Best Practices Desaereator[1]

Section 2.7

DEAERATORS AND FEEDWATER TANKS

CONFIDENTIAL, For Nalco Employee Use Only ©2007 Nalco Company

Boiler Water Applications and Best Practices Guide – Book 314 (10-07)221

Deaerators and feedwater tanks provide the most common means of deoxygenation in theplants/systems we service. Mechanical removal of oxygen should be the primary means ofdeoxygenation, with removal by oxygen scavenging secondary. It is, therefore, necessary tobe able to survey and troubleshoot deaerators and feedwater tanks, determine whether perfor-mance is appropriate, and make recommendations if improvements are needed.

DEAERATORS

THE SURVEY

The first step in servicing or troubleshooting adeaerator is to complete a Mechanical-Opera-tional-Chemical (MOC) survey. The survey formshould include questions in each of the MOCareas. Answers to these questions will helpestablish expected versus existing performanceof the deaerator, as well as identify potentialproblem areas.

Mechanical questions:

• For what boiler(s) or system does this specificdeaerator provide feedwater?

• How many deaerators are present in thissystem?

• What is the type of deaerator (spray, tray,atomizing, co-current, counter-current, etc.)?

• What is the metallurgy of the main deaeratorcomponents (shell, sprays, and trays)?

• Is there an external vent condenser?

• What is the design flow rate?

• What are the design temperatures and flowsfor all makeup and condensate streams?

• Do condensate and makeup mix before thedeaerator or enter separately?

• What is the return (entry) point for high-temperature condensate streams?

• What is the design pressure?

• Where is the dome pressure sensor located?

• Where are the sample points located, and whatare their estimated lag times based on linelength and diameter?

• What is the sample line metallurgy?

• What dissolved oxygen concentration is thedeaerator designed to supply?

Operational questions:

• What is the actual water flow through thedeaerator?

• What is the condition of the vent (steady,sputtering, an invisible space between the topof the vent and the plume, etc.)?

• What are the source, pressure, and temperatureof the steam to the deaerator?

• What is the actual operating pressure?

• Does the temperature fluctuate, if so by howmuch?

• Does the pressure fluctuate, if so by howmuch?

• What are the dome temperature and pressure?

• What is the storage tank temperature?

• What is the temperature and flow rate of allhigh-temperature condensate streams?

• What is the temperature and flow rate of allmakeup and low-temperature condensatestreams?

• Do sample lines flow continuously or only asneeded? If only as needed, how much flushtime is provided before sampling?

• Is there an oxygen meter present, and howoften is it calibrated?

Section 2.7




• What are the date and findings of the mostrecent oxygen testing done without scavenger?

• What are the date and findings of the lastinternal inspection, e.g., condition of the spraynozzles, trays, storage tank, welds, etc.?

• What are the date and findings of the last weldintegrity testing?

Chemical questions:

• What products/chemicals are fed to thedeaerator or in the deaerator recirculation line?

• What is the dosage of these products?

• Are the products fed neat or diluted?

• Are multiple products mixed together beforeor at the injection point?

• What are the point(s) of injection for allproducts fed into the deaerator, e.g., todeaerator neck, storage, dropleg (afterdeaerator storage), recirculation line, etc.?

• Is a quill or extended feed header used, andhow far does it extend into the feed point?

• What is the metallurgy of the feed system?

• What is the measured dissolved oxygenconcentration in the deaerator effluent withoxygen scavenger?

• What is the feedwater pH out of the storagetank after scavenger injection?

TROUBLESHOOTING POOR

PERFORMANCE

In troubleshooting a deaerator, an open mind anda systematic plan of attack are necessary.Working your way through the following actionsshould reveal whether or not the deaerator youare working with is operating properly. If it isnot, these procedures should pinpoint theproblems. When you conclude your survey,make sure your observations and recommenda-tions fit the facts that you observed.

Evaluate Original Design and Installation

Obtain original design parameters anddeaerator specifications – A deaerator is arelatively simple mechanical device, which isdesigned to remove dissolved oxygen (DO) fromboiler feedwater down to maximum concentra-tions of 7-40 ppb. The level of removal dependson deaerator design. When a deaerator does notfunction properly, it is due to one of two causes:

1. Deviations from original designspecifications

2. Mechanical failure

Deviations from specifications can be caused bychanged plant operating conditions, designretrofits, or engineering changes that haveoccurred since design. Start your survey at theoriginal deaerator operating manual. Record allmanufacturer’s design criteria and specifica-tions. (Refer to mechanical questions previouslylisted.) Obtain an original drawing of the vessel.Carefully note the positions of all inlets, outlets,baffles, gauges, and sensors.

Obtain current system diagram and determinecurrent operating parameters – Go to thedeaerator, and prepare a drawing of the unit, as itcurrently exists. Carefully note the position ofall inlets, outlets, gauges, thermocouples, andother sensors. Obtain accurate flow and tempera-ture data for all influent streams and for thesteam supply. Note flow swings. Do not averageflow information, but record the flow ranges.Diagnostic TRASAR® can be used for calibra-tion of flow meters and verification of estimatedflow rates.

Evaluate Current On-Line Performance

Check pressure of deaerator dome – Theoperating pressure should be well regulatedand not vary more than 0.5 psig (3.4 kPag).Pressure fluctuations can be caused by any ofthe following:

• Rapid flow changes in inlet water

Section 2.7




• Rapid changes in condensate return flow ortemperature, especially of high-temperaturereturn streams

• Insufficient or fluctuating steam supply

• Improper instrument tuning of the pressurereducing valve (PRV)

• Improper size or design or excess pressuredrop across the PRV

• Improper location of pressure sensing devicesfor PRV

In general, the pressure sensor for a PRV shouldbe located on the deaerator proper, not on thesteam piping coming to the deaerator. Locationsother than the deaerator result in a lag in re-sponse time to pressure changes in the dome,resulting from changes in load or inlet tempera-tures. This lag allows continuous rapid pressureand temperature changes in the unit and con-tinuous changes in the stress level appliedto the tank.

Check temperature of deaerator dome – Thetemperature of the deaerator dome sectionshould be within 1-2°F (0.6-1.1°C) of thetheoretical saturated steam temperature at theoperating pressure. Atmospheric pressure(14.7 psia; 101.3 kPaa in most locations) mustbe added to the gauge pressure before referringto most steam tables for the temperature value.

EXAMPLE 1

Gauge pressure on deaerator dome = 10.0 psig (68.9 kPag)Atmospheric pressure = 14.7 psia (101.3 kPag)Absolute pressure for steam table = 24.7 psia (170.2 kPaa)

Therefore, the dome temperature should be238.5°F ± 2°F [114.7 ± 1.1°C] based on steamtable in Appendix A. Too low a temperaturewill indicate an internal malfunction, insufficientventing, too little steam flow, too muchfeedwater flow, or too low an inlet watertemperature.

It is important to note that a fluctuating domepressure will result in fluctuating temperatures.This will increase the probability that thetemperature is below saturation for a portion ofthe cycle. Oxygen solubility increases signifi-cantly, whenever the temperature drops belowsaturation. The result is typically high oxygencontent in the effluent (feedwater). The effect offluctuating pressure and temperature is morepronounced in a spray type deaerator becausethere is less water volume within the dome.

Location of the deaerator thermometer (ifpresent) is just as important as location of thepressure sensor. It is not uncommon for thethermometer to be located very close to thesteam inlet. When this is the case, the thermom-eter is actually measuring the temperature of theinlet steam and not the temperature of the dome.This is of particular importance when the steamused is superheated. The difference betweentheoretical (from steam tables) versus measuredwill then be very large, with measured being thelarger number.

Check temperature of deaerator storagesection – The temperature of the deaeratorstorage section should be within 1-4°F(0.6-2.2°C) of the deaerator dome temperature.A temperature difference greater than thisindicates that there was insufficient contactbetween the water and steam as the watermade its way through the dome. The storagesection temperature will be the lower of thetwo temperatures.

The temperature difference between dome andstorage may be very large if the dome tempera-ture is actually the temperature of superheatedsteam because of an improperly positionedprobe. Under these circumstances, you will needto estimate the dome temperature from the steamtables using the dome pressure and compare thattemperature to the storage temperature.

Section 2.7




Typically, anything that results in unevendistribution of water through the sprays or acrossthe trays will result in insufficient contact. Thiscan include broken spray nozzles and cracked,plugged, broken, or shifted trays.

Separate introduction of return condensate andcold makeup into the deaerator water box orpreheater sprays can result in a slightly differentdistribution problem. The amount of mixingthat actually occurs in the water box is minimalso that the water spraying out of the water boxtends to divide the preheater and tray sectionsinto hot and cold zones. This creates an im-balance of steam flow through the tray stack,with the steam gravitating to the cold side.The performance impact of this conditionwill depend on the temperature differencebetween zones.

