters. Consider a suspension of particles of different sizes and number: di = diameter of particles of size class i ni = total number of particles in size class i N = total number of particles of all sizes Consider particle of size di collide with size dj particles forming particles of size dk. At the same time, aggregates of size dk may break up into smaller aggregates due to hydrodynamic shearing forces. The rate of flocculation is given by the general form:

35

⎥⎦⎤

⎢⎣⎡ βα= jiijij nn21 -

dtdN (4.18)

where αij = attachment efficiency βij = collision frequency The attachment efficiency αij is defined as the ratio of successful to unsuccessful particle collisions, and has a range of values between 0 ≤ α ≤ 1. Successful particle collisions occur when two particles remain attached after the collision. It depends on the effectiveness of destabilization. (α = 1 for perfectly destabilized particles). The collision efficiency βij is a function of the probability that an individual particle will collide with another in a unit of time. It depends on the transport processes. Note that the sign in Eq. 4.18 is negative because each successful event of collision/adhesion reduces the total number of particles N. The value of βij can be estimated for different size classes using different transport mechanisms:

βij = βPK + βOK + βDS (4.19) where βPK = collision frequency due to Brownian motion (micro-flocculation or perikinetic

flocculation) βOK = collision frequency due to shear motion (macro-flocculation or orthokinetic flocculation) βDS = collision frequency due to differential settling (or sedimentation)

The volume of floc particles per unit volume of water i

3i n

6dπ

=Ω (4.20)

where Ω = floc volume fraction in cm3 of particles per cm3 of water

di = particle diameter, cm ni = number of particles per cm3 or mL of water (if particle diameter is in cm)

Macro-flocculation or Orthokinetic Flocculation, OK In water treatment, orthokinetic flocculation predominates for particles > 1 µm. Even in quiescent waters, orthokinetic mechanism also predominates for bigger particles (> 10 µm). In order to accelerate particle aggregation for colloidal particles and achieve acceptable flocculation rates, mechanical mixing must be employed. For spherical particles moving under laminar flow conditions, the collision frequency function for orthokinetic flocculation is given by:

iOK GV8π

=β (4.21)

36

where G = average velocity gradient, s-1 Vi = volume of particles of size di = π d3/6, cm3 if d in cm Substituting Eqn. 4.21 into 4.18 gives the following result for the orthokinetic flocculation (aggregation) rate of spherical particles of a uniform size:

N G 4 dtdN

Ωαπ

−= (4.22)

where Ω = floc volume fraction = volume of floc/volume of water = N Vp = N 6d3π

Micro-flocculation or Perikinetic Flocculation, PK Brownian motion affects the movement of colloidal particles of 1 µm or smaller. For a suspension of single-sized spherical particles, the collision frequency function is given by:

⎟⎟⎠

⎞⎜⎜⎝

⎛μ

=βTk

38 B

PK (4.23)

where βPK = collision efficiency for perikinetic flocculation

kB = Boltzmann constant = 1.38 x 10-23 J/K T = absolute temperature, K (= 273 + oC) µ = dynamic viscosity of water, Ns/m2 or kg/m.s Substituting Eqn. 4.23 into 4.18 gives the following result for the perikinetic flocculation (aggregation) rate:

2B N Tk34

dtdN

⎟⎟⎠

⎞⎜⎜⎝

⎛μ

α−= (4.24)

Differential settling (or Sedimentation) The velocity of particles of similar densities settling in a water column is proportional to the size squared. Thus, differential particle motion occurs in heterogeneous suspensions during sedimentation, providing an additional transport mechanism for promoting flocculation.

]dd][)dd[( 72

g ji

3jiDS −+

μαρΔπ

=β (4.25)

where Δρ = difference in density between the particle and water = ρp – ρw

37

Example After addition of 50 mg/L of alum coagulant, the floc that is formed has a density of 1.02 g/cm3 and a diameter of 0.1 mm. It is treated in a flocculation basin for a hydraulic retention time of 20 min. Assume that the floc particles have a α = 0.3 and velocity gradient in the flocculation basin G = 50 s-1, determine the number and size of floc particles leaving the basin. Note: Assume for floc particles: crystalline Al(OH)3 solids s.g. = 3; water s.g. = 0.997 Consider 1 cm3 floc: mass of floc = mass of Al(OH)3 solids + mass of water Vfloc ρfloc = Vsolids ρsolids + Vwater ρwater 1 cm3 x 1.02 = (Vsolids x 3) + (Vwater x 0.997) Also Vfloc = Vsolids + Vwater → Vwater = (1 cm3 – Vsolids) Therefore 1 cm3 x 1.02 = (Vsolids x 3) + 0.997 (1 cm3 – Vsolids) Vsolids = 0.0115 cm3 Mass of Al(OH)3 solids = 0.0115 cm3 x 3 g/cm3 = 0.034 g Solution:

