Water and Wastewater Engineering

Подождите немного. Документ загружается.

8-10 WATER AND WASTEWATER ENGINEERING

potential of causing physical damage to the resin beads, the hydraulic requirements of the resin

rather than the kinetics for ion exchange govern the selection of the resin particle size.

Ion exchange resin beads are spherical. They are produced in particle diameters ranging from

0.04 to 1.0 mm. In the United States, the particle sizes are sold by standar

d sieve screen or “mesh”

sizes. A table of U.S. Standard Screen sizes and their equivalent diameters is given in Appendix B.

The common sieve size ranges used are 16 to 50 and 50 to 100. The smaller number is the largest

diameter sieve, and the larger number is the smallest diameter sieve. The manufactu

rer’s specifi-

cation is generally given the notation 16  50 or 50  100. Thus, for a 16  50 resin, all of the

resin beads will pass the number 16 sieve, and none will pass the number 50 sieve.

Other data provided by the manufacturer includes the effective size ( d

) and the uniformi ty

coefficient. The effective size is the mesh size that passes 10 percent of a sieved sample. The

uniformity coefficient is the ratio of the d

to the d

resin sizes. These data are provided to

facilitate hydraulic design.

Structural Stability and Service Life. As noted above, high pressure drops through the bed

have the potential to cause resin bead compression. This, in turn, can cause inadequate liquid

distribution and reduced flow. In addition the resin beads are al

so susceptible to swelling, shrink-

ing, and abrasion from excessive backwashing. These effects reduce the structural integrity of the

resin and shorten its operating life.

O xidation of the resin beads, especially strong acid sulfonated polystyrene-DVB resins, from

chlorination prior to ion exchange will significantly re

duce service life. If prechlorination is essential,

resins with high cross-linking are recommended (MWH, 2005).

E xcessive concentrations of iron and manganese, if oxidized, will form precipitates that will

foul the resin. GLUMRB (2003) specifies that iron, manganese, or a combination of the two

should not exceed 0.3 mg/L in the water applied to the re

sin. Organic compounds may foul the

resin by irreversibly binding to strong base anion exchange resins.

T urbidity should not exceed 5 NTU in water applied to cation exchange softeners (GLUMRB,

2003).

Some of these issues are remedied with the selection of an appropriate resin and proper

arrangement of the

sequence of pretreatment processes.

8-3 PROCESS OPERATION

To contact the water with the ion exchange resin, it is passed through a columnar pressure vessel

as shown in Figure 8-3 . The water is passed through the colu mn until the effluent no longer

meets the treatment objective. The column is then regenerated. The two common methods for

regeneration (cocurrent and countercurrent) are used to i

dentify the operating schemes.

Cocurrent Operation

In this scheme the regeneration step is conducted in the same flow direction as the treatment

flow. The direction of both flows is usually downward. For softening operations where some

leakage of hardness in the effluent can be tolerated, this operational scheme is frequently chosen.

It is usually the lowest cos

t design and the simplest to operate. The domestic water softener is a

familiar example of this type of design (Brown, 1998).

ION EXCHANGE 8-11

The following steps are used in the ion exchange cycle :

• Service. The raw water is passed downward through the column until the hard-

ness exiting the column exceeds the design limits. The column is taken

out of service and another column is brought on line.

• Backwash. A flow of water is introduced thro

ugh the underdrain. It flows up through

the bed sufficient to expand the bed by 50 percent. The purpose is to

relieve hydraulic compaction (Gottlieb, 2005), and to move the finer

resin material and fragments to the top of the column and remove any

suspended solids that have accumulated d

uring the service cycle.

• Regeneration. The regenerating chemical, for example, sodium chloride, flows downward

through the bed at a slow rate to allow the reactions to proceed toward com-

plete regeneration.

• Slow rinse. Rinse water is passed through the column at the same flow rate as

the regenerating flow rate to pu sh the regenerating chemical through

the bed.

Upper manifold

Nozzles

Resin

Graded quartz

Lower manifold

Strainer nozzles

Backwash

outlet

Sight glass

Backwash controller

Meter

Water inlet

Water outlet

Regenerant

FIGURE 8-3

T ypical ion exchange resin column.