Check deaerator vent and vent line – Insuffi-cient venting is one of the most common causesof poor deaerator performance. Noncondensablegases not vented from the deaerator will con-centrate in the vapor phase, increasing theirpartial pressure. This in turn, increases theirconcentration in the liquid phase or deaeratoreffluent (feedwater).

Proper venting of a deaerator will result in aninvisible or clear area between the vent pipedischarge to the atmosphere and the plume.There are numerous opinions on how long thisclear space should be. There are also numerousopinions on plume height, ranging from6-36 inches (152-914 mm). Acceptable plumeheight is actually deaerator specific and isaffected by existing weather, such as tempera-ture and wind speed. It is better to adjust thevent rate to achieve an acceptable oxygenconcentration (verified by measurement) andthen note general plume appearance for future

reference. The plume should be visible (evenwith a vent condenser), strong, steady (notpuffing), and without entrained water.

Common causes for insufficient venting arechoked flow, plugged orifices, intentionalreduction for conservation of energy, convolutedor excessively long vent piping, and rapidvariations of feedwater flow or temperature,which results in a temporary upset condition.

Energy conservation measures may also resultin the installation of heat exchangers, coolingjackets, or other devices to reclaim the heatcontent of the vent steam containing thenoncondensables. Many of these devices cause abackpressure on the vent line and thus restrictflow. The result is higher noncondensable gasesin the deaerator effluent (feedwater).

Irregular spitting of water with the vent dis-charge can indicate condensation, internal leaks,water entrainment from broken or damagedspray nozzles, erratic spray valve action, insuffi-cient vent opening, or incorrect vent piping. Thiscondition, whatever the cause, can choke thevent gas flow. Cooling water leaks in an externalvent condenser can cause excess water accumu-lation in the shell. If the drainpipe cannotdischarge the excess liquid or is plugged, gasflow through the vent will be restricted.

Typically, the maximum quantity of steamrequired for venting a properly operatingdeaerator is less than 0.2% of the feedwaterflow. (See Figures 2.7.1 and 2.7.2.) This valuevaries relative to the percentage of fresh makeupused, but is less than 0.2% for tray or spray typedeaerators operating with 100% makeup. Somedeaerator manufacturers might conservativelyrecommend 0.5% to assure sufficient venting.

Section 2.7




Figure 2.7.1 – Vent rate for a tray type

deaerator (U.S.)

Figure 2.7.1 – Vent rate for a tray type

deaerator (Metric)

Figure 2.7.2 – Vent rate for a spray type

deaerator (Metric)Figure 2.7.2 – Vent rate for a spray type

deaerator (U.S.)

Section 2.7




The vent valve may have a 1/8 inch (3.2 mm)hole drilled in the seat to prevent the valve frombeing tightly closed during operation. For verylight loads or small noncondensable loading,this 1/8 inch (3.2 mm) hole is sometimes largeenough to act as a throttling orifice, but this isan unusual situation. The vent valve should notbe operated in this closed position (with only the1/8 inch [3.2 mm] hole as vent), unless adequateventing is indicated by temperature and oxygenchecks.

To determine the correct amount of openingrequired, the vent valve should be openedapproximately one or two turns and the effecton the operating temperature noted. If noappreciable effect on the temperature is notedafter one hour, oxygen tests should be done todetermine the effectiveness of venting, i.e., isthe appropriate oxygen removal achieved.

The vent rate can be reduced by tightening(closing) the vent valve. If after reducing theopening, the deaerator operating temperaturedrops or the difference between dome andstorage temperatures increases, venting is notadequate, and the vent valve must be openedfurther.

When loads are small or where uniform opera-tion (flow and pressure) is expected, a fixedorifice can be used. This usually consists of adrilled pipe cap or orifice plate mounted above avent valve. The vent valve should be fully openand not drilled with the 1/8 inch (3.2 mm) hole.The optimum size of the orifice can be found bychecking dissolved oxygen in the effluent andverifying that the deaerator temperature matchesthe theoretical saturation temperature for theoperating pressure. If the oxygen is high or thetemperature is low, the hole size in the orificeneeds to be increased. Orifice hole size can alsobe estimated based on feedwater flow andpressure. (See Figure 2.7.3 and Table 2.7.1.)

0

200

400

600

800

1000

1200

1400

1600

1800

0.0 10.0 20.0 30.0 40.0 50.0 60.0Pressure (psig)

Flo

w (

lb/h

r)1/8 in

5/8 in

3/4 in

7/16 in

3/8 in

5/16 in

1/4 in

3/16 in

9/16 in

1/2 in

13/16

11/16 in

Figure 2.7.3 – Estimated steam flow vs. orifice

size (U.S.)

0

20

40

60

80

100

120

0 100 200 300 400Pressure (kPag)

Flo

w (

kg

/hr)

8 mm

4 mm

3 mm

2 mm

5.5

6 mm

6.5 mm

7 mm

7.5 mm

4.5

5 mm

2.5 mm

3.5

Figure 2.7.3 – Estimated steam flow vs. orifice

size (Metric)

Section 2.7




1/8

3/1

61/4

5/1

63

/87/1

61/2

9/1

65/8

11

/16

3/4

13

/16

7/8

15/1

61

.0

1.5

10.2

22.9

40

.863.7

91.7

12

4.8

163

.020

6.3

254.7

308.2

366.8

430

.4499

.257

3.1

65

2.0

2.2

10.6

24.0

42

.666.5

95.8

13

0.4

170

.321

5.6

266.1

322.0

383.2

449

.8521

.659

8.8

68

1.3

2.9

11.1

25.0

44

.469.4

99.9

13

6.0

177

.622

4.8

277.6

335.9

399.7

469

.1544

.062

4.5

71

0.6

3.6

11.6

26.0

46

.272.3

104.0

14

1.6

185

.023

4.1

289.0

349.7

416.2

488

.4566

.465

0.3

73

9.8

4.4

12.0

27.0

48

.175.1

108.2

14

7.2

192

.324

3.4

300.4

363.5

432.6

507

.7588

.967

6.0

76

9.1

5.1

12.5

28.1

49

.978.0

112.3

15

2.8

199

.625

2.6

311.9

377.4

449.1

527

.1611

.370

1.7

79

8.4

5.8

12.9

29.1

51

.780.8

116.4

15

8.4

206

.926

1.9

323.3

391.2

465.6

546

.4633

.772

7.4

82

7.7

6.5

13.4

30.1

53

.683.7

120.5

16

4.0

214

.227

1.1

334.7

405.0

482.0

565

.7656

.175

3.2

85

7.0

7.3

13.8

31.2

55

.486.5

124.6

16

9.6

221

.628

0.4

346.2

418.9

498.5

585

.0678

.577

8.9

88

6.2

8.7

14.8

33.2

59

.092.3

132.9

18

0.8

236

.229

8.9

369.1

446.6

531.4

623

.7723

.383

0.4

94

4.8

10

.21

5.7

35.3

62

.798.0

141.1

19

2.0

250

.831

7.5

391.9

474.2

564.4

662

.4768

.288

1.8

10

03.3

11

.61

6.6

37.3

66

.41

03.7

149.3

20

3.3

265

.533

6.0

414.8

501.9

597.3

701

.0813

.093

3.3

10

61.9

13

.11

7.5

39.4

70

.01

09.4

157.6

21

4.5

280

.135

4.5

437.7

529.6

630.2

739

.7857

.898

4.8

11

20.4

14

.51

8.4

41.4

73

.71

15.1

165.8

22

5.7

294

.737

3.0

460.5

557.3

663.2

778

.3902

.710

36.2

11

79.0

17

.42

0.3

45.6

81

.01

26.6

182.3

24

8.1

324

.041

0.1

506.3

612.6

729.1

855

.6992

.311

39.2

12

96.1

20

.32

2.1

49.7

88

.31

38.0

198.7

27

0.5

353

.344

7.1

552.0

668.0

794.9

932

.9108

2.0

12

42.1

14

13.2

23

.22

3.9

53.8

95

.61

49.4

215.2

29

2.9

382

.648

4.2

597.8

723.3

860.8

101

0.3

117

1.7

13

45.0

15

30.3

26

.12

5.7

57.9

103

.01

60.9

231.7

31

5.3

411

.952

1.3

643.5

778.7

926.7

108

7.6

126

1.3

14

47.9

16

47.4

29

.02

7.6

62.0

110

.31

72.3

248.1

33

7.7

441

.155

8.3

689.3

834.0

992.6

116

4.9

135

1.0

15

50.9

17

64.5

36

.33

2.1

72.3

128

.62

00.9

289.3

39

3.8

514

.365

0.9

803.6

972.4

1157

.2135

8.1

157

5.1

18

08.2

20

57.3

43

.53

6.7

82.6

146

.92

29.5

330.5

44

9.8

587

.574

3.6

918.0

1110

.81

321

.9155

1.4

179

9.3

20

65.5

23

50.1

50

.84

1.3

92.9

165

.22

58.1

371.6

50

5.9

660

.783

6.2

1032

.41

249

.21

486

.6174

4.7

202

3.4

23

22.8

26

42.8

1. B

ased

on N

apie

r's f

orm

ula

Gau

ge

Pre

ssu

re

ps

ig

NO

ZZ

LE

DIA

ME

TE

R (

in)