1 mg alum produces 0.26 mg Al(OH)3 solids or 33

cm1034.0

1026.0×

× −= 7.65 x 10-3 cm3 floc

50 mg/L alum dose produces 50 x 7.65 x 10-3 cm3 = 0.382 cm3 per L of water Ω = floc volume fraction = (0.382 cm3 x 10-3 L/cm3)/L = 0.382 x 10-3

From Eq. 4.20, i

3i n

6dπ

=Ω → I33

3 N6

)101.0(10382.0−

− ××π=×

NI = 7.38 x 108 floc particles per m3 of the coagulated water (particle diameter = 0.0001 m) Assume flocculation basin to be a completely mixed flow reactor (CMFR) or continuously stirred tank reactor (CSTR) where there is a continuous feed of coagulated floc particles entering the basin and a continuous discharge of flocculated particles from the basin. The mass balance equation for the basin is: Particle accumulation = input + flocculation rate in basin – output For steady state, there is no accumulation of particles in basin:

0 = Q NI + VdtdN – Q NE

38

where Q = flow rate, V = flocculation basin volume, NI = influent floc number, NE = floc number in effluent, dN/dt = flocculation rate in basin at effluent NE (More on CSTR: (1) Reynolds & Richards – Unit operations and processes in environmental engineering, 2nd ed., Chapter 3 (2) Crittenden – Water treatment principles and design, Chapter 6) Divide equation by Q and substituting t = V/Q and using Eq. 4.22 (based on first-order reaction rate)

NI – NE = – (t dtdN ) = EN G 4t Ωα

π

Consider 1 m3 of water, 7.3 x 108 – NE = (20 x 60 s) E3- N 10382.03.0504

××××π

NE = 7.08 x 107 particles per m3 water

Efficiency of particle aggregation in basin = INEN1− = 0.9 = 90%

Assume volume of floc per m3 water remains the same after flocculation (only size and number changed), hence Ω = floc volume = = 0.382 x 10-3

From i

3i n

6dπ

=Ω → E3

3 N6

)d(10382.0 ×π=× −

Substituting NE = 7.08 x 107 particles per m3 water, d = 0.218 mm Flocculation Systems: Design of Flocculators The degree of agitation requirements in flocculators is much less than in rapid-mixing systems. The rate of particulate collisions is proportional to G and sufficient gradient must be furnished to achieve the desired rate of collisions. However, G is also related to the shear forces in water: large G produces appreciable shear forces that can prevent floc formation. Typical G values for flocculation are between 10 and 100 s-1. Flocculation basins are normally designed with multiple mixing compartments in a series, with velocity gradients successively lower in each compartment. This is known as tapered flocculation and can produce a more uniform and tough floc that will settle readily. The higher G value in the first compartment allows the formation of a higher density floc. The lower G values in the subsequent compartments avoid breaking up of already formed floc; but can promote the build-up of progressively larger-size floc. The average velocity gradient Geff for the entire flocculation unit can be calculated from the different compartment G values:

39

5.023

22

21

eff 3GGGG

⎥⎥⎦

⎤

⎢⎢⎣

⎡ ++= (4.26)

The detention time (t) in flocculation basins is much longer than that in rapid-mix basins. Typical (t) values are between 20 and 60 min. The key design factor in flocculation basin is the value of Gt (velocity gradient x detention time). The number of particle collisions is directly proportional to the product of G and the detention time, t. Typical Gt values range from 10,000 to 150,000. There are two general groups of flocculation units: hydraulic and mechanical flocculators. The hydraulic flocculators utilise cross-flow baffles or 180o turns to produce the required turbulence. The main design objective is to achieve gentle, uniform mixing that will not shear the floc. They are only used when flow rate is relatively constant and are rarely used in medium- and large-sized water treatment plants. The mechanical flocculators typically used are paddle-wheel mixers, walking beam flocculators, flat-plate turbines, and axial flow propellers or turbines.

Fig. 4.8 Three-stage flocculation basin (Qasim et al.)

Mechanical Flocculators The velocity gradient for mechanical mixing can be calculated from Eq. 4.13. In the case of paddle-wheel mixers, the water power is given by:

2vAC

P3

D ρ= (4.27)

40

where P = power imparted to water, W CD = drag coefficient, depending on length-to-width ratio of paddle blades (For L/W = 5, CD = 1.2; L/W = 20, CD = 1.6; L/W =∝, CD = 1.9) A = total area of paddles, m2 v = velocity of the paddle relative to the water, m/s (about 0.75 of paddle tip velocity) ρ = mass density of water, kg/m3

Figure 4.9.

Fig. 4.10 Flocculation basins using paddle-wheel flocculators.

41

Mechanical Mixers The power imparted to the water by a mechanical mixer at laminar flow conditions (Re < 10) is given by Eqn. 4.28 where Np is from Table 4.1. Re can be determined using Eq. 4.15.

P = Np μ n2 d3 for Re < 10 (4.28) The motor power requirement (P’) is calculated from the efficiencies of the gear box (Egear) and the flocculator bearings (Ebearings):

bearingsEgearsEP'P

+= (4.29)

Hydraulic Flocculation (Baffled flocculation tanks)

The baffles can be arranged to provide flow patterns of 180o turns (Fig. 4.11) or cross-flow baffles (Fig. 4.12). The action of the baffles is to intensify the velocity gradient by enforcing changes in the direction of flow. The characteristics of such tanks are large head losses (0.2 m to 1.8 m) and little short-circuiting. They are cheap to build and maintain but their efficiency is dependent on the flow rate. For such tanks, the head loss (h) through the flocculation tank can be determined from: P = Q ρ g h (4.30) The velocity gradient can be determined based on Eqs. 4.13 and 4.30:

5.0

V h g QG ⎟⎟⎠

⎞⎜⎜⎝

⎛μρ

= (4.31)

Fig. 4.11 Tapered horizontal baffled hydraulic flocculator (Crittenden)

42

Fig. 4.12 Baffle-type hydraulic flocculator. The number of baffles in each maze-type flocculator can be calculated from:

3

12

QG L H

f)(1.44 t 2n

⎥⎥⎦

⎤

⎢⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎟⎠

⎞⎜⎜⎝

⎛+ρ

μ= (4.32)

where: Q = flow rate, m3/s t = flocculation time, seconds ρ = density of water = 1000 kg/m3 µ = absolute viscosity of water = kg/ms f = roughness coefficient of baffles (for timber = 0.3)

H = depth of water in flocculator, m L = length of flocculator, m (the side that is perpendicular to the baffles) G = velocity gradient, sec-1

43

5. Sedimentation 5.1 Introduction 5.2 Types of sedimentation 5.3 Settling of discrete particles 5.4 Ideal sedimentation basin 5.5 Removal rate for a suspension of discrete particles 5.6 Settling of flocculent particles 5.7 Sedimentation Basins 5.8 Potential operational problems 5.9 Useful design parameters References: Reynolds & Richards, “Unit operations and processes in environmental engineering”, 2nd ed., Chap 9. 5.1 Introduction Sedimentation is the process of removal of suspended particles that are heavier than water by gravitational settling. The relevant terms with respect to sedimentation are: (a) Plain sedimentation: refers to the separation of impurities from water by the action of

natural forces alone i.e. by gravitation and natural aggregation of settling particles. Examples are settling of sand during filtration and the settling of grits and sandy and silty particles in pre-treatment. Such particles are usually greater than 10 μm in size.

(b) Coagulation and Flocculation: the addition of chemicals or other substances are added

to induce or hasten aggregation and settling of finely divided suspended matter, colloidal substances, and large molecules. Examples are the removal of colour and turbidity in water.

(c) Chemical precipitation is the addition of chemicals to remove dissolved impurities such

as hardness, Fe, Mn, etc. out of solution. 5.2 Types of Sedimentation The sedimentation process can be used to treat raw water and reclaimed wastewater containing suspensions ranging from a very low concentration of nearly discrete particles to a high concentration of flocculent solids. There are 3 classes of particles according to their characteristics during settling: Class 1: the settling of relatively low concentration of discrete particles that will not readily flocculate or grow in size. An example is the settling of granular particles after backwashing. Class 2: the settling of relatively low concentration of flocculent material. An example is the settling of coagulated water. Class 3: the settling of relatively high concentration of materials. Also known as hindered settling. The material may be flocculent. An example is sludge thickening.