( Source: U.S. EPA, 1981.)

8-12 WATER AND WASTEWATER ENGINEERING

• Fast rinse. This is a final rinse step. The fast rinse flows at the same flow rate as the

service flow rate to remove any remaining regenerating solution.

• Return to service. The column is put back in use.

Countercurrent Operation

In this mode of operation the regenerant is passed though the resin in the opposite d irection to

that of the water being treated. Generally, the mode of operation is raw water flowing downward

and regenerant flow upward. In most cases, countercurrent operation will result in lower leak-

age and higher chemical efficiency than cocurrent operation. However, countercurrent opera-

tion is a more expens ive de

sign and is more com plicated to operate. Countercurrent operation is

used where (1) high purity water is required, (2) chem ical consumption must be minimized, or

(3) waste volu me must be minimized .

Bypass

A s noted previously, there will be some leakage of hardness through the column because the

passage of the saturation wave through the column is spread out, as shown in Figure 8-2 , and

because the high concentration of regenerant being released from the upper levels of the column

will “regenerate” lower portions of the column where polyvalent ions were not co

mpletely removed

in the regeneration cycle. The amount of leakage is usually less then 5 mg/L as CaCO

(Clifford,

1999). Thus, the treated water is softened far more than is necessary for normal consumer use.

Thus, passing the entire flow to satisfy demand through the column results in a larger column than

is necessary as well as consuming larger amounts of regeneration chemic

als. In addition, very soft

water is often corrosive.

To improve the stability of the water and make it less corrosive while reducing costs, a por-

tion of the flow is bypassed around the column and blended with the treated water to achieve the

design hardness. The bypass flow is calculated by solving the mass balanc

e for hardness at the

point where blending takes place. The mass balance of hardness is

QC QC Q C

treated treated bypass bypass blended bl



eended

(8-17)

and the flow balance is

QQQ

treated bypass blended



(8-18)

where Q

treated

 flow rate of raw water entering column for treatment, m

/ d

bypass

 flow rate of water that is not treated, m

/ d

blended

 total design flow rate

treated

 concentration of hardness in the treated water, mg/L as CaCO

bypass

 hardness of the raw water, mg/L as CaCO

blended

 design final hardness, mg/L as CaCO

The bypass flow rate is determined by simultaneous solution of these two equations.

ION EXCHANGE 8-13

Multiple Columns

With the exception of home water softeners that may be shut down for a short period at night or that

may be replaced by a service technician, the exhausted column that is taken out of service must be

replaced by bringing another column on line. Although there are many alternate arrangements, three

schemes are more common than others

. They are (1) a standby column, (2) columns in series (known

in the trade as the “merry-go-round” system), and (3) columns in parallel (the “carousel” system).

In the standby system there is a minimum of two columns. One is in service while the other

is being regenerated and pla

ced in standby. The operating time of each column must be long

enough to allow for regeneration of the out-of-service column. This system does not provide any

redundancy if only two columns are provided. A three column arrangement provides one extra

column in the rotation and allows for backup during

maintenance.

A s show in Figure 8-4 , the first column in the merry-go-round system serves as a roughing

column and a second column serves as a polishing step.

In the carousel system, three columns are run in parallel while one is out of service. The three

column

s are in various stages of exhaustion: u p to and including breakthrough, less than break-

through, and substantially less than breakthrough. The water from the three columns is blended

to achieve a consistent product water. This system is more likely to be used to meet an MCL

requirement for a toxic con

stituent than for softening.

8-4 ION EXCHANGE PRACTICE

T ypical design criteria for cation and anion exchange systems are summarized in Table 8-2 . The

following paragraphs elaborate on the design parameters.

Resin Selection

There are several hundred different res ins available from United States and European manu-

facturers. Of these, the resins based on the polystyrene-DVB matrix are most widely used. The

operating capacity ( meq/mL as CaCO

) serves as the primary selection criterion. This differs

Column 1 Column 2 Column 3

FIGURE 8-4

Two columns in series with one column as standby. After exhaustion of column 1, it will be taken out of service and regener-

ated. Column 2 will become “lead” and column 3 will follow in series. When column 2 is exhausted, it is taken out of service

and column 3 bec

omes “lead.” Its effluent is passed through the regenerated column 1. This system has been called a “merry-

go-round” system.

8-14 WATER AND WASTEWATER ENGINEERING

from the exchange capacity in that it is a measure of the actual performance of the resin under

a defined set of conditions such as the raw water composition, empty-bed contact time (EBCT)

or service flow rate (SFR), and degree of regeneration. The operating capacity is always less

than the advertised exchange capacity be

cause of incomplete regeneration, early leakage (break-

through) that causes termination of the operational cycle to meet design limits, and efficiency of

regeneration (measured as eq NaCl/eq CaCO

S m all laboratory columns (1.0 to 5.0 cm inside diameter) have been effective in analyzing

alternative resins. Thes e columns can be scaled directly to full scale design if the loading rate

and EBCT are the same. Because the resin beads are small compared to the column diameter,

the error du

e to channeling of the water down the walls is small. The hydraulics of full scale

operation cannot be modeled by these small scale columns (MWH, 2005).

Strong acid Strong base

Parameter cation resin anion resin Unit

Exchange capacity 3.6–5.5 1.8–2.0 meq/g as CaCO

1.6–2.2 0.8–1.4 meq/mL as CaCO

Operating capacity

50–70 40–60 % of exchange capacity

Moisture content 40–80 35–80 %

Shipping weight (moist) 640–930 670–720 kg/m

Screen size 16  50 16  50

Service flow rate 200–2000 200–2000 m

/d · m

of resin

8–40 8–40 BV/h

Surface loading rate 400–800 400–800 m

/d · m

or m/d

Backwash rate 12–20 5–7.5 m

/h · m

or m/h

Backwash duration 5–155–20 min

Bed expansion 50 50–75 %

Regenerant NaCl NaCl

Regenerant concentration 5–10 2–15 %

Regenerant dose 80–320 80–320 kg NaCl/m

of resin

Regeneration flow rate 60–120 60–120 m

/d · m

or m/d

2–5 2–5 BV/h

Rinse volume2–5 2–10 BV

pH range 0–14 0–14 units

Max. operating temp. 140 OH



form  60; C



form  100

Turbidity limit 55 NTU

Iron limit 5 0.1 mg/L as Fe

Total Fe  Mn 0.3 0.3 mg/L

Chlorine limit 1.0 0.1 mg/L of Cl

Operating capacity depends on the method of regeneration and amount of regenerant applied.

Sources: Clifford, 1999; GLUMRB, 2003; MWH, 2005.

TABLE 8-2

Typical ranges for design data and criteria for strong acid cation and strong base anion resins

ION EXCHANGE 8-15

The larger the laboratory or pilot scale column, the better will be the results from scale-up.

Although 60 cm long, 1 to 5 cm diameter columns are adequate for laboratory studies, larger

diameter columns (for example, 10 cm) that have res in bed depths greater than 1 m are recom-

mended (Reynolds and Richards, 1996; MWH, 200

5).

Breakthrough curves are obtained from laboratory scale or pilot scale data such as that shown

in Figure 8-5 . The design breakthrough concentration, shown as

in Figure 8-5 , may be used to

estimate the capacity of the resin by calculating the area between the influent concentration ( C

)

and the effluent concentration and dividing by the mass of resin in the column.

Flow Rates

The flow rate through the column affects the kinetics of the absorption bed. The longer the water

is in contact with the resin, the greater is the opportunity for the mechanisms of the exchange

process to come into play. Thus, the longer the contact time, the longer the time to reach break-

through. There are two parameters used to control the contact time: (1) empty-bed contact time

(EBCT) and (2) service flow rate (SFR) or exhausti

on rate. The EBCT is calculated as the vol-

ume occupied by the resin (

) divided by the flow rate:

EBCT

(8-19)

The service flow rate is

SFR 

(8-20)

The EBCT and SFR are used for ease of calculation. An ac tual detention time in the bed

would have to account for the porosity. Typic al EBCTs range from 1.5 to 7.5 min and SFRs

range from 200 to 1,000 m

of water per day for each cubic meter of resin (m

/ d · m

). The SFR

012

3 4 5 6

Volume, m

C, meq/L as CaCO

Area  6.572 meq/L  2,350 L

 15,444 meq

FIGURE 8-5

Ion exchange softening breakthrough curve.