FL

OW

RA

TE

(lb

/h)1

Ta

ble

2.7

.1 –

Ste

am

flo

w t

hro

ug

h n

ozzle

s (

U.S

.)1

Section 2.7




2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

9.0

10

.0

10

1.8

22

.84

4.0

95.5

77

.28

9.2

11

1.3

713.7

616.3

71

9.2

222.2

925.5

82

9.1

136

.84

45.4

8

15

1.9

02

.97

4.2

85.8

27

.61

9.6

31

1.8

914.3

817.1

22

0.0

923.3

026.7

43

0.4

338

.51

47.5

5

20

1.9

83

.10

4.4

66.0

87

.94

10.0

51

2.4

015.0

117.8

62

0.9

624.3

127.9

03

1.7

540

.18

49.6

1

25

2.0

73

.23

4.6

56.3

38

.27

10.4

61

2.9

215.6

318.6

02

1.8

325.3

229.0

63

3.0

741

.85

51.6

7

30

2.1

53

.36

4.8

46.5

88

.60

10.8

81

3.4

316.2

519.3

42

2.7

026.3

330.2

23

4.3

943

.52

53.7

3

35

2.2

33

.49

5.0

26.8

38

.93

11.3

01

3.9

516.8

820.0

82

3.5

727.3

431.3

83

5.7

145

.19

55.7

9

40

2.3

13

.62

5.2

17.0

99

.26

11.7

21

4.4

617.5

020.8

32

4.4

428.3

532.5

43

7.0

346

.86

57.8

5

45

2.4

03

.74

5.3

97.3

49

.59

12.1

31

4.9

818.1

221.5

72

5.3

129.3

633.7

03

8.3

448

.53

59.9

1

50

2.4

83

.87

5.5

87.5

99

.92

12.5

51

5.4

918.7

522.3

12

6.1

830.3

734.8

63

9.6

650

.20

61.9

8

60

2.6

44

.13

5.9

58.1

010.5

813.3

81

6.5

219.9

923.8

02

7.9

332.3

937.1

84

2.3

053

.54

66.1

0

70

2.8

14

.39

6.3

28.6

011.2

414.2

21

7.5

621.2

425.2

82

9.6

734.4

139.5

04

4.9

456

.88

70.2

2

80

2.9

74

.65

6.6

99.1

111.9

015.0

51

8.5

922.4

926.7

63

1.4

136.4

341.8

24

7.5

860

.22

74.3

4

90

3.1

44

.90

7.0

69.6

112.5

515.8

91

9.6

223.7

428.2

53

3.1

538.4

544.1

45

0.2

263

.56

78.4

7

100

3.3

05

.16

7.4

310

.12

13.2

116.7

22

0.6

524.9

829.7

33

4.8

940.4

746.4

65

2.8

666

.90

82.5

9

120

3.6

35

.68

8.1

811

.13

14.5

318.3

92

2.7

127.4

832.7

03

8.3

844.5

151.1

05

8.1

373

.58

90.8

4

140

3.9

66

.19

8.9

212

.14

15.8

520.0

62

4.7

729.9

735.6

74

1.8

648.5

555.7

36

3.4

180

.26

99.0

8

160

4.2

96

.71

9.6

613

.15

17.1

721.7

32

6.8

332.4

738.6

44

5.3

552.5

960.3

76

8.6

986

.94

107

.33

180

4.6

27

.22

10.4

014

.16

18.4

923.4

02

8.8

934.9

641.6

14

8.8

356.6

365.0

17

3.9

793

.61

115

.57

200

4.9

57

.74

11.1

415

.17

19.8

125.0

73

0.9

537.4

644.5

75

2.3

160.6

769.6

57

9.2

410

0.2

9123

.82

250

5.7

89

.03

13.0

017

.69

23.1

129.2

53

6.1

143.6

952.0

06

1.0

270.7

781.2

49

2.4

411

6.9

9144

.43

300

6.6

010.3

214.8

520

.22

26.4

133.4

24

1.2

649.9

359.4

26

9.7

380.8

792.8

41

05.6

313

3.6

9165

.05

350

7.4

311.6

016.7

122

.74

29.7

137.6

04

6.4

256.1

666.8

47

8.4

490.9

7104

.44

118.8

215

0.3

9185

.66

1. B

ase

d o

n N

ap

ier's f

orm

ula

Gau

ge

Pre

ssu

re

(kP

ag

)

NO

ZZ

LE

DIA

ME

TE

R (

mm

)

FL

OW

RA

TE

(kg

/h)1

Ta

ble

2.7

.1 –

Ste

am

flo

w t

hro

ug

h n

ozzle

s (

Me

tric

)1

Section 2.7




If the feedwater temperature is too cold, steamdemand increases. This can create excessivesteam flow demand to maintain operatingpressure/temperature. The excessive steam flowcan damage vessel internals and cause feedwaterspray carried through the vent. It can also causeflooding of tray type deaerators.

When superheated steam is used instead ofsaturated steam, the significantly higher heatcontent of the steam results in a lower steamflow to heat the water to saturation temperature.The lower steam flow means that there is alower volume of steam available for scrubbingthe incoming water and for removing thenoncondensable gases through the vent. Thismay result in a higher dissolved oxygen concen-tration in the effluent. In order to overcome thissituation, the vent rate should be increased sothat oxygen can be removed. In such cases,desuperheating the steam should be considered.

Heating cold, air-saturated makeup water maygenerate free air in the line due to its decreasedsolubility at elevated temperatures. This free airmust be removed from the makeup water beforeit reaches the deaerator, or free air will bereleased into the deaerator. This can cause airblanketing, which can impair heating anddeaeration. A vented atmospheric receiver willeliminate this problem, if located after all heatrecovery equipment and any condensate returnthat will be blended with the makeup.

Check flows of water entering deaerator –A modulating control valve should be used tomaintain water level in the deaerator. Modulat-ing control will give stable operating conditionsversus on/off valves, which will provide asudden rush of colder water and may profoundlyaffect pressure control.

All water supplies entering the deaerator musthave sufficient pressure to overcome any loss ofhead caused by pipe friction, control valves, ventcondenser, and spray valves. If the pressure istoo low, sufficient water will not enter the

The presence of a 1/8 inch (3.2 mm) hole in theseat of the vent valve will always preventoxygen free deaerator lay-up.

Check temperature of water enteringdeaerator – Deaerators are designed to operatewithin specific water temperature ranges. Theyare designed for a certain amount of makeupwithin a specific temperature range and a certainamount of condensate within a specifictemperature range. Operating outside of thesedesign parameters (flow or temperature) canresult in poor deaerator performance. Changesin any one of the flows or correspondingtemperatures will affect the mass and energybalance of the entire system. Gradual changescan be expected and, as long as they are withinspecifications, should cause no decrease inperformance.

Abrupt flow or temperature changes can alsoadversely affect performance and even causeinternal deaerator damage. The cause of theseabrupt changes should be determined andcorrected or eliminated. Abrupt flow changescan be mollified by adding surge capacity.

Heat recovery projects are sometimes added assystem upgrades after the deaerator has beendesigned and installed. This can cause deaeratorperformance issues. If steam demand andconsumption are lessened by a retrofit heatrecovery system, which now delivers feedwaterat a significantly higher temperature, steamdemand may decrease to a point where anadequate steam flow for stripping oxygen andother noncondensables is not maintained. Atemperature differential of 30-50°F (17-28°C) istypically required between the incoming water(makeup plus condensate) and the operatingtemperature for spray type deaerators. Thisdifferential is typically 10-20°F (6-11°C) fortray type deaerators. Check the original equip-ment specifications for the actual differentialrequired.

Section 2.7




deaerator, and distribution through the spraynozzles will be poor. If the pressure is too high,difficulty may be experienced with the inletcontrol valve due to an excessive pressure drop.

Many plants are currently operating underreduced loads. Feedwater flow rates signifi-cantly below design specifications will notatomize properly through the spray nozzles,causing poor deaerator performance. Invertingselectively chosen nozzles or retrofitting thedeaerator with smaller nozzles (contact theequipment manufacturer for proper sizing) maybe required if a spray deaerator is operatedconsistently below 25% of design capacity.Tray type deaerators can typically operateeffectively down to 10% of design capacity.

The feedwater flow through the deaerator shouldnot be higher than the design water flow. At flowrates in excess of design, there is too much waterflowing through each spray nozzle to obtain thefine mist/film required for good steam/watercontact, and effluent oxygen rises.

Check location of condensate return lines –Condensate return and makeup water should beintroduced into the deaerator in such a manneras to avoid extreme temperature interfaces. Theseparate introduction of condensate return andcold makeup streams into the deaerator waterbox MUST be avoided. The amount of mixingthat actually occurs in a water box is minimal;thus, the water spraying out of the water boxinto the heater area tends to divide the heaterinto a cold zone and a hot zone. This affects traydeaerators severely, creating an imbalance ofsteam flow through the tray stack, with thesteam gravitating to the cold side. Performanceloss will vary directly with the temperaturedifference.

Condensate returns that are hotter than theoperating temperature in the deaerator dome canbe expected to generate flash steam. This cancreate water hammer at the interface between the

flashed steam and lower temperature water, ifthis condensate is returned to the water box.A vented atmospheric receiver for all returnsand makeup can go a long way towards theattainment of system stability. An atmosphericreceiver should be located downstream of allheat recovery equipment.

If condensate at flashing temperatures isreturned to the deaerator, it should be introducedinto the heater section outside the water boxthrough a T baffle or other energy absorbingdevice. Larger volumes may be returned to thedeaerator dome at a location approximatelyequivalent to the steam inlet. The hot condensatewill flash, adding to the stripping steam present.Problems can arise if the supply of this flashsteam varies significantly. In such cases,the steam supply valve must be capable ofhandling the fluctuations quickly so that domepressure does not fluctuate by more than0.5 psig (3.4 kPag).

Check steam pressure and flow rate – A modu-lating control valve is typically used to regulatethe steam supply to the deaerator. This valve ismodulated by means of a pressure controller. Afast acting, pneumatically actuated control isused in most systems, although a pilot operatedpressure control valve may be suitable in smallerapplications and a self-acting, diaphragmactuated control valve may be used in systemswhere the load is constant. Remember, deaeratortemperature is actually the result of how well thepressure is controlled, so it is very important thatthe valve be able to maintain a constant pressure.

Instruments should be tuned with a relativelynarrow proportional band and a very slow resetto minimize pressure and temperature variations.When flash steam from hot condensate return isused as part of the steam supply, a pressuresensor at the supply source should be installed tosignal the pressure reducing valve if supply islost.

Section 2.7




The amount of steam required by any deaeratoris a combination of the amount consumed toheat all the incoming water to the saturatedsteam temperature within the unit, plus a smallamount that is vented with the noncondensablegases, less any flashed steam generated from hotcondensate or trap returns. This can be estimatedwith a heat balance of all incoming and outgoingstreams. (See Examples 2.7.1 and 2.7.2.)

A general rule of thumb for a typical plant is that15% of the boiler feedwater is condensed steamfrom the deaerator. In a utility or where makeupwater is hot process softener effluent, thisnumber will be lower due to the higher tempera-ture of the makeup water. If superheat exists inthe heating steam, flow is decreased approxi-mately 1% for every 20°F (11°C) superheatpresent. The steam requirement will be higherwhere there is a high percentage of cold makeupwater used. Approximately 0.05-0.5% of thetotal steam requirement is needed for venting.

Check dissolved oxygen – Dissolved oxygenmonitoring at the ppb concentration requiresexacting sampling and analytical techniques toachieve accurate results. Dissolved oxygenmonitors require regular maintenance and shouldbe checked by wet chemical methods on aregular basis. Acceptable wet chemical methodsare the CHEMetrics ampoules and the HachIndigo Carmen dissolved oxygen tests. Thefrequency of the wet chemical testing should bebased on operating experience, but should not beless frequent than once per month, ideally onceper week. Sampling and operator testing tech-niques are very important for accurate, preciseoxygen values. All operators running dissolvedoxygen tests should be well trained.

Dissolved oxygen should be determined withand without oxygen scavenger present. In theabsence of scavenger, the oxygen study isreferred to as a deaerator efficiency test, since itfocuses only on equipment performance orefficiency. It will typically take 2-3 hours to

dilute the oxygen scavenger totally out of thestorage section, even if the holding time in thestorage section is relatively short. This is be-cause mixing is usually very limited in thedeaerator storage section. Figure 2.7.4 illustratesan oxygen scavenger bypass feedpoint, whichshould ideally be installed so deaerator oxygentesting can be performed without shutting off theoxygen scavenger feed. Testing should still bedone even if a bypass is not present.

Dissolved oxygen samples should be taken fromthe storage section dropleg. If two deaeratorsfeed into a common storage section, samplelines from each deaerator dropleg or neck to thestorage section should be installed, in addition tothe storage section dropleg sample. A dissolvedoxygen sample should also be available afterthe boiler feedwater pump, because this is acommon contamination point.

Section 2.21 discusses proper sampling in detail.In general, all sample lines must be cooled tobelow 100°F (38°C) ideally below 90°F (32°C),they should be stainless steel construction,and they should flow continuously for at least2-3 hours before taking a sample. Copperconnections and lines should never be used,because copper is catalytic to most scavenger/oxygen reactions. Any non-metal connection

Figure 2.7.4 – By-pass oxygen scavenger

feedpoint

Section 2.7




EXAMPLE 2.7.1 – DEAERATOR ENERGY BALANCE – U.S. UNITS

System Given From Steam Tables

Boiler 145 psig hg = 1195 Btu/lb

Steam load 21164 lb/hr

Deaerator 4.4 psig, 225°F hf = 193.7 Btu/lb, h

g = 1155.4 Btu/lb

Condensate 194˚F, 75% returned hf = 162 Btu/lb

Makeup water 77˚F hf = 45 Btu/lb

At 25 cycles of concentration (COC), the boiler will need 22046 lb/hr of water supplied bythe deaerator.

Feedwater = (Steam)(COC)/(COC - 1) = (21164 lb/hr)(25)/(25 - 1) = 22046 lb/hr

The enthalpy of the makeup water and condensate are found in the steam tables and are shownabove. These streams must have their enthalpy raised to 193.7 Btu/lb, which is the operatingcondition of the deaerator.

Heat needed by makeup = mmu

(hf da

- hf mu

) = (25%)(22046 lb/hr)(193.7 - 45) Btu/lb= 820000 Btu/hr

Heat needed by condensate = mcond

(hf da

- hf cond

) = (75%)(22046 lb/hr)(193.7 - 162) Btu/lb= 520000 Btu/hr

Total heat needed = 820000 + 520000 = 1340000 Btu/hr

The heat is supplied by condensing steam at 145 psig to water at 4.4 psig and 225°F. Heat suppliedby the steam = (h

g stm - h

f da) = (1195 - 193.7) Btu/lb = 1001.3 Btu/lb

Therefore, the steam flow required to the deaerator for heat = 1340000 Btu/hr/1001.3 Btu/lb= 1338 lb/hr.

A small portion of the steam is also vented. This is typically less than 0.2% of the feedwater flow,or 44 lb/hr for this example. Since it is vented as steam, we use h

g instead of h

f. This number is

small compared to the amount required for heat and can be omitted in most circumstances.

Steam vented = (mvented

)(hg da

)/hg stm

= (44 lb/hr)(1155.4 Btu/lb)/1195 Btu/lb = 43 lb/hr

The total steam flow to the deaerator is then a combination of the steam required to heat themakeup and condensate plus the amount lost to venting.

Total steam flow = 1338 + 43 = 1371 lb/hr

Section 2.7




EXAMPLE 2.7.2 – DEAERATOR ENERGY BALANCE – METRIC UNITS

System Given From Steam Tables

Boiler 1000 kPag hg = 2780 kJ/kg

Steam load 9600 kg/hr

Deaerator 30 kPag, 107°C hf = 450 kJ/kg, h

g = 2687 kJ/kg

Condensate 90°C, 75% returned hf = 377 kJ/kg

Makeup water 25°C hf = 105 kJ/kg

At 25 cycles of concentration (COC), the boiler will need 10000 kg/hr of water supplied bythe deaerator.

Feedwater = (Steam)(COC)/(COC - 1) = (9600 kg/hr)(25)/(25 - 1) = 10000 kg/hr

The enthalpy of the makeup water and condensate are found in the steam tables and are shownabove. These streams must have their enthalpy raised to 450 kJ/kg, which is the operatingcondition of the deaerator.

Heat needed by makeup = mmu

(hf da

- hf mu

) = (25%)(10000 kg/hr)(450 - 105) kJ/kg = 860000 kJ/hr

Heat needed by condensate = mcond

(hf da

- hf cond

) = (75%)(10000 kg/hr)(450 - 377) kJ/kg= 550000 kJ/hr

Total heat needed = 860000 + 550000 = 1410000 kJ/hr

The heat is supplied by condensing steam at 1000 kPag to water at 30 kPag and 107°C. Heatsupplied by the steam = (h

g stm - h

f da) = (2780 - 450) kJ/kg = 2330 kJ/kg

Therefore, the steam flow required to the deaerator for heat = 1410000 kJ/hr/2330 kJ/kg= 605 kg/hr.

A small portion of the steam is also vented. This is typically less than 0.2% of the feedwater flow,or 20 kg/hr for this example. Since it is vented as steam, we use h

g instead of h

f. This number is

small compared to the amount required for heat and can be omitted in most circumstances.

Steam vented = (mvented

)(hg da

)/hg stm

= (20 kg/hr)(2687 kJ/kg)/2780 kJ/kg = 19 kg/hr

The total steam flow to the deaerator is then a combination of the steam required to heat themakeup and condensate plus the amount lost to venting.

Total steam flow = 605 + 19 = 624 kg/hr

Section 2.7




should have low oxygen permeability. (SeeTable 2.7.2.) Testing should continue until threesamples provide approximately equivalentvalues over a period of 30 minutes.

If high oxygen is noted and confirmed bytesting, the troubleshooting process shouldbegin. Vent rates, dome temperature versuspressure, and dome temperature versus storagetemperature are usually the first items checked.A number of other potential causes have alsobeen discussed.

Two items that have not yet been mentioned area leaking sample line and a leaking deaeratorbypass valve. The sample line might leak airinto the sample, or it might draw cooling water(with oxygen) into the sample. The sample lineshould be throttled only at the outlet to maintainpressure throughout its entire length. A leakingdeaerator bypass valve will allow undeaeratedmakeup water to short circuit the deaerator intothe feedwater line. Many new systems no longerhave the ability to bypass the deaerator, so thisproblem is typically (but not always) noted onolder construction.

Evaluate Chemical Feed and Control

Procedures

Use proper chemical program – Consider theneeds of the system, and then fit the best productfor the application. Sulfite should typically beused for most systems operating below 250 psig(1.7 MPag). These systems usually place littlevalue on differentiating benefits such as passiva-tion or low product solids contribution providedby carbohydrazide, diethylhydroxylamine(DEHA), erythorbic acid, or hydroquinone(HQ). Sulfite should not be used at pressuresabove 900 psig (6.2 MPag), because break downto the corrosive gases hydrogen sulfide (H

2S)

and sulfur dioxide (SO2) can be appreciable.

Sulfite cannot be used in feedwater that is usedfor attemperation or desuperheating, because thesulfite solids will form deposits.

Check corrosion products in feedwater –Corrosion product concentrations in the feed-water should be determined regularly usingNalco AP-088, the extended low-level total ironprocedure, and AP-022, the low-level Fe (II)procedure. The Fe (II) test measures the freshproducts of corrosion reactions in low oxygenenvironments. The Fe (II) test does not reactwith iron in the Fe (III) oxidation state or iron inthe form of oxide particles. Fe (II) reacts withoxygen to form Fe (III), so this test has limiteduse during deaerator upsets.

Sampling technique is critical. Sample linesmust be stainless steel, and samples must becooled and must flow continuously for manyhours before the sample is taken. Additionaldetails on sampling procedures are discussed inSection 2.21 of this manual.

Sample points should be located at the inlet andoutlet of all feedwater heaters and economizers.The goal of any oxygen scavenger program is tominimize corrosion. Without measuring corro-sion products, you will not know if you areachieving this goal or not.

Check chemical dosage – Calculate the theoreti-cal oxygen scavenger dosage based on dissolvedoxygen versus the actual amount being fed. Theactual amount fed is best measured by a pumpdrawdown cylinder mounted on the suction lineof the chemical feed pump. Calculations for thetheoretical dosage of a given product are foundin the Confidential Product Profile (CPP).Deviations from theoretical can be caused bydecreased deaerator performance (higher oxy-gen), oxygen inleakage at the feedwater pump orminimum flow line, and higher feedwater flows.

Check chemical feed point – Feed all oxygenscavengers to the neck between the dome andstorage section of the deaerator using aNALQUILL® injector. This provides optimummixing and retention time for the scavenger.Injection into the neck, however, requiresthe injection quill shown in Figure 2.7.5

Section 2.7




Table 2.7.2 – Leakage of oxygen through plastic tubing

Common plastic materials used for sample tubing show leakage of oxygen into water

flowing inside the tubing due to the permeability of the material. The following list shows

least permeable to most:

O2 pickup in water; O

2 pickup in water;

Material ppb/meter ppb/lineal foot

Polyvinylidene chloride (Saran) 0.02 0.006

Nylon 0.03 0.009

Polychloro trifluoroethylene (Kel-F) 0.05 0.015

Polyvinyl fluoride (Tedlar) 0.05 0.015

Polyvinylidene fluoride (Kynar) 0.1 0.030

Polyethylene Terephthalate (Mylar) 0.12 0.037

Polyvinyl chloride (Non-plasticized) 0.14 0.043

Polyacetal (Delrin) 0.2 0.061

Ethylene/Monochlorotrifluoroethylene

copolymer (Halar) 0.43 0.13

Ethylene/Tetrafluoroethylene copolymer

(Tefzel) 1.7 0.52

High density polyethylene (opaque) 2.04 0.62

Polypropylene 3.2 0.98

High density polyethylene (clear) 3.9 1.2

Polycarbonate (Lexan) 5.1 1.6

Polystyrene 5.3 1.6

Low density polyethylene 8.5 2.6

Fluorinated ethylene/propylene (FEP) 13 4.0

Tetrafluoroethylene (PTFE) 19 5.8

Natural rubber (Latex) 60 18

Silicone rubber (Silastic) 1700 518

Data reported in Application Note 1.06, Orbisphere Laboratories

Section 2.7




(231-P4615.88). A retractable version is alsoavailable (231-P4635.88).

If the particular deaerator design does not havean external drop leg (neck) between the domeand the storage section or if there is any questionwhether the drop leg is commonly full of water,scavenger can be fed into the deaerator storagesection. Multiple feedpoints may be necessarywhen feeding to the storage section in order toadequately distribute the product. This is accom-plished by feeding to both ends of the deaeratoror by means of an internal header that distributeschemical to both ends. (See Figure 2.7.6.)

All products must be fed continuously. A by-pass should be installed around the oxygen

sample point so oxygen sampling can occurwithout shutting off the scavenger feed. (SeeFigure 2.7.4.)

Check chemical preparation and feeding – Allliquid products should be fed neat. Dry productsmust be prepared properly to minimize activityloss. Turn the agitator off as soon as mixing iscomplete. Use a floating cover on the solutiontank. Use high quality condensate or demineral-ized makeup as the water source wheneverpossible. Never mix other products in with anoxygen scavenger. This is especially importantfor catalyzed sodium sulfite/bisulfite productssince polymers, chelants, phosphates, andalkalizing agents (e.g., caustic or amines) willconsume or precipitate the sulfite catalyst.Failure to follow these recommendations willcause a loss in product activity.

Check feedwater pH after scavengerinjection – Most chemical scavengers requirea minimum pH of 8.5 for optimum oxygenremoval. Acid regenerant breakthrough,demineralized or reverse osmosis quality make-up, raw makeup changes, condensate problems,untreated condensate, removal of amines fromcondensate treatment program, and high oxygenscavenger dosages can reduce the feedwater pHto unacceptable levels.

DAILY OPERATION CHECKS

The following items should be checked daily toensure that the deaerator is functioning properly.

Temperature of the water – The temperature ofthe dome should be within 1-2°F (0.6-1.1°C)of the theoretical saturation temperature forthe operating pressure of the deaerator. Thetemperature of the dome should be within 4°F(2.2°C) of the storage temperature.

Deaerator venting – The plume should bevisible, strong, steady (not puffing), and withoutentrained water.

Figure 2.7.5 – NALQUILL with no-drip shield

(231-P4615.88)

Figure 2.7.6 – Alternate oxygen scavenger

feedpoint. 1/2 inch (13 mm) pipe with 1/8 inch

(6.4 mm) hole in end caps

Supports

Drop leg

Water line

Deaerator

Oxygenscavenger

supply

stainless steelpipe or tubingcapped at theends with holein end caps

Deflection guard

Drip guard

Reference mark to indicate

downstream side of quill

3/16”(0.5 cm)

(30.5 cm)12” Ref.

5/8” (1.5 cm)

60°

2 1/2” Ref.(6.4 cm)

Section 2.7




Water level gauge glasses – These should bechecked to ensure that an adequate storage levelis being maintained by the makeup regulatingvalve.

Overflow valve – The overflow valve should beclosed and not bleeding off steam. If the unit hasa loop seal, check that this has not blown andthat steam is not being lost for lack of a waterseal.

Oxygen scavenger level – Correlate scavengerresiduals/demand to normal deaerator operation.A severe change in treatment demand maysignify a deaerator problem.

Feedwater oxygen – Feedwater oxygen shouldideally be checked once/day minimum withoxygen scavenger on.

Deaerator pH – Deaerator (feedwater) pHshould be checked once/day minimum andmaintained at 8.5 or above.

Usually, these simple checks will verify that thedeaerator is operating correctly.

INSPECT DEAERATOR DURING

OUTAGES

In many plants, a single deaerator serves mul-tiple boilers. The boilers are dropped off-lineindividually for inspection or maintenance, butthe deaerator stays in service. Sometimes, yearsgo by without a single deaerator inspection.Deaerators are vital pieces of plant equipmentand should be inspected annually. In fact, someinsurance companies are now requiring yearlyinspection of the deaerator as well as the boilers.

Although it may require some temporary plumb-ing, provision can usually be made to bypass thedeaerator if it is necessary to keep the boilersrunning. Higher concentrations of oxygenscavenger are typically fed during this time.

Usually, the deaerator turnaround is kept as shortas possible; however, a temporary deaerator ordeoxygenation equipment can be brought intothe plant for longer outages or repairs.

During the inspection, be sure to check thefollowing.

Confirm all safety precautions beforeentering – Inspection of a deaerator requiresentry into a confined space and all appropriatesafety procedures, such as “lock out, tag out”must be followed. Check all valves to make surethat they are tightly closed and tagged properly.Check for safe oxygen levels before enteringdeaerator storage or dome sections. Follow allapplicable safety procedures.

Determine deaerator type – There are threemain types of deaerators – spray, tray, andatomizing. (See Figures 2.7.7-2.7.9.) These arecovered in detail in the PAC-2 Boiler WaterTechnology Manual.

The spray type uses spring loaded spray valvesto break the water into a thin film for primarydeaeration. This water then falls to a secondarystage, where the pressure is dropped to cause thewater to boil and remove the remaining gases.

In a tray type unit, the water enters the same asin the spray type for primary deaeration. Thewater then falls over a series of trays to break itinto a thin film so that steam can remove finaltraces of gases.

The atomizer type sprays the water as tinyparticles into an atmosphere of steam. Fromthere, it falls to the atomizer where a very highvelocity steam jet turns the water to a mist,providing final deaeration.

Check trays for proper positioning andcleanliness – The trays in a tray type deaeratorcan fill with corrosion products or can be tornfrom their proper positions by water hammer or

Section 2.7




Figure 2.7.7 – Spray type deaerator

Section 2.7




Figure 2.7.8 – Tray type deaerator

Section 2.7




Figure 2.7.9 – Atomizing type deaerator

Section 2.7




rapid flow changes. Check all trays to make surethat they are secure in their positions and level(unless otherwise specified). Make sure alltrays are clean and not clogged with corrosionproducts.

Check each spray nozzle – Spray nozzles havethree parts – mounting, gasket, and nozzle(Figure 2.7.10). All nozzles should close tightly.A hanging nozzle indicates a broken or weakspring. The nozzle area should be free of depos-its. Spring pressure should be checked on eachnozzle to insure proper operation. All nozzlesshould have an equal spring pressure. If not, thewater will channel through the loose nozzles,and optimal water distribution will not beachieved. Spring calibration weights are nor-mally provided by the manufacturer when thedeaerator is installed. When the weight is hungfrom the nozzle, it should just begin to open.

Spring tension should be adjusted accordinglyon any nozzle, which either does not open oropens too much.

All gaskets should be inspected. They should bein place and secure. There should be no crackedgaskets.

A dark or shiny area on a wall or around anozzle can indicate a problem with the spraypattern. This can be caused by overloading thedeaerator or by a malfunctioning spray nozzle.

Record the number of nozzles repaired.

Inspect for corrosion and deposits – Theamount and location of all corrosion anddeposits should be recorded.

Inspect vessel welds for cracking – All weldsmust be visually inspected for cracking. Inaddition to the visual inspection, the weldsshould be inspected by the non-destructive WetFluorescent Magnetic Particle technique on aregular basis. Inspection data accumulated bythe NACE International have shown that morethan 35% of the vessels, which had properinternal inspection, have been cracked severelyenough to require weld repairs.

Inspect injection and recirculating ports –These ports should be free from obstructions.A dark spot across from the port may indicateexcessive velocity of the flow.

Check vent – Make sure the vent is straight upfrom the deaerator and is free from restrictions.Make sure any vent condensers are free of leaks.

Check waterline position – Insure that thewaterline in the deaerator storage section is at itsproper level.

Figure 2.7.10 – Typical deaerator spray nozzle

assembly

Section 2.7




EVALUATE LAY-UP PROCEDURES

Lay-up procedures for deaerators are covered inthe American Society of Mechanical Engineers(ASME) document CRTD Volume 66, Consen-sus for the Lay-Up of Boilers, Turbines, TurbineCondensers, and Auxiliary Equipment. Thisdocument is posted in the Knowledge Manage-ment (KM) database.

The following paragraphs are from this consen-sus document. They are copyrighted by ASMEand reprinted here with their permission.

12.2 Lay-up of Deaerator and Storage TankA number of methods may be used to properlylay-up the deaerator and deaerator storagetank. It is preferred to maintain a steamblanket on the equipment, as this ensures somehot deaerated water for the subsequent startup.If steam is not available then the deaerator andstorage tank may be:(1) Blanketed with a small continuous flow (30 scfh; 0.85 m3/h) of nitrogen(2) Drained while hot and maintained dry with dehumidified air or desiccant(3) Filled to the vent with water containing volatile oxygen scavenger and either ammonia or amine, as previously described for feedwater heaters

12.1 Lay-up of Feedwater HeatersThe tubeside of feedwater heaters should betreated the same as the boiler during lay-upperiods. The shellside normally is steamblanketed or dried and pressurized withnitrogen during lay-up. However, the shellsidemay be flooded with treated condensate orcondensate-quality water. All-steel systemsshould be laid up with condensate containing200 mg/L (ppm) volatile oxygen scavenger orequivalent, and ammonia or suitable aminesuch as cyclohexylamine as needed to main-tain the pH at 10.0 minimum. For copper orcopper alloy systems, use 50-100 mg/L (ppm)of volatile oxygen scavenger, and adjust thepH to 9.5.

EVALUATE START-UP PROCEDURES

Start-up procedures introduce an upset conditionregardless of how the start-up is accomplished.The method used should minimize the upset asmuch as possible. Improper, rapid start-ups aftercold or hot lay-ups can produce conditionsfavoring corrosion fatigue cracking at the weldsand can damage vessel internals, which willprevent good deaeration right from the start.

The following comments are not meant toreplace the equipment manufacturer’s recom-mendations for equipment start-up procedures.

Provide initial feedwater near saturationtemperature – An auxiliary recirculation loopfrom deaerator storage to the main inlet water-line, as shown in Figure 2.7.11, will purge steamfrom the water box, which is created during ahot stand-by condition. Recirculation willeliminate those problems caused by suddensteam pressure collapse from cold feedwaterentering the water box. Start the feedwater flowslowly (10-15% of normal flow). Increase theflow of cold feedwater gradually, watching thedome temperature and pressure and not allowingthe pressure to vary more than ±0.5 psig

Figure 2.7.11 – Deaerator water box purge loop

Section 2.7




(3.4 kPag) from normal. The purging loop isgradually shutdown as the feedwater flowincreases and its temperature approaches thesaturated steam temperature.

Introduce steam first – Start the flow of inletwater and slowly increase from 50-60% ofdesign rate. Place the deaerator steam controlvalve on manual, and heat the unit gradually.Care must be taken as the water temperaturepasses atmospheric boiling and pressure is

established in the deaerator. Flashing in theboiler feedwater suction line may result insevere upsets if boiler feed pump net positivesuction head is marginal. Once pressure isestablished in the deaerator, it should be in-creased at a slow rate so the deaerator storagetank water temperature is maintained within 9°F(5°C) of the saturation water temperature at thepressure in the deaerator.

DEAERATOR PERFORMANCE OPTIMIZATION

Deaerator optimization using data logging oxygen analyzers such as the OrbispherePowerLogger can provide significant customer value. Optimization of the deaerator is usuallylimited to operational parameters such as vent rate, steam flow, feedwater flow, deaeratorpressure, and feedwater pump(s) in use. Chemical parameters such as scavenger, scavengerfeed rate, and feedwater pH can be optimized. Mechanical parameters such as vent size,internal nozzle type and condition, and tray type and condition cannot be changed during plantoperation and are not candidates for performance optimization.

The actual process consists of connecting the analyzer with logger to a cooled sample point inthe feedwater. Scavenger feed should be discontinued approximately 2-3 hours before the startof the study. This allows baseline data on current deaerator performance to be collected. Theoperational and chemical parameters of interest are then changed one at a time and the effecton feedwater oxygen monitored. Oxygen scavenger may be restarted sometime during thisportion of the study, if the intent is to note the effect of these parameters on scavenger perfor-mance. (See Figure 2.7.12.)

Oxygen must be logged over a period long enough to produce a well-defined correlation to theparameter that was changed. The optimum conditions will produce feedwater with 7-10 ppbdissolved oxygen, which is the commonly accepted performance standard for a pressuredeaerator.

Dissolved oxygen greater than 10 ppb increases the chance of system corrosion, especially ifthe feedwater system includes an economizer, and increases the cost of scavenging. Dissolvedoxygen less than 7 ppb results from increased steam use, venting, and operating pressure.Cost savings relative to increased steam use and venting can easily be quantified based onExamples 2.7.1 and 2.7.2.

Section 2.7




FEEDWATER TANKS

Feedwater tanks are typically installed in low-pressure plants (usually less than 100 psig;689 kPag). They are a much simpler piece ofequipment than a deaerator and mechanicalremoval of dissolved oxygen will be much lessefficient.

THE SURVEY

The first step in servicing or troubleshooting afeedwater tank is to complete an MOC survey.The list of appropriate questions is somewhatshorter than the list for deaerators. The surveyshould include questions in each of the MOCareas. Answers to the MOC questions will helpestablish expected versus existing performanceof the feedwater tank, as well as identifypotential problem areas.

Mechanical questions:

• For what boiler(s) or system does this specificfeedwater tank provide feedwater?

• How many feedwater tanks are present in thissystem?

• What is the design flow rate through thefeedwater tank?

• What are the design temperatures and flowsfor all makeup and condensate streams?

• Does condensate and makeup mix before thefeedwater tank or enter separately?

• What is the return (entry) point for any high-temperature condensate streams?

• Where are the sample points located, and whatare their estimated lag times based on linelength and diameter?

• What is the sample line metallurgy?

Figure 2.7.12 – Deaerator performance study

Section 2.7




• What dissolved oxygen concentration is thefeedwater tank designed to supply?

Operational questions:

• What is the actual water flow through thefeedwater tank?

• What is the condition of the vent (steady,sputtering, an invisible space between the topof the vent and the plume, etc.)?

• What is the source, pressure, and temperatureof the steam used by the feedwater tank?

• What is the tank temperature?

• What is the temperature and flow rate of allhigh-temperature condensate streams?

• What is the temperature and flow rate of allmakeup and low-temperature condensatestreams?

• Do sample lines flow continuously or only asneeded? If only as needed, how much flushtime is provided before sampling?

• What are the date and findings of the latestoxygen testing done without scavenger?

• What are the date and findings of the lastinternal inspection, e.g., tank, welds, etc.?

Chemical questions:

• What products/chemicals are fed to thefeedwater tank?

• What is the dosage of these products?

• Are the products fed neat or diluted?

• Are multiple products mixed together beforeor at the injection point?

• What are the point(s) of injection for allproducts fed into the feedwater tank?

• Is a quill used, and how far does it extend intothe feed point?

• What is the metallurgy of the feed system?

• What is the measured dissolved oxygenconcentration in the feedwater tank effluentwith oxygen scavenger?

• What is the feedwater pH in the feedwatertank after scavenger injection?

TROUBLESHOOTING POOR

PERFORMANCE

Figure 2.7.13 shows a well designed feedwatertank. Condensate returns through a sparge linenear the bottom of the tank. Makeup entersthrough a sparge line just under the waterline.The density difference between the cold makeupand hot condensate allows some natural circula-tion to occur. A steam sparge should also bepresent to control tank temperature and willprovide additional mixing.

Sample points should be available for all incom-ing and outgoing lines, e.g., condensate,makeup, and feedwater discharge. Any sampleconsistently above 100°F (38°C) requires asample cooler.

Evaluate Original Design and Installation

Obtain original design parameters andfeedwater tank specifications – Troubleshootingbegins at the same point used for deaerators –evaluation of existing equipment and installationagainst the original design and installation. Aspreviously, obtain original design parametersand specifications for the feedwater tank.Record all manufacturer’s design criteria andspecifications on the survey form. Obtain anoriginal drawing of the vessel. Note the posi-tions of all inlets, outlets, baffles, gauges, andsensors.

Obtain current system diagram and determinecurrent operating parameters – Go to thefeedwater tank and prepare a drawing of theunit, as it currently exists. Note the positionof all inlets, outlets, gauges, thermocouples,and other sensors. Obtain accurate flow and

Section 2.7




temperature data for all influent streams and forthe steam supply. Note flow swings. Do notaverage flow information but record the actualflow ranges. Diagnostic TRASAR® can be usedfor calibration of flow meters and verification ofestimated flow rates.

Evaluate Current On-Line Performance

Check operating temperature – The feedwatertank must be kept at a temperature as high aspossible. Since the tank is not pressurized, thiswill be less than 212°F (100°C). Higher tem-peratures minimize the content of dissolvedoxygen and other gases. (See Figure 2.7.14).A constant temperature above 185°F (85°C),ideally above 195°F (91°C), should be main-tained with a supplementary steam sparge, ifnecessary.

If a large portion of makeup is used, heating thefeedwater can substantially reduce the amount ofoxygen scavenger required by lowering theoxygen content. (See Example 2.7.3.) Not onlyis there a savings in the amount of scavengerfed, but also a reduction in the amount of solidspresent in the feedwater. (Sulfite becomessulfate when it reacts with oxygen.) The lowersolids reduce the amount of blowdown neededto maintain chemical control limits in the boiler.

The cost of heating the feedwater tank withsteam is offset by the higher energy content(temperature) of the feedwater entering theboiler. The overall system energy requirementsfor the system stay essentially the same, exceptfor a small increase in radiation and vent lossesfrom the feedwater tank. Radiation losses can be

Figure 2.7.13 – Properly designed feedwater tank

Section 2.7




Figure 2.7.14 – Dissolved oxygen solubility vs. temperature

minimized by maintaining good insulation onthe tank.

As temperatures approach 212°F (100°C), thepossibility of feedwater pump cavitation in-creases. Water close to its boiling point can flashto steam in the low-pressure area at the eye ofthe pump impeller. Bubbles of steam are formedin this area and then collapsed when the pressurerises again at the pump outlet. Cavitation isnoisy and can quickly damage the pump. Rais-ing the feedwater tank as high as possible abovethe boiler and generously sizing the pipework onthe suction to the feedwater pump increases thetemperature at which cavitation is likely tooccur.

Check feedwater tank metallurgy – Cast ironand mild steel are the most common materials of

construction for feedwater tanks, but with thetypical operating conditions present, both areprone to oxygen corrosion. Type 304L stainlesssteel will greatly improve the life expectancy ofthe feedwater tank.

Check feedwater tank capacity – The feedwatertank serves as a hot condensate receiver, coldwater makeup addition point, hot feedwaterstorage reservoir, and feedpoint for variouschemical treatments. Ideally, the tank should belarge enough to hold a 1-hour supply offeedwater at maximum steaming rate. It shouldalso be large enough to accommodate peakcondensate returns or surges. Lastly, the largerthe tank, the longer the time available foroxygen/oxygen scavenger reactions to occur.

Ox

yg

en

, p

pm

Oxyg

en

, ccL

Section 2.7




EXAMPLE 2.7.3 – FEEDWATER TANK SULFITE DEMAND CALCULATION

Feedwater Tank Sulfite Demand Calculation

Oxygen Concentration

Temperature from Figure 2.7.14 (ppm)

Current 140°F (60°C) 4.7

New (Goal) 190°F (88°C) 1.5

The boiler is operating at 120 psig (827 kPag) and 20 cycles of concentration (COC).It is producing 15000 lb/hr (6804 kg/hr) of steam. Based on 20 cycles, it will require15789 lb/hr (7162 kg/hr) of feedwater.

Feedwater = (Steam)(COC)/(COC - 1) = (15000 lb/hr)(20)/(20 - 1) = 15789 lb/hr= (6804 kg/hr)(20)/(20 - 1) = 7162 kg/hr

Sulfite usage under these two conditions can be calculated using the oxygen and residual sulfitefactors found in the product CPP. For this example, we will use the following:

• Oxygen factor = 22.8 ppm product/ppm dissolved oxygen (DO)• Residual factor = 4.56 ppm product/ppm sulfite residual• Boiler water sulfite target = 45 ppm sulfite, as SO

3–2 (30-60 ppm SO

3–2)

Current Sulfite Usage

Sulfite Needed to Scavenge Oxygen

Feedwater dosage = (feedwater DO)(oxygen factor) = (4.7 ppm DO)(22.8 ppm product/ppm DO)= 107.2 ppm product

Sulfite Needed to Provide Boiler Residual

Feedwater dosage = (boiler sulfite residual/boiler cycles)(residual factor) = (45 ppm sulfite/20cycles)(4.56 ppm product/ppm sulfite) = 10.3 ppm product

The total sulfite product required in the feedwater is the sum of the sulfite needed to scavengeoxygen and the sulfite needed to provide the desired boiler residual.

Total feedwater dosage = 107.2 + 10.3 = 117.5 ppm product

Section 2.7




New Sulfite Usage

Sulfite Needed to Scavenge Oxygen

Feedwater dosage = (feedwater DO)(oxygen factor) = (1.5 ppm DO) (22.8 ppm product/ppm DO)= 34.2 ppm product

Sulfite Needed to Provide Boiler Residual

Feedwater dosage = (boiler sulfite residual/boiler cycles)(residual factor)= (45 ppm sulfite/20 cycles)(4.56 ppm product/ppm sulfite) = 10.3 ppm product

The total sulfite product required in the feedwater is the sum of the sulfite needed to scavengeoxygen and the sulfite needed to provide the desired boiler residual.

Total feedwater dosage = 34.2 + 10.3 = 44.5 ppm product

Sulfite Product Difference

The difference between the current sulfite dosage and the new (goal) is 117.5 - 44.5 = 73 ppmsulfite product.

The difference in the sulfite product usage is:

U.S. Units(73 ppm)(15789 lb/hr feedwater)/1000000 = 1.2 lb/hr or 28 lb/day

Metric Units(73 ppm)(7162 kg/hr feedwater)/1000000 = 0.52 kg/hr or 13 kg/day

EXAMPLE 2.7.3 – FEEDWATER TANK SULFITE DEMAND CALCULATION – (CONTINUED)

Plants with larger steam loads may not be able tosize the feedwater tank to meet the characteris-tics listed above. They should have an additionaltreated water storage tank or a condensate tankcapable of accommodating condensate returnvolumes.

Check vent – The feedwater tank must be ventedto prevent buildup of pressure and to dischargethe noncondensable gases removed from thewater. The vent should be fitted with a venthead, which incorporates an internal baffle toseparate entrained water.

The plume should be visible and steady, al-though it will typically not be as large or strongas for a deaerator.

Check entry point of makeup and condensate –As previously mentioned, condensate shouldenter through a sparge line close to the bottom ofthe tank. Makeup should enter through a spargeline just below the water surface of thefeedwater tank. This provides some naturalcirculation. Cold makeup that enters along thebottom will likely be drawn directly into thefeedwater line take-off, sending colder water tothe boiler.

Section 2.7




Both sparge lines should be constructed of 304Lstainless steel to prevent early replacement fromoxygen corrosion.

Check dissolved oxygen – The dissolved oxygenconcentration in the effluent of a feedwater tankwill depend on the temperature of the tank. Thehigher the temperature, the lower the oxygen.Refer to the previous discussion of dissolvedoxygen testing under Deaerators.

EVALUATE CHEMICAL FEED AND

CONTROL PROCEDURES

Check oxygen scavenger – Catalyzed sulfite isthe only acceptable chemical oxygen scavengerfor these systems. Systems vary widely. Steamsparging to a temperature above 185°F (85°C),residence time at temperature in excess of 20minutes, and product stoichiometry up to fivetimes theoretical may be required in individualcircumstances. Such high sulfite feed rates willrequire additional caustic feed to counteract theboiler water alkalinity consumption, if an acidicsulfite product is used. Adjusting the pH of themakeup water before the sulfite addition ensuresa rapid chemical reaction rate.

Check chemical dosage – Calculate the theoreti-cal dosage based on dissolved oxygen versus theactual amount being fed. The actual amount fedis best measured by a pump drawdown cylindermounted on the suction line of the chemicalfeed pump. Calculations for the theoreticaldosage of a given product are found in theConfidential Product Profile (CPP). Deviationsfrom theoretical can be caused by decreasedfeedwater tank temperature (higher oxygen),oxygen inleakage at the feedwater pump, andhigher feedwater flows.

Check treatment chemical feedpoint(s) –Oxygen scavenger must be fed to the feedwatertank, not before it or after it. Feed of scavenger

ahead of the tank will increase scavengerconsumption. Feed of scavenger to the feedwaterline after the tank will reduce scavenger con-sumption, but increase residual oxygen present.The presence of residual scavenger does notguarantee that all oxygen has been removed.

Treatment injection should be at the pointfarthest from the feedwater take-off so that theentire tank receives treatment and maximumreaction time is available. It should be injectedbelow the waterline using a NALQUILL®

injector.

All products must be fed continuously. A bypassshould be installed around the oxygen samplepoint so that oxygen sampling can occurwithout shutting off the scavenger feed.(See Figure 2.7.4.)

Check chemical preparation and feeding – Allliquid products should be fed neat. Dry productsmust be prepared properly to minimize activityloss. Turn the agitator off as soon as mixing iscomplete. Use a floating cover on the solutiontank. Use high quality condensate or demineral-ized makeup as the water source wheneverpossible. Never mix other products in withcatalyzed sulfite. Polymers, chelants, phos-phates, and alkalizing agents (e.g., caustic oramines) will consume or precipitate the sulfitecatalyst. Failure to follow these recommenda-tions will cause a loss in product activity.

Check feedwater pH after scavenger injection –Most chemical scavengers require a minimumpH of 8.5 for optimum oxygen removal. Rawfeedwater changes, condensate problems,untreated condensate, removal of amines fromcondensate treatment program, and high oxygenscavenger dosages reduce the feedwater pH tounacceptable levels.

Section 2.7




DAILY OPERATION CHECKS

The following items should be checked daily toensure that the feedwater tank is functioningproperly.

Temperature of the water – The temperature ofthe water in the dome should be above 185°F(85°C), ideally above 195°F (91°C), and shouldbe maintained with a supplementary steamsparge, if necessary.

Feedwater tank venting – The plume should bevisible, steady (not puffing), and without en-trained water.

Water level gauge glasses – These should bechecked to ensure that an adequate storage levelis being maintained by the makeup regulatingvalve.

Overflow valve – The overflow valve should beclosed and not bleeding off steam. If the unit hasa loop seal, check that this has not blown andthat steam is not being lost for lack of a waterseal.

Oxygen scavenger level – Correlate scavengerresiduals/demand to normal feedwater operation.A severe change in treatment demand versusnormal may signify an operational problem.

Feedwater oxygen – Feedwater oxygen shouldbe checked once/week minimum with oxygenscavenger on.

Tank pH – Tank pH should be checked once/dayminimum and maintained at 8.5 or above.

Usually, these simple checks will verify that thefeedwater tank is operating correctly.

INSPECT THE FEEDWATER TANK

DURING OUTAGES

Feedwater tanks are important to the plantoperation and should be inspected annually.During the inspection, be sure to check thefollowing:

• All injection and recirculating ports should befree from obstructions. A dark or shiny spotacross from the port may indicate excessivevelocity of the flow.

• Make sure the vent is straight up from thefeedwater tank and is free from restrictions.

• Insure that the waterline in the tank is at itsproper level.

• The amount and location of all corrosion anddeposits should be recorded.

• All welds should be visually inspected forcracking.

Section 2.7




This page has been intentionally left blank.

For your notes:

Section 2.7





For your notes:

Section 2.7





For your notes:

Best Practices Desaereator[1]

Documents

intentionally

average flow

damage vessel

manufacturers

dissolved

multiple products

chemical scavengers

check chemical