44

Fig. 5.1 Settling characteristics of solids in water (Montgomery)

5.3 Settling of discrete particles (Class 1) A discrete particle is one that does not alter in size, shape or weight when settling. In falling through water in a quiescent condition, such a particle will accelerate until the gravitational force (FG) is equaled by the sum of particle drag (FD) and buoyancy forces (FB):

FG = FD + FB (5.1)

ρs g Vp = 12

CD Ac ρw vs2 + ρw g Vp (5.2)

where ρs = density of particle, kg/m3

ρw = density of liquid, kg/m3 Vp = volume of particle, m3 CD = drag coefficient, dimensionless

Ac = cross-sectional area of particle, m2 vs = velocity of settling particle, m/s g = acceleration due to gravity = 9.81 m/s2 Thereafter, the particle settles at a uniform velocity known as the terminal velocity, which is an important hydraulic characteristic of the particle. The terminal velocity vs of a discrete particle depends on its size, shape, and density, and also the density and viscosity of the liquid. For spherical particles,

Vp = 6d 3π (5.3)

45

Ac = 4d 2π (5.4)

By substituting Eqs. 5.3 and 5.4 into Eq. 5.2, the general equation for sedimentation of discrete spherical particles as described by Newton’s Law is:

w

ws

Ds

)(Cgd

34v

ρρ−ρ

= or ( )1sSDC

gd34

sv −= (5.5)

where Ss = specific gravity of particle = ρs/ρw. The drag coefficient CD is dependent on the settling velocity and diameter of the particle, and the density and viscosity of water, which are represented by the dimensionless Reynolds number, Re:

Re24CD = for Re ≤ 1 (5.6)

34.0Re3

Re24CD ++= for 1 < Re ≤ 104 (5.7)

υ

=dv

Re s (5.8)

where d = particle diameter, m

υ = kinematic viscosity of water, m2/s

Figure 5.2

46

Note: (1) For laminar flow where Re ≤ 1, CD tends to be 24/Re (see Fig. 5.2)

Equation 5.5 becomes υ−

=18

)1S( d gv s

2

s or μ

ρ−ρ=

18)s( 2d g

sv (5.9)

Equation 5.9 is also known as Stokes Law. It is applicable for small diameter particles (up to about 1 mm diameter), and Reynolds number less than 1. (2) Particles in water are not spherical. In such cases, the equivalent diameter d’ is

computed as:

d’ = ψ d (5.10) where d’ = effective spherical diameter to be used in the computations, m d = diameter of actual particle, m ψ = shape factor (3) The effect of an irregular shape is not pronounced at low settling velocities and most

sedimentation devices are designed to remove small particles which settle slowly.

Table 5.1 Settling velocities for particles of s.g. 2.65 in water at 10oC (Tebbutt, 1992).

Particle size, μm Settling velocity, m/hr 1000 600 100 20 10 0.3 1 3 x 10-3

0.1 1 x 10-5 0.01 2 x 10-7

Example (a) Determine the settling velocity of alum (aluminium hydroxide) flocs of diameter d = 1

μm and ρs = 1020 kg/m3. Take υ of water = 0.89 x 10-6 m2/s at 25oC.

Assume laminar flow condition. Applying Eq. 5.9: vs = 12247 d2 (d in metres)

For d = 1 μm, vs = 12247 d2 (d in metres) = 1.22 x 10-8 m/s

Check Re = 10-8 < 1 ok. Laminar flow condition.

47

(b) Determine the settling velocity of alum flocs of diameter d = 1000 μm and ρs = 1020 kg/m3.

Assume laminar flow condition: for d = 1000 μm, vs = 0.01225 m/s

Check Re = 13.8 > 1 Stoke’s law not valid: need to use Eq. 5.5.

(c) Determine the settling velocity of sand grains of diameter d = 1 mm and ρs = 2650 kg/m3.

Assume laminar flow condition: vs = 1010393 d2 (d in metres) = 1 m/s Check Re = 1135 > 1 Stoke’s law not valid: need to use Eq. 5.5.

Note: solving Part (b) and (c) in the above example using Eq. 5.5 will require trial and error. Figure 5.3, which is based on Eq. 5.5 can be used to determine settling velocity for a given particle diameter and specific gravity.

Fig. 5.3 Settling velocity of particles. (Reynolds & Richards).

48

5.4 Ideal sedimentation basin The study on settling of discrete particles can be described using an ideal sedimentation tank such as that shown in Fig. 5.4. There are four distinct zones in the ideal tank: inlet settling, sludge, and outlet.

Fig. 5.4 Ideal sedimentation basin. (Montgomery)

The following assumptions are made in developing the tank’s removal efficiency equations: • horizontal flow in the settling zone • uniform horizontal velocity in the settling zone • uniform concentration of all-size particles across a vertical plane at the inlet end of the

settling zone • particles are removed once they reached the bottom of the settling zone • particles settle discretely without interference from other particles at any depth The theoretical design of sedimentation processes is based on the concept of the ideal settling tank. A schematic tank with flow and dimensional parameters is shown in Fig. 5.5. All particles in the settling zone travel in a straight line path.

Fig. 5.5 Schematic diagram of an ideal rectangular sedimentation basin.

Consider a discrete particle entering the tank at height hs from the base, having a horizontal velocity vh (= Q/L.ho) and a vertical settling velocity vs (determined using Eq. 5.5 or 5.9). The particle must fall a depth of hs during time t it is in the tank in order to be removed (i.e. reaching the sludge zone).

49

For particles entering at the top of the tank, they must fall a distance equal to the depth of the tank ho during time t if they are to be removed. Hence, for removal, the settling velocity of particles

th

v os ≥ (5.11)

Note that Q

LWhQVt o ××== and plane or surface area of tank A = L x W

Therefore, oo

os v

AQ

LWQ

QLWh

hv ==×

=××

≥ (5.12)

where Q = rate of flow, m3/s V = volume of tank, m3

W = width of tank, m ho = depth of water through which particle has to fall, m

L = length of tank, m vo = critical settling velocity of design particle in tank, m/s The term Q/A is known as the surface overflow rate (SOR) or simply overflow rate. The overflow rate of the design tank corresponds to a critical settling velocity vo of a specified particle to be removed by the tank. Note that: • All particles having settling velocity vs greater than vo or SOR will be removed (i.e.

settled in the sludge zone). • The removal efficiency of discrete particles is dependent on the SOR, but not the depth of

the ideal sedimentation tank. • Even those particles whose settling velocity vs is less than vo can be removed in

horizontal flow tanks if they enter the tank at a height hs less than ho (see Fig. 5.5). The proportion of particles with vs less than the design vo but can also be removed in a horizontal-flow ideal sedimentation tank is given by:

Xr = v v

tv tv

hh

o

s

o

s

o

s == (5.13)

where Xr = fraction of particles removed in an ideal tank vs = velocity of specified particle size, m/s vo = critical velocity defined as ideal tank surface overflow rate, m/s

Limitations of sedimentation theory The analysis of an ideal sedimentation basin indicates that for discrete particles, only the plan area of the basin is important. This theoretical conclusion is subject to practical modifications dictated by wind currents, short-circuiting and the need for uniform distribution of flow.

50

5.5 Removal rate for a suspension of discrete particles A large variation of particle size will exist in a typical suspension. Hence, the entire range of settling velocities has to be evaluated in order to determine the overall removal for a given design settling velocity or overflow rate. A batch settling column test, in which samples are withdrawn at various time and a particular depth and the solids concentrations are determined, is employed. From the settling column test results, a graphical procedure is often used to determine an approximation of overall removals of discrete particles in a suspension.

Fig. 5.6 Typical discrete particle settling curve.

Procedure to determine overall removal efficiency of a suspension of discrete particles: (1) Prepare a discrete settling curve as shown in Fig. 5.6 (2) Integrate the area to the left of the curve representing overall removal efficiency:

Fraction removed: Pv

v1)P1(

dP vv

)P1(X

i s

oo

oP

0 o

sor

Δ+−=

+−=

∑

∫ (5.14)

Xr = I

ECC1− (5.15)

where Po = fraction of particles with vs ≤ vo

(1 – Po) = fraction of particles with vs > vo CE = effluent concentration CI = influent concentration

Pvv1 or dP

vv

i s

o

oP

o o

s Δ∑∫ = fraction of particles with vs < vo but are removed

Fraction remained = I

ECC = Po Pv

v1

i s

oΔ− ∑ (5.16)

51

Example Determine the removal efficiency of a discrete suspension in a tank with a surface overflow rate of 4 m/hour. The distribution of particles is given in the table below. ρs = 2650 kg/m3 µ = 1x 10-3 kg/m.s Size interval (μm) 0-10 10-

20 20-30

30-40

40-50

50-60

60-70

70-80

80-100

Weight fraction 0.05 0.2 0.25 0.2 0.15 0.075 0.05 0.02 0.005 Average size, d (μm)

5 15 25 35 45 55 65 75 90

v s (m/s) x 10-3 0.02 0.2 0.56 1.1 1.8 2.7 3.8 5 7.3 v s (m/hr) 0.08 0.72 2 4 6.5 9.7 13.7 18 26.3 P, weight fraction < v s

0.05 0.25 0.5 0.7 0.85 0.925 0.975 0.995 1

Assume laminar flow :μρ−ρ

=18

d)(gv

2s

s = 0.9 x 106 d2 m/s

Check Re for largest sand size d = 100 μm, v s = 0.009 m/s, μ

ρ=

dvR sw

e = 0.9 < 1 ok

Plot graph of settling velocity of particles v s (m/hr) versus the respective weight fraction, P.

Suppose vo = 4 m/hr. From plot, Po = 0.7.

Mass fraction removed = (1 – 0.7) + Pvv1

i s

oΔ∑ ≅ 0.3 + 0.25 = 0.55 or 55%

52

5.6 Settling of flocculent particles (Class 2) Flocculent particles have a tendency to coalesce into a bigger particle during settling. There are two principal causes of flocculation during sedimentation: (1) particles with faster velocities overtake those that settle more slowly and coalesce with

them. (2) velocity gradient within the water that cause particles in a region of a higher velocity to

overtake those adjacent stream paths moving at slower velocities. Flocculent settling has two advantages over discrete settling: (1) the combination of smaller particles to form larger ones results in a faster settling particle

because of increase in size (2) flocculation tends to have a sweeping effect in which large particles settling at a velocity

faster than slow particles tends to sweep some of them from suspension. Tiny particles which otherwise would not settle are removed.

The design of sedimentation tanks in removing flocculent particles can also be based on settling column tests. The depth of the column is normally equal to or greater than the depth of the proposed tank. The diameter of the column is about 150 to 300 mm and sampling ports are provided at 500 mm intervals along the depth. A suspension is poured into the column and gently mixed with a perforated plunger to obtain a uniform dispersion of particles. At predetermined time intervals, samples are removed from the ports and analyzed for suspended solids concentrations. Experimental data of suspended solids removal are then plotted on a time-depth graph as shown:

Fig 5.7 Settling trajectories for a flocculent suspension.

53

The overall removal is given by:

nn1-no

1-n21

o

21

o

1T R)R-(R

hh....)R-(R

hh)R-(100

hhR ++++= (5.17)

where: h1, h2, …., hn-1 = vertical distance from the top of the settling column to the mid-point between two consecutive lines of iso-removal at the desired detention time ho = desired design side water depth R1, R2,….., Rn-1 = consecutive iso-removal curves in percent removal RT = total removal, percent

The accuracy of estimation can be improved by decreasing the interval between iso-concentration lines and adding more terms to the removal equation. The test results allow the overflow rate and detention time for a sedimentation tank to be determined. Non-ideal conditions in actual sedimentation tanks such as short-circuiting, inlet and outlet turbulence, density, and temperature-induced currents result in reduced removal efficiency compared to that obtained from settling column test. To compensate for the non-ideal conditions, a factor of safety equal to 0.65 to 0.85 for the overflow rate and 1.75 to 2 for the detention time are recommended. (Hence, corrections are to have a smaller SOR and a longer retention time). Since flocculent particles tend to grow in size during settling, the depth of the settling tank and detention time are important design parameters. Deeper tanks improve the possibility of bigger particles to sweep smaller particles from suspension. Deeper tanks therefore perform better in removing flocculent particles.

54

Example The results of a column settling test are given in the following table. Find the overall removal for a settling basin 1.75 m deep with an overflow rate of 105 m/d.

Water depth from

surface

Percent removal of particles at various time 5 min 10 min 20 min 30 min 40 min 50 min

0.5 m 14* 17 56 66 74 79 1.0 m 13 16 49 59 71 - 1.5 m 12 17 43 47 70 72

* Percent removal = [(TSSI – TSSF)/TSSI] x 100% where TSSI = initial suspended solids concentration in column at time t =0, mg/L

TSSF = suspended solids concentration in sample after certain time t, mg/L Indicate the numerical percent removals on a plot of time versus water depth as shown in Fig. 5.7. Draw the iso-removal lines at 15% intervals, by interpolating the numerical values. Surface overflow rate = 105 m/d = ho/t For ho = 1.75 m, the required detention time t = 1.75m/(105 m/d) = 24 min From Fig. 5.7, 45% of particles are having settling velocities greater than SOR = 105 m/d The overall particles removal can be computed using: Equation 5.17: h1 = 0.1m, h2 = 0.35m, h3 = 1.1m, ho = 1.75m, R1 = 100%, R2 = 75%, R3 = 60%, R4 = Rn =45%