8-16 WATER AND WASTEWATER ENGINEERING

may also be expressed as bed volumes of water per hour (BV/h). The usual range of values is 8 to

40 BV/h. EBCTs greater than 7.5 min and SFRs less than 200 m

/ d · m

are acceptable because

they provide greater time for the reactions. Shorter EBCTs and higher SFRs will result in earlier

breakthroughs.

The surface loading rate (SLR) is limited to control the pressure drop across the bed and

thereby control breakage of the resin beads. It is expressed as

SLR 

(8-21)

where A

 the cross-sectional area of the resin bed, m

In general, the maximum allowable pressure drop across the bed is about 140 kPa. Because

the pressure drop increases over time as the bed is operated, the design value for pressure drop

is usually about 35 to 70 kPa less than this . This results in a maximum SLR of about 880 cubic

meters per day per square meter of cross-sectional area (m

/ d · m

or m/d). SLRs range from 175

to 880 m/d (Gottlieb, 2005). GLUMRB (2003) specifies that the rate should not exceed 400 m/d.

Typical manufacturers’ design curves for pressure drop are shown in Figure 8-6 .

Backwashing

A s noted previously, cocurrent beds are backwashed to relieve com pression and remove parti-

culate matter (often called “fines”). The backwash rate for strong acid cation resins is in the range

12–20 m

/h · m

of bed surface area. The backwashing period is on the order of 5 to 15 minutes.

Bed expansion during backwash is typically assumed to be 50 percent, but some authors report ex-

pansions up to 100 percent of the operating depth (Reynolds and Richards, 1996; MWH, 2005).

Estimation of Resin Volume

There are several methods for estimating the required resin volume. The one that will be described

here depends on the results of column studies. Column studies on the raw water provide a better

estimate of the kinetic behavior of the resin to the actual constituents in the water than either

synthetic water data or equilibrium data provided by m

anufacturers.

0 200 400 600 800 1000

Surface loading rate, m

/d-m

Pressure drop, kPa/m of resin

5C

10C

20C

25C

FIGURE 8-6

T ypical ion exchange resin pressure drop curves at various water temperatures. Actual

manufacturer’s data should be used for design.

ION EXCHANGE 8-17

The column should be operated long enough to achieve complete saturation of the bed

through several cycles of service and regeneration. To determine the optimum SFR, the flow rate

must be varied during the saturation loading tests. The main goal in determining the optimum

SFR is to reduce the capital cost of the colum

In the simplest expression, the resin volume required to treat a given flow rate of water is



SFR

(8-22)

One method for estimating the resin mass is based on the principle of mass balance. It is

illustrated in the following example.

Example 8-2. As part of the preliminary design for a softening plant, a sodium-based ion ex-

change column is to be evaluated. For the evaluation of alternatives, estimate the mass of moist

resin required to soften the Hard Times water (Example 7-6) to a hardness of 80 mg/L as CaCO

The design flow rate is 275 m

/ d. Assume that there is no leakage from the column, that is,

treated

 0.0 mg/L as CaCO

, that the moisture content of the resin is 44%, and the operat-

ing temperature is 10  C. Also assume that iron and turbidity c oncentrations are negligible. The

manufacturer’s resin operating capacity to breakthrough is 67% of the exchange capacity.

The laboratory scale column was 7.5 cm in diameter and the height of the resin in the col

umn

was 150.0 cm. The resin density on a moist basis is 0.85 g/cm

. The moisture content is the same

as the full scale column. The flow rate through the column was 0.18 m

/h.

Solution:

a . Begin by computing the meq of hardness removed per g of resin on a dry weight basis.

The mass of dry resin in the column is computed from the column dimensions, the unit

weight of the resin, and the moisture content of the resin.



()

()( )(

7 5

150085 1044

cm g/cm

⎡

⎣

⎢

⎤

⎦

⎥

 )) 3 154 35,.g

b. From the breakthrough curve of the laboratory column ( Figure 8-5 ) the meq of hardness

removed at breakthrough (

) was 15,444 meq.

c. The meq/g of dry resin is

15 444

3 154 35

489

meq

meq/g

d. The total hardness of the Hard Times water is equal to the sum of the Ca

2 

and Mg

2 

238 mg/L as CaCO

 90.6 mg/L as CaCO

 328.6 mg/L as CaCO

e. A material balance on the flow downstream of the ion exchange column is used to deter-

mine the bypass flowrate.

8-18 WATER AND WASTEWATER ENGINEERING

QC QC Q C

treated treated bypass bypass blended bl



eended

With no leakage C

treated

 0.0 mg/L and material balance equation is

treated bypass

mg/L as Ca() ( ) (0 328 6 2

.CO775 80

275

m /dmg/L as CaCO

m /

bypass

)( )

(

Q 

ddmg/L as CaCO

mg/L as CaCO

)( )80

328 6

 .. 9 5

m /d

f. The flow rate passing through the ion exchange column is

275 66 95 208 05

33 3

m /dm/dm/d..

g. The meq/L of hardness to be removed is

()328 6

6 572

..mg/L as CaCO

meq/mg

⎛

⎝

⎜

⎞

⎠

⎟

 meq/L

where 50 meq/mg is the equivalent weight of CaCO

h. The meq of hardness to be removed in one day is

()()()()6 572 208 05 1 000 1

..,meq/L m /d L/md113710

.  meq

i. The mass of dry resin required is

()()1 3710

489

6 3





meq

kg g

⎛

⎝

⎜

⎞

⎠

⎟

/ 2279 61 280.or kg

j. The mass of resin on a moist basis is

()280

1044

500kg kg





⎛

⎝

⎞

⎠

k. To account for the manufacturer’s operating capacity to breakthrough, increase the

mount of resin to

067

500 746 27 750

.()kg or kg

Comments:

1 . The estimated resin was given in meq hardness/g of dry resin. This estimate could also be

made as meq of hardness/mL of moist resin. The resin is shipped and installed moist. The

ION EXCHANGE 8-19

volume should be estimated in the test column after the test column has been backwashed

and settled over several cycles.

2. The assumption of zero leakage is not realistic. As noted above, it will usually be some

concentration less than 5 mg/L as CaCO

3. The estimate of the hardness removed by the resin is determined by computing the area

under the breakthrough curve at the “design breakthrough.” F or a two column system

the design breakthrough is some hardness concentration above the leakage. For three

colum ns in series, the design breakthrough

may be as high as complete bed exhaus-

tion. Likewise, in the parallel system with four columns, exhaustion of the bed may be

selected as the design breakthrough.

Regeneration

Resins operated on the sodium cycle are usually regenerated with a 5 to 10% brine solution. The

mass loading ranges from 80 to 320 kg NaCl/m

of resin with 80 to 160 kg NaCl/m

of resin

being typical. The liquid flow rate is 60 to 120 m

/ d · m

of surface area or in terms of bed vol-

umes, about 2–5 BV/h (Reynolds and Richards, 1996; MWH, 2005).

Slow Rinse

The water rinse to push the regenerate through the bed is at the same flow rate as the regeneration.

Cycle Time

A minimum of two columns is recommended for redundancy: one in service and one in regenera-

tion or standby. One column in service with storage is an alternative, but it provides no redun-

dancy for mechanical or resin rehabilitation. Even with two columns, the out-of-service time must

be less than the operating tim

e for the in-service column to reach breakthrough. The following

may be used to estimate the out-of-service time (Clifford, 1999):

ttttt

os bw r srfr



(8-23)

where t

 out-of-service time

 time for backwashing, 5 to 15 min

 time for regeneration, 30 to 60 min

 time for slow rinse, 10 to 30 min

 time for fast rinse, 5 to 15 min

Using the maximum estimate for each of these steps, the total out-of-service time is about

two hours.

Example 8-3. An alternative three column design for Hard Times ( Example 8-2 ) is to be evalu-

ated. In this alternative, the columns will be in series and exha

ustion of the resin is the “design

breakthrough.” The total out-of-service time is estimated to be two hours.