FROM OXYGENATION OF TURBINE DISCHARGES

Ri. J. RUANE . Tennessee Valley Authority, Chattanooga,

DR. SVEIN VIGANDER Tennessee Valley Authority, Norris, Tennessee.

ABSTRACT the use of oxygen injection The Tennessee Valley Authority is investigating oxygen concentrations in through fine-pore diffusers for increasing dissolved Dam. Oxygen injection is the turbine discharges from Fort Patrick Henry from the turbine intakes, being considered either immediately upstream into the turbines, or immedi- so that the oxygenated water is drawn directly ately downstream from the "boil" area in the tailrace. ) selection of a diffuser for Research is being conducted in.two phases: (1 best location for injecting the *injecting the oxygen, and (2) selection of the are being evaluated for oxygen. In Phase I, various commercial diffusers efficiency. Field and * possible plugging problems and for oxygen absorption apparently is not a significant laboratory tests indicate that diffuser-plugging approaching 100 percent can problem and that oxygen absorption efficiency with several commercially be achieved in a 42-foot height of bubble rise system will be evaluated available diffusers. In Phase II, a large-scale diffuser fall of 1973. at Fort Patrick Henry Dam during the summer and if results of this research A full-scale system will be installed at the dam and economically feasible. show this method of reaeration to be technically

INTRODUCTION through fine-pore diffusers TVA is investigating the use of oxygen injection in the turbine dis- for increasing dissolved oxygen (DO) concentrations charges from Fort Patrick Henry Dam. River at river mile 8.2 in The dam is located on the South Fork Holston from Kingsport, Ten- Sullivan County, Tennecssee, about 2.5 miles upstream (Figure 1). At nessee, and 10.4 river miles downstream from reservoir extends 10.3 miles normal maximum pool (elevation 1,263), the 291 Docket# 5' '3't-@ /3Ve Control #7 7i 70 112. Date 5!!/ 77 of Document. REGULATO1Y DOCKET FfLE tz' is:.

16

- 12 X D,8

, 4

0F JAN

Flour.

by the Stal that exists After an *servoir relc or oxygen correcting methods tL could be c dam. This Figure l Map of Project Area voir and p: Oxygen upstream and has a total volume of 27,100 acre-feet. The reservoir has 4,300 of the higl acre-feet of useful controlled storage between normal maximum pool level high initiaI and normal minimum pool level (elevation 1,258). Average streamflow at diffused a Kingsport was 2,522 cubic feet per second during the period 1925-1970. plants (4 The dam is equipped with two hydraulic turbines, each of which has a such an il generating capacity of 18,000 kw. Normal maximum discharge through each the needc turbine is 4,400 cfs. in.r.-wage The maximum depth of the pool immediately upstream from the dam is about 101 approximately 70 feet. The thermocline during summer and early fall is us- creasing: ually 5 to 20 feet below the surface of the pool. oxygen iT Low concentrations of DO occur in the turbine releases from Fort Pat- 000 for, rick Henry Dam duringtthe summer and fall each year (Figure 2). This prob- $220,00( lem is the combined result of water low in DO that enters Fort Patrick Henry sidered e Reservoir through low-level power intakes at Boone Dam upstream and an aeration additional depletion of oxygen in the hypolimnion of the stratified Fort Pat- This I. rick Henry Reservoir. This low DO problem adversely affccts a reach of Scale 0: stream immediately below the darn that has been classified as a trout stream fusers, ( 1* 292

t I 'I ~

CONCENTRATION SATURATION i 4/-MEAN i I i I5•1 I AI

1 o (960-71

NOV DEC JUN JUL AUG SEP OCT JAN FEB MAR APR MAY in the Turbine Discharge from FIGURE 2. Dissolved Oxygen Conditions Fort Patrick Henry Dam problem also compounds a water pollution by the State of Tennessee and that exists downstream from Kingsport. in re- methods for increasing DO levels After an evaluation of various air aeration in the tailrace (5), it was concluded that diffused servoir releases for are the most promising methods or oxygen injection into the releases aeration Patrick Henry Dam. Only those correcting this problem at Fort of the turbine releases not increase the temperature that would from the SBanCmethods of the cold-water fishery downstream could be considered because reser- eliminated destratification of the dam. This consideration automatically might be more economical. voir and possibly other methods that investigation, primarily because injection was selected for further Oxygen five percent of the plant output) and the high power requirements (about of DO by 4,300 of to achieve high concentrations has high initial investment cost required pool level air aeration in sewage treatment aeration. Using data on diffused diffused air estimated that amflow at to tailwater conditions, it was plants (4) and extrapolating to drive 1970. require a minimum of 3,000 horsepower such an installation would used rich has a fine-bubble diffusers such as those the needed compressors and 15,000 each ough tailrace would cover an area plants. The diffusers in the estimates for in- e din sewage treatment 250 feet downstream. Preliminary about 100 feet wide and cost for he dam is to 6 mg/1 indicate the total annual creasing'DO in the turbine release fall is us-. , injection and $220,- between $120,000 for upstream is about oxygen injection ranges diffused air aeration in the tailrace Fort Pat- 000 for downstream injection; was con- ortiPatb- investment cost for oxygen injection $220,000. The lower initial on his prob- of the limited information available sidered especially important because ick Henry aeration of reservoir releases. Full- m and an main sections: (I) Proposed This paper is divided into the following Pat- Available Dif- Fort System, (2) Evaluation of Commercially Scale Oxygenation Conclusions. Lreach of Tests, (4) Discussion, and (5) fusers, (3) Field Oxygenation Ut stream 293 ......