45)45-(601.751.1)60-(75

1.750.35)75-(100

1.750.1R T +++= = 58.9%

55

5.7 Sedimentation Basins The common shapes of conventional sedimentation tanks in water treatment and reclamation are rectangular, circular, or square. In rectangular basins, the influent flow is distributed across one end of the tank by an inlet baffle structure, which provides energy dissipation and uniform flow distribution. The outlet structure usually comprises a system of effluent launders located on the opposite end of the tank. Circular basins are fed from a central inlet or from the peripheral. Effluent structures are normally consisting of a V-notch weir constructed at the outside perimeter of the tank. Diameters of tanks are calculated from the overflow rates. Square basins have the advantage of common-wall construction as in rectangular tanks. Effluent launders are also constructed along the perimeter of the basins. Few square basins are actually constructed because of difficulties in sludge removal from the corners of the basins.

Fig. 5.8 Types of sedimentation basins (a) Rectangular (b) Square

(c) Circular, centre feed (d) Circular, peripheral feed. Careful design of inlets and outlets is important to assure reasonable performance of sedimentation tanks. The ideal inlet reduces the entrance velocity to prevent pronounced currents towards the outlet. It should distribute the water as uniformly as possible across the width and depth of the tank and mixes it with the water already in the tank to prevent density currents. Outlets frequently consist of free-falling weirs discharging into effluent launders. Weir loading rates are limited to prevent high approach velocities near the outlet. Outlets are placed as far from the inlet as possible and frequently consist of V-notch approximately 50 mm in depth.

Weir overflow rate = length weir Total

Q rate Flow (m3/m.d) (5.18)

56

Figure 5.9 Outlet weirs (Wagner & Pinheiro).

Fig. 5.10 Plan view of rectangular sedimentation basins (Wagner & Pinheiro).

Sedimentation basin innovations Variations in basic sedimentation basin design have been employed to enhance the performance of the sedimentation process. Laminar-flow devices Sedimentation can be accelerated by increasing particle size or decreasing the distance a particle must fall for removal. This can be achieved by coagulation and flocculation to increase particle size. A shorter fall distance can be achieved by provided parallel plates (commonly called plate settlers) or square-shaped tube (commonly called tube settlers) near the outlet of the basin. The plate spacing or tube sizes and feed rates are set to maintain laminar flow at all times. Alum coagulated sludge can remain deposited in the tubes at an angle as steep as 60o from the horizontal.

57

Fig. 5.11Tube settler systems (Montgomery).

Solids-contact devices They employ a sludge blanket at the bottom of the basin to promote flocculation and enmeshment of incoming solids. The same removal efficiency as in conventional sedimentation process can be achieved at higher loading rates. The flow, however, does not pass through the sludge blanket. An example of this kind of clarifiers is a centre feed circular tank where coagulation and flocculation occur in a central conically shaped compartment.

Fig. 5.12 A solids-contact clarifier.

For sludge blanket clarifiers (Fig. 5.13), there is a distinct layer or blanket of suspended solids which acts as a filter, trapping smaller particles which would otherwise following the up-flowing water out of the tank.

Fig. 5.13 A sludge blanket clarifier.

58

Dissolved air flotation systems Commonly used in waste treatment for sludge thickening, dissolved air flotation (DAF) has also been used in water treatment. It is a unit operation where solids are removed from the liquid by attaching to rising air bubbles. There are three steps in the flotation process: bubble generation, attachment of solids to the bubbles, and solids separation. The figure shows a DAF system where raw water is first coagulated and flocculated prior to entering the DAF basin. The water enters the basin near the bottom beneath a baffle to prevent short-circuiting. At the same entry point, a cloud of air bubbles called white water (typically 10 to 100 μm) is released and adhere to floc particles and causing them to float. The layer of solids formed on the water surface, known as float, is collected at the effluent end of the basin and is removed into a collection trough by a mechanical skimmer. Clarified water is removed through a perforated pipe system near the bottom of the basin.

Fig. 5.14 Schematic of a DAF system (Crittenden). Generally, DAF is most effective when it involves the removal of: (a) low-density particulate matter such as algae (b) water with dissolved organic matter e.g. natural colour (c) low-density flocs resulting from coagulation and flocculation of low- to moderately-turbidity waters (d) low-temperature waters as particulates are more difficult to settle

59

Proprietary sedimentation systems These systems are mostly modifications of the solids-contact clarifier. One is the Degremont pulsator clarifier (Fig. 5.15), which is used in Singapore. Water is fed through the bottom laterals. The feed rate is not constant. Instead, a specially designed vacuum chamber produces a pulsating flow.

Fig. 5.15 Pulsator clarifier.

Two-tray system

Fig. 5.16 Two-tray sedimentation basin.

60

This configuration offers a large sedimentation area in a relatively compact space by effectively stacking one basin on top of another. Flocculated water is fed through a perforated wall into a bottom chamber. The water flows horizontally to the end of the basin, then upwards into the upper chamber. The water then flows horizontally to the outlet zone in the end of the upper chamber. A chain-and flight sludge collection system sweeps the sludge along the floor into the lower chamber, where it settles and then raked to the sludge pit to be pumped out of the system. 5.8 Potential operational problems A number of factors not considered in the design of sedimentation basins can affect basin performance. These are temperature gradients, wind effects, inlet energy dissipation, outlet currents, and equipment movements. Temperature differentials: when warm influent water overflows into a sedimentation basin of cold water, it flows over it and reaches the outlet weirs in a much shorter time than the theoretical hydraulic retention time. This leads to the phenomenon of short-circuiting where bigger suspended particles can be carried away without settling due to the higher flow velocity. On the other hand, cold water entering a basin of warmer water tends to flow along the tank bottom as they are heavier (e.g. at 4oC ρw = 1000 kg/m3; at 20oC ρw = 998 kg/m3) and rise at the outlet end. .

Fig. 5.17 Short-circuiting of flow due to density current (Montgomery).

Turbidity effects: density currents may also be caused by changes in influent turbidity as a result of flash flood or strong winds on reservoir surfaces which increase the suspended solids (SS) concentration in water. Increase in turbidity will increase the density of influent causing it to dive to the bottom of the basin. (e.g. at 100 mg SS/L, density of mixture =1000.06 kg/m3; at 10000 mg SS/L, density of mixture = 1006 kg/m3) Wind effects: strong wind can cause water to overflow the outlet weir in addition to causing a surface current in the direction of the wind. There will be a return flow of water at the bottom of the tank in the opposite direction. Thus, wind actions can also cause short-circuiting of flow from inlet to outlet as well as scouring of settled particles at the bottom of the basin. Inlet energy dissipation: influent water normally enters the basin via a pipe at a sufficient velocity to keep the flocculated particles in suspension. At the inlet, the flow is distributed

61

across the basin section for the sedimentation process and the flow slows down significantly. This sudden reduction in energy is achieved by using baffles to dissipate the energy. If the baffles are not properly designed, density and eddy currents will be created and may cause short-circuiting of flow. Outlet currents: if the weir length is too short, outlet currents may form and sweep settleable particles into the effluent. V-notch weirs are normally used to allow better lateral distribution of outlet flow when basin level is imperfect. Equipment movement: the movement of equipment within the basin can affect its performance. Chain-and-flight scrapers, bridge-mounted scrappers, or hydraulic suction units are often used to remove settled sludge from the basin. Their movements, if excessive, will introduce currents which can stir up the settled particles and upset the sedimentation process. 5.9 Useful design parameters (i) Overflow rates (rectangular and circular clarifiers)

Type of material vo (m/hr) SOR = Q/A (m3/d/m2) Alum floc 0.6 – 0.9 14 – 22 Ferric floc 0.9 – 1.2 22 – 29 Lime precipitate 0.9 – 1.8 22 – 43 Grit 40 960 Raw sewage solids 1.0 24 Activated sludge 1.7 41

(ii) Detention time

2 to 6 hours - Permits further flocculation to occur - Controls horizontal velocity (usually less than < 30 m/h) - Determines tank depth - Does not control settling velocity

(iii) Water depth: 1.5 m to 4.5 m (iv) Length to width ratio: Rectangular tanks: approximately 4:1 to minimise short

circuiting (v) Sludge removal: Mechanical sludge removal equipment is preferred. (vi) Inlet design: Flow should be uniformly distributed. Multiple ports and/or baffles break

momentum of flow, but not to break up floc. (vii) Outlet design: submerged or free flowing weirs are often used. Flow rate over weir

should be less than 370 m3/d/m to prevent excessive velocities and particle carry-over.