PROPOSED FULL-SCALE OXYGENATION SYSTEM Oxygen can probably best be injected into the turbine rcleases from Fort Patrick Henry Dam by diffusing the oxygen through finc-pore diffusers. a Amberg et al. tested injection of oxygen into a turbine system by means of the perforated sparge ring that encircled the turbine above the runner blade; the maximum absorption efficiency obtained was only 40 percent (1). In proposed system, oxygen would be injected through diffusers placed either intakes, so that the oxygenated water immediately upstream from the turbine OXYC- is drawn directly into the turbines (Figure 3), or immediately downstream ox~ 0 0000oo U F from the "boil" area in the tailrace (Figure 4). The oxygen injection system through the turbines. 0 would only operate when water was discharged 00.01 The most economical location for diffuser injection appears to be upstream of from the turbine intakes. The greater depth of water has the advantages of allowing more contact time between the water and the gas bubbles and of increasing the ratio of the partial pressure of oxygen in the bubbles to that dissolved nitrogen so that oxygen transfers into the water faster than dissolved

nitrogen transfe sure of oxygen t partial pressure HEADWATER dissolved nitrog of these factors thus reducing t' A disadvant- if all the oxygt 0 0 0 possible probic: 0 'o O.HYPOLIMNION turbine and thl 0 0 0 00 * 0 0 0 used to m,: 000- , ° :-OXYGEN0 taps Injecting w:: ...... BUBBLES affect the turi; 0o 00 0 0 0000 lower because injection inclu DIFFUSER: of gas (result oxygen per un injection), wC recirculation, depending on A liquid w Figure 3. Oxygen Injection Upstream from Turbine Intakes

294

0 -1

ases from Fort pore diffusers. by means of a incr blade; the it (I). In the ;placed either Igenatcd water

' downstream OXYGEN BUBBLES ýjection system bines. 000000 0 t0 0 0 0 0 0 00 D 0 be upstream o.oo%. o -oo o.0a 'dvantages of ubbles and of bles to that of Ian dissolved Figure 4. Oxygen Injection Downstream from Turbine Releases

nitrogen transfers into the oxygen gas bubbles. The ratio of the partial pres- sure of oxygen to that of nitrogen increases with increased depth because the partial pressure of oxygen in a bubble increases with depth while that for dissolved nitrogen generally remains the same at all depths. The combination of these factors allows more gas to be injected through a unit area of diffuser, thus reducing the size of the required diffuser system. 'LINE A disadvantage of upstream injection is that several problems may develop if all the oxygen is not absorbed by the time it reaches the turbines. These IMNION possible problems include a reduction in turbine efficiency, corrosion of the turbine and the associated discharge system, and adverse effects on pressure taps used to measure turbine discharge. )XYGEN Injecting oxygen through diffusers downstream from the dam would not BUBBLES., affect the turbine discharge system, but oxygen absorption efficiency may be lower because of the shallower water. Other disadvantages of downstream injection include the need for a larger diffuser system to inject larger volumes of gas (resulting from both lower absorption efficiency and less mass of oxygen per unit volume of gas, because of less pressure compared to upstream injection), waste of oxygen to the atmosphere or the need for capture and recirculation, and perhaps a somewhat greater consumption of electric power depending on the number of vaporization units required. A liquid oxygen storage tank would be the source of oxygen gas because

295 An

0 0 Year 0

.- JI 1960 1961

0 3962 1.9634 .1965

MAY JUN JUL AUG SEP OCT NOV DEC 3.966 106 FIGuRE 5. Average Monthly Oxygen Requirements at Fort Patrick Henry Dam (Based on 1960-1970 Records).. 3196T

the oxygen requirement is seasonal, highly variable, and relatively low. The monthly demand for oxygen during an average year ranges from about 2:970 pounds to 1.1 million pounds, as shown in Figure 5. Oxygen would 25,000 Averages normally be required only six months of each year. The annual oxygen re- quirement ranges from 3.5 to 6.3 million pounds per year to achieve a DO concentration of 6 milligrams per liter and from 1.7 to 4.1 million pounds per Diffusers, year to obtain a concentration of 5 milligrams per liter (Table 1). silt and othe riods when, EVALUATION OF COMMERCIALLY AVAILABLE DIFFUSERS on the diffu, into tur- Since only limited information is available on injecting oxygen pounds in tF bine releases by means of diffusers, special studies are required before design- when water V. ing a full-scale injection system. TVA is conducting these studies in two phases. possible pro In Phase I, commercially available oxygen diffusers are being test.ed in the was injectec field and in the laboratory to select a diffuser for Phase II. Diffuser selection Figure 6) will be based mainly on operation and maintenance considerations, oxygen characterist transfer characteristics, and cost. For Phase II, which is discussed in the next in pressure main section, a large-scale diffuser system capable of increasing the dissolved uated perke oxygen concentration 2-3 milligrams per liter in the releases from one turbine ging proble changes in will be installed at the dam and tested.

296

a - ---- ,------.-.------.- - .-.--. -

T -- TABLE I Annual Oxygen Requirements for Turbine Discharges from Fort Patrick Henry Dam-1960-1970

Oxcygen Required to Provide the Following Year Mnlmium ownstrewn Disrolved Cb-ygen Concentrations, Ibs/yr- 5 reil 6 ; 1l 6 x 106 X 1o

1960 3.51 5.73 1961 2.26 4.50 3962 3.20 5. 15

1963 1964 3.13 5.33 1965 2.32 3.86 EC 1966 3.68 5.58 Dam 1967 2.64 14.05 .1968 . ow. The 1969 2.57 4.25 m about :n would 1970 2.42 14.146 4.8o ygen re- Aver•ges 2.86 ve a DO unds per -Operation and Maintenance Tests Diffusers can be clogged or plugged by one or more of the following: (1) silt and other solid material settling onto the diffusers, especially during pe- riods when oxygen is not being injected, (2) biological organisms growing reduced ions or chemical com- into tur- on the diffusers, (3) oxidation of chemically e design- pounds in the water on the diffusers, and (4) the diffusers behaving as filters system during idle periods. To evaluate these o,phases., when water enters the diffuser field study was conducted in which oxygen :d in the possible problems, a small-scale was injected for various lengths of time through 14 diffusers (Table 2 and selection Figure 6) located on the bottom of the reservoir (Figure 7). Clogging oxygen characteristics of the various diffusers were evaluated on the basis of change the next in pressure drop across each diffuser (Figure 8). The' diffusers were eval- fissolved uated periodically during an eight-month period, and no apparent plug- turbine ging problems developed. Several of these diffusers will be tested for possible changes in transfer efficiency. Sediment has accumulated on the diffusers

297

. . - - q MI TAuLE 2

Diffusers Evaluated Under Reservoir Conditions ForPossible AMaintenance Problems Width or wall Pore Testing Size Diffuser Length OD Thickness TIMs Xiffuser (daV3) Number Dercription (in) (in) 246 1 Steel Disk 8-3/4 2/8

2 Steel Disk 8-3/4 1/8 10 152 2146 3 Alumina 6 2-9/16 11/16 14 15-r-20 Cylinder 27-50 2.6 4 CEVIAN-N 20 2-5/8 1/2 Plastic" 152 5 Alumina 24 2-1/2 2/16 60 Cylinder

6 Alumina 10-1/8 10 1 60 75 Plate

7 Alumina 10-1/4 3.0-1/2 31 164 Plate 8 Alumina 24. 2-1/2 n1/16 240 11 Cylinder 34 9 Aluminum Alloy 12-1/2 2-2/2. - Plate

10 Aluminum Alloy 12-1/2 2-1/2 - . -1/2-2 164 Plate 122 21. Aluminum Alloy 12-1/2 2-1/2 2-3 Plate 1-:1/2-2 142 :12 Aluminum Alloy 38 2-1/2 Plate Figure 6. Ph: 23 Alumina 6 2-1/2 n/6 5-10 924 scribed in Tabh Cylinder 8-15 924 A4 Alumina 6 2-3/2 n/16 (Figure 9) t1 Cylinder tested interm been removed by applying a high gas flow rate hours. The o: 6.. several times but has always the eight ard cubic fee to the diffusers, In addition, no plugging or clogging resulted when the same as tl diffusers having the smallest pores were left idle for two weeks. are amou The water quality conditions under which the diffusers were evaluated The will be con- justed so thal summarized in Table 3. Field testing of these diffusers probably 7.4 milligrav tinued during the summer and fall of 1973. tank by addi To supplement the field tests for evaluating diffuser plugging that might oxygen oxidation of be caused by oxidation of chemically reduced ions and compounds, of water ard. .was injected through the smallest pore diffuser (Table 2) in a tank

298

f

MRvwI .mIV. 4 _ wf w " ~U ~

Testing Time

22.6 152

24.6

26.6

152

75

75

21

64.

.2 164

X21

-2 V22 those diffusers de- (Numbers on diffusers refer to Figure 6. Phase I Diffuser System scribed in Table 2) did not clog after being contained ferrous iron. The diffuser (Figure 9) that for a total of 317.5 8 hours a day for about 2 months tested intermittently the test was 0.093 stand- flow rate per diffuser unit during ;as flow rate, hours. The oxygen which is about per square foot of diffuser surface, the eight ard cubic feet per minute ten the manufacturer. same as that recommended by the added to the tank were ad- of ferrous iron and fresh water valuated are The amounts were held at about 1.6 and ferrous and total iron concentrations will be con- justed so that about 7 was maintained in the per liter, respectively. A pH of 7.4 milligrams the acid formed from sufficient phosphate buffer to neutralize ; that might tank by adding the ferrous iron stand- iron and the acid used to preserve mnds, oxygen oxidation of ferrous ank of water ard. 299 BARGE (APPROX. 10' x 22') WITH WINCH AND BOOM wVal

TI.

3-a'-72 SALO

*.2v-72 9: -V

5.n-72 2KOC

.30-72 a:1! 6-6-72 6,33 6-13-72 10Mo,: 6.20-72 6•47-72 86? 6-27-?2 9:5'

Patrick Henry Dam Aeration Project General Lay-Out Figure 7. Fort :UAI 10:1 8-k-72 7-U 11 NO.12 7-18-72

20: -DIFFUSER NO.11 30 IxJ 7-25-7n I- 6-1-72

_J 20 6-53-72

Ir1- 8_W272 NO.14' NO. I ILI

300 100 200 400 OXYGEN FLUX, SCFH/FT" pressure drop and oxygen flow rate per unit - - FIGURE 8. Graphical relationships between I'. tests area of diffuser for several of the diffusers tested during the field maintenance (Refer to Table 2 for descriptions of the diffuser numbers.)

10-10-'Wi Although the pressure drop across the diffuser did not change significantly, 10-17/-72 "C an iron plate formed on the diffuser surface. Tests are planned to determine 10-29J-l7 whether this plating will affect the transfer efficicncy of the diffuser. A possi- *I~remts? ble solution to this plating problem may be the periodic injection of hydrogen chlorine gas for dissolving the iron on the diffuser. 300

", "., ý-' .1 ".-1. I 'Wpv' x q ý . ý-'- - 1 NV.

TADLE 3 WITFI Henry Reservoir Water Quality Conditionsin Fort Patrick (South Fork Mile 8.2)

j . __ I••r- :: 20*0. 02 07- at,.. V.ft 7.2

.... - 1%.0 7.3 0.3,9 0.2. 5-2%.W 6:40 0.10 00.01 0.7 0••.10 ?.SJ -.1 0.!7! C.;' 9:20 .... 1 29.0 12.6 0.i3 0o.,o 0.02. O.7 0.0t. 0.12. 9-tl-12 11.2 3.6 00.01, 0.01 0.01. 0.63 8.8 0.21 0.64 0.01 0.a 10.2 o.'0l o0.0 0.0. 0.09 3.5 0.17 0.18 0.01 0.63j 0.33 0.64 0.UL S13.L 7.28.7 2.7 0.16 0.13 0.01. 12.0 11:00 S... 71 61.9 0.13 0.30 1.1 0.2? 0.13• 0.01 0.71 0.64 0.t 20 p... 13.0 1.8 0.,1 0.13 0.01 0.71, 9-2•1-7•2 3.1 "-23-72 2,01 p.M. * 13.0 T.A 12.0 6.7 9-30-72 8C1I s. 0.0 -U 6,:30 5... 15.0 114.0 7.8 4-13-72 303 *... * 6.9 6-20-72 * 15.0 0..7 0.03 0.06 4-27-2/2 8:.'o S... * - 5.0 0.16 0.15 0.01 0.03 0.10 o.14 0.1 o0.0 o0.72 6.3 1.k 6-272 9:0 .5 . 1 17.6 0.72 0.31"• 0.11I 10 1 .7 1.2 0.20 0.11 0.72 00 14.3 1.9 1.1 0. 0C.3 0.09 0.13 0.10 0.71 0.60 16 3 1. 0.01 0.03 0.C6 0.o 8.:7 <0.0 0.16 8.6 1.0 12,05.. 7'1 13.7 1".0 6.9 3-%-72 10:.3 *..* * h.,. 6.2 7-11-72 21A58 a... 15.0 0.02 0.64 • .•a.. 13.0 3.6 a.01 0.0 o0.11 7-18-72 8:.3 0.!6 0.00 0.09 a.13 0o. O.f0 Me.6 22.1 11.9 6.6 0.9:1 0.3 0.*4 t-23-72 10,20 a... 1 0.2. 0.0. 0.00, 15.3 5.1 ?.2 0 0. 0.9 s 0910 10 L.9 2.5 0.29 0.0 0.09 20 13.0is .0 0.26 0.09 0.00 1. 2.7 0.0 1.1 60. 0.1h 2.3 0.93. 0.0 0.13 0.0. 0.01 1.1 0.03 35 1'..6 4.0 2.3 0.12 .0 .1 1.6 2.6 9.32 ,0-04Om 1.1 12300 VA.. 65 1%.5 12.0 h.% 7"1-T72 8:135... * 16.0 8.7 6-1-72 9:00 :.9 64--2 1.151 s.c. 17.0 1.3 1o. 6-132 8.45 as. * 16.0 11-22-72 8,O .ac. * 17.0 81.9 0.C3 0.72 0.02 0.11 2.6 3 0.CA <0.01 0.C1 0.09 1120 cs.. 202.1 -. 01 . 0.01 0.94 6-29-72 11.9 1.6 1* 0.!4 0.97 0.0C 0.10 10 17.3 1.3) 41 0.39 <0.01 0.01 17.3 0.01 0.9? 0.cc 0.0! 20 1.3 S 0. ý 0.01 0.03 80 16.1 .01. 0.01 0.10 0.08 16.0 1.0 6 0.:20 1 0.05 0.17 0 13 0.29 0.02 003 13.3 -a.0 0.03 1.0 0.0) 0.23 G0 1.1 34 0.32 0.02 31.00 #A. T0 13.3 16.0 9-9-73 12.30 P-c. 400 8, s A.- 16.0 9-12-72 3.6 216.3 9-29-7Z 9:30 0s.. L.7 17.0 1. 2 9-26-72 9:30 cc. 8.32 30-3-72 2•3,0 p.-. * 10.0 unit' <0.01 0.CA 0.-6 C."2 rate per 10 O.O 0.1.2 9,30• .•. 1 21.3 0.22 G0.C1 0.02 0.;) 20-3-72 16.7 1.7 10 o.92 0.11 mnance tests' ` 1o 8.1 7 0.20 0.01 0.02 80 13 .9 1.62.0 o.C0 0.9! 0.12 1.8 a 0.22 00.0l 3.0 7 0.1 00.0 0.1 .9 02.03 0.01 0.* 1.2 10 0.12 0.90 0.12 l 0 10 0 ý.00.01 0.01 60 153.135. 1.3 0.01. 0.2 0.17 1.6 Is 0.20 0o.o0 7.09Y..- 70 13.3 1.0 ,nificantly, 101072 6,13 .. 17.0 determine 10-17-72 0.00.s.. * :r. A possi- 10-5-7.2

Ihydrogen 03trlon 1. OtN f- lOU4. 4V1tftdt. V6.I08. 0.. " w/I

301

I

t

m ~ p'0~Y3~9Sq.IqI ~. -.-~v- ~ ¶U59'098.¶~' 2 8 0;~, ~ - ,~'.644 -. q73O~.~ .....

TABLE 3 (continued) Water Quality Conditionsin Fort Patrick Henry Reservoir (South Fork Holston River Mile 8.2)

(00733 ~t '~5i" ~

C,80.31 0.1o 33..0,, 73773. :2 ~ a.m onc ao.• 312 7I :IN/ 137 p 22.0., 0.07 .73.03 7.0 31 07 INO OVERFLO.' 73 05 0.0 0.%1 88 05 0.37.0 0.00.3 1C.a.0.31 1030 3-30.73 0,3w o...* 8~l.42 08,3 e..t.. 1: 33oxIN 4-33.7310.3

*.A •.0 35 73.0. 73.o050.0 73.737.0. •07 5.5. .3 82 79 .6.4 0.01 0.3io T.,. 00 3 o.oW •17 13 13 20 8-17-72 8.30.. SIMo . U* Oo1 0.7,73.•7- 0.01 0.1t2 88.1 .0 .5 7. 230 149 7. 042 .04 5. S O.aWl0,.i. OA.7.0 29. V-187. SAS APO 3 11 0.05 am 3 9373.0. 1 l.9 8.9I -r0 r- 9.1oN IS• ." u 8.0 IT 1 4.10 0 8.5 01 310 6.35.41 3949o~. 73 13 64."07 I'm0.. 73o.7 o.3,0 -1.07% SIM0.. I ••i,.050 0.73, o. .m.€0.20 7306.:d o 3.0) 00.07 , .o•. 0.20. duicting labý o.09 0.09 73.7 0.0 0.-a o. .1.• ý.c the diffuser 0.. 0.31 73.73 P3"`3 160a 310 • oxygen abs, 0).05O.AS0.-1 73).01'a.7 7.3 160 10 3 MS."37 SIM I-.. set conditic: 8.0 O.Or oo S.)."3 3.730P.0. D.3 In, 0..3 ciency inch.

-3. 0.0. *.00.12.30 0.51 7.0+o centration c 89-21-7t M,1OO . 04'o.a 0.08 tank, and i\ 0.9 0.,0 7.731 7.0 03 Transfer 00 .3 0.73 0•.05 23 0. IN3 00 1.3 05 3'" 073 23 12. 073 01 O..73 O.!S 0.0 a I" 73 I 1-1 01 14 05 IS 8. .7 .61 T

where (Ci) ively, for th weight of 0w Oxygen Transfer Efficiency Tests" The oxygen transfer efficiency of various diffusers depends mainly on pore Tr: size, uniformity of pore size, spacing of pores, and gas flow rate per unit of diffuser. The theoretical aspects of gas transfer have been discussed by nu- where (dc/ merous investigators (2) (3) (6) but information is not available for predict- developed 1 ing the aeration transfer efficiencies of various commercially available dif- fusers having pore sizes in the range desirable for use with oxygen. Traz The oxygen transfer efficiency for diffusers can best be determined by con-

302 .'L ý boftasý N.',

Figure 9. Chemical Plugging Test Tank

ducting laboratory studies using a tank to simulate the conditions under which the diffusers will be used. The transfer efficiency is defined as the ratio of the oxygen absorbed by the water to the oxygen supplied through the diffuser for set conditions of DO and temperature. Other factors that affect transfer effi- ciency include height of bubble rise, gas flow rate through. the diffuser, con- centration of dissolved nitrogen, geometry and hydraulic characteristics of the tank, and water quality conditions in the tank. Transfer efficiency may be determined by the following three techniques:

absorbed _ (C 2 -C1 )Vt Transfer Efficiency =- total oxygen (1) total oxygen supplied NV concentrations, respect- where (Ci) and (C 2 ) are the initial and final DO ively, for the test; (Vt) is the volume of water in the tank; and (W) is the weight of oxygen supplied. absorption _ (dc/dt)V, Transfer Efficiency = rate of oxygen (2) y on pore oxygen supplied W/t ,r unit of rate of Ad by nu- where (dc/dt) is the rate of oxygen absorption determined by methodology r predict- developed by Ippen and Carver (2), and (t) is the period of oxygen supply. lable dif- - 1 -- Vp Transfer Efficiency =1 - total oxygen unabsorbed (3) total oxygen supplied W d by con- 303

J 2 ......

where (Vr) is the volume of oxygcn gas collectcd at the top of the water column, and (p) is the gas density. The first technique is easiest to use but the results are applicable only to those situations where the initial and final DO concentrations are the same LI as those for the original tests. The applicability of this technique is cspecially limited for air aeration, because transfer efficiency is very dependent on DO concentrations; however, when oxygen is used in place of air, this technique is applicable over a wvidcr range of DO concentrations. The second technique is perhaps the most difficult to use in oxygenation tests because it requires measuring the oxygen uptake rate, which in turn requires measuring the DO in the tank during the test. Measuring DO in the WORK PL tank is difficult if oxygenation occurs in only part of the tank and if this ox- ygenated water is not mixed with that in the remainder of the tank. The ad- vantage of this technique, however, is that the results are applicable for all DO concentrations within the range tested. The third technique requires collecting gas at the top of the water column VOLUM and analyzing it for oxygen content. This technique should be used in com- 12,400, bination wvith the first to provide a check on the results. An important ad- 46,8001 vantage of this technique for oxygenation tests is that the amount of dissolved nitrogen stripped from the water during testing may be determined easily without the need for on-site instruments for continual dissolved nitrogen anal- ysis. One disadvantage of this technique as applied in the past is that the re- sults are applicable only in the same manner as those from the first technique because only initial and final measurements are taken, and therefore are not applicable for intermediate concentrations of DO. To compare the relative oxygen transfer efficiencies for various commercial diffusers, a tank (Figure 10) was constructed at TVA's Engineering Labor- atory in Norris, Tennessee. The tank was designed for a 45-foot depth to sim- ulate approximately the minimum height of the hypolimnion in Fort Patrick Henry Reservoir. A diameter of 7 feet was selected so that the volume of water in the tank would be sufficient to limit the rate of DO increase during testing to about 1 milligram per liter per minute so the rate of DO increase might be measured. The size of this tank should also reduce effects of the wall of the Figurn tank and thus possibly reduce the amount of error in scaling up results. (This deter The tank is equipped with a pump for mixing the contents, a filter for clean- the results, ing the water, a nitrogen gas injection system for adjusting the concentration of dissolved nitrogen in the water before tests, and a gas collector that may be tested.) adjusted to various heights of the water column. The gas collector system The am( allows continuous measurement and analysis of unabsorbed gas while a test urements fi is in progress which may allow the determination of rate of oxygen absorption. used from

304

I ic water COLLECTION AND ANALYSIS OF only to UNABSORBED GAS hc same specially t on DO hnique is genation i in turn OXYGEN )0 in the WORK PLATFORI f this ox- SAMPLING The ad- LOCATIONS ,le for all r column VOLUME I in com- 12,400 GALS. OR rtant ad- 46,800 LITERS- dissolved Led easily gen anal- at the re- ,echniquc •e are not mmercial ig Labor- WINDOWS- ,th to sim-, rt Patrick e of water ng testing Smight be Figure 10. Tank for Measuring Aeration Transfer Efficiencies of Diffusers rall of the4 Ilts. above, and (This determination would allow use of the second equation noted for clean- the range the results would be applicable to all concentrations of DO within centration at may be tested.) can be measured by integrating meas- tor system The amount of oxygen gas injected and by weighing the amount of oxygen bhile a test urements from the orifice flow meter each test, DO is reduced to bsorption. used from a small medical supply bottle. Before

305

. . * -l.. rpm""______o- ......

-~w'- '~-~

a cobalt ducted, anid about 0.5 mg/l by chemical reduction (sodium sulfite is used with will be dct( catalyst). dif- of the rates Much of this equipment was just recently added to the tank and final tests, however, Oxygen trai fuser evaluations have not yet been conducted. Preliminary determining a reasonab! have been conducted on six diffusers by the first technique for injection 45-foot dep transfer cfficiency without the benefit of the gas collector, nitrogen gas in- phere withii system, and weight measurement system for the amount of oxygen each test was allow the g: jection. The approximate final DO in the tank water following measure- which possi' 6.0 mg/l . Another limitation to these results is that gas flow rate prelim- the draft tul ment at very low rates (<0.2 scfm) apparently was in error. These rise the trans- inary tests, however, indicate that for a 42-foot height of bubble tested. fer efficiency is essentially the same (Figure 11 ) for all the diffusers 100 per- Since hil In general, the results indicate that transfer efficiencies approaching is 50 to selection of cent are possible at low gas flow rates and the minimum efficiency bubble size nance, and 60 percent for high gas flow rates. The results also indicate that rise and that installation is not significant within the range tested for a 42-foot height of is forming Engineerin! a significant amount of oxygen transfer occurs while the gas bubble through th, on the diffuser especially at low gas flow rates. be con- velocities, ( After all equipment has been installed, more accurate tests will fuser, and other plugg DESCRIPTIONS 1.4F DIFFUSER tion and m (D ALUMINUM ALLOY PLATE I-2U.- of idleness, ( ALUMINUM ALLOY PLATE 2-3M fuser. The IiJ 1.2 )ALUMINUM ALLOY PLATE 3-4, - gated becat U_ @ALUMINA TUBE 5-•IOI- to be levelei U.W I - . Other tL.cc: 5ALUMINA TUBE 8-i5M cc 1.0 ease of b1.. @ALUMINA TUBE •15--20OM.- and diffusers be up to the n z fuser No. 6 the san I- cost clean becat: 0 N is not as f

I InAt I" 2 3 4 5 6 7 O After a ( OXYGEN FLOW-SCFM .capable of (Diffuser num- Figure 11. Preliminary Results of Oxygen Transfer Efficiency Tests numbers liter in the bers 1-3 were plates with a surface area of about 0.2 square feet and diffuser I foot.) Phase II. P 4-5 were tubes with an outside diameter of 22 inches and a length of

306

-P 00IPCP S," ., - .- . W -ý -. --nA*'5

Oxygen transfer efficiency ducted, and the preliminary tcsts will be chccl:cd. t cobalt if possible, the ratio will be dectermined by the gas balance technique and, technique will be used. of the rates of oxygen absorption to oxygen supply inal dif- that then will be conducted at various depths to insure owever, Oxygen transfer tests oxygen absorption is attained within the entire rmining a reasonable distribution of supersaturation of oxygen with respect to the atmos- njection 45-foot depth to avoid 20 feet of bubble rise. Such supersaturation might gas in- phere within the first 10 to from solution aftcr passing through the turbines test was allow the gas to be stripped loss of oxygen and increased rates of corrosion of ýieasure- which possibly could cause *prelim- the draft tube. he trans- Other Considerationsfor Selecting Diffusers s tested. for several diffusers, Since high transfer efficiencies have been measured 100 per- system costs, mainte- selection of a diffuser probably will be based on total is 50 to Several possible, nance, and possible problems of installation and operation. bble size in the tank at the installation and operational problems will be evaluated and that to obtain gas flow Engineering Laboratory to determine (1) time required forming (2) effect of horizontal through the diffusers following periods of idleness, of inclination of the dif- velocities, (3) effect on efficiency of a small degree [ be con- on the diffuser and fuser, and (4) the effect on efficiency of iron oxidation during the field opera- other plugging that may have occurred on the diffusers diffuser during periods tion and maintenance tests. When water enters the pass through the dif- of idleness, it may have to be displaced before gas can will be investi- fuser. The effect of diffuser inclination on transfer efficiency if the system has gated because diffuser system installation might be expensive to be levcled. should include cost Other considerations for choosing a particular diffuser yet available for all the and ease of cleaning. Although cost estimates are not increasing pore size diffusers being evaluated, cost apparently decreases with transfer efficiency (Dif- up to the maximum pore size thus far evaluated for larger pores apparently fuser No. 6 in Figure 11); after that, diffusers with to be the easiest to cost the same regardless of pore size. Tube diffusers appear made of CEVIAN- •clean because they can be dismantled. The plastic diffuser N is not as fragile as the alumina tube diffusers.

FIELD OXYGENATION TESTS 7 diffuser system, After a diffuser has been selected in Phase I, a large-scale 2-3 milligrams per capable of increasing the dissolved oxygen concentration Ifuser num- in the reservoir for liter in the releases from one turbine, will be installed ser numbers and fall of 1973. The Phase II. Phase II will be conducted during the summer 307

i

-'.-- P77 '- P". -- U...... - - - Oxyge of a diffuser system that can be moved field installation probably will consist (or peaki of a barge similar to that used in Phase I (Fig- to various locations by means more ecoi 7). ure for maini I I includes the following: Work to be done in Phase rates are: for injecting the oxygen, i.e., in front of the 1. Determine the best location air destr. the tailrace. It is important to deterniine the best loca- turbine intake or in system si. from the turbines to avoid bubbles entering tion for the diffusers upstream jecting o If bubbles enter the turbine system and affect power the turbine system. techniqu, problems develop with upstream injection, operations, or other unforeseen mediatel, injection will be evaluated. downstream plementa transfer efficiencies under field conditions. 2. Evaluate oxygen system is with oxygenation by this method, such as 3. Evaluate problems associated to investi turbine discharge system or reduction in power production corrosion of the clog the, efficiency. used to c for use in designing a full-scale demonstra- 4. Obtain additional information a combir tion project. for diffu a full-scale installation if results obtained show this 5. Schematically design A con: technically and economically. method of reaeration to be feasible ising tecl voirs wil DiscussIoN standard alternatives, ap- vast sum: Oxygen injection through diffusers, compared with other in the pears to be an economical method of increasing DO concentrations must not be turbine releases from reservoirs where downstream temperatures results of this study indicate (1) maintenance and increased. Preliminary 1. After, is not a major problem in that clog- operation of the diffusers apparently investi diffusers has not occurred during testing, (2) ap- ging or plugging of the diffus absorption of the oxygen injected can be attained proximately 100 percent for co and in sufficient depths, and (3) oxygen injection is at at low gas flow rates was sc air aeration in the tailrace. least as economical as diffused requir appears to be the least cost alternative for Fort Pat- Although this method 2. Upstr to note that this method is not the most eco- rick Henry Dam, it is important injecti Each dam needs to be investigated for its specific physical nomical in all cases. tem, d (5). For dams where there is a large an- conditions and DO requirements 3. Liqui, diffused air aeration with its high nual requirement for addition of DO, on-sito more economical because of low operational costs. Oxy- capital cost may be able,*• operational cost but lower capital cost, is more gen injection, with high 4. Fourt requirements are relatively small. Long-term op- economical where oxygen Hlemr, techniques of aeration probably will be eration of several very promising and 11 necessary to determine the economics of such systems.

308 ~r '½1~--~

as a supplementary Oxygen injection into turbine releases also may be used moved methods that may be (or peaking) system in conjunction with other aeration I (Fig- not be dependable more economical for normal turbine releases but that may turbine release for maintaining a minimum DO concentration when higher be used with diffused rates are necessary. For example, oxygen injection could it of the a separate diffuser air destratification either by injecting oxygen through est loca- Dam or by in- system similar to the one proposed for Fort Patrick Henry entering The latter jecting oxygen through the diffusers used for destratification. :t power volume of water im- technique would increase the DO throughout a large ijcction, the cost of a sup- mediately upstream from the dam and would not require a separate diffuser plementary diffuser system for peak discharges. Before it might be necessary system is used for oxygen injection in a peaking system, such as growths might to investigate the possibility that sedimentation and biological :xduction might also be clog the diffusers if left idle over long periods of time. Oxygen in the tailrace. Such used to enrich the gas mixture for diffused air aeration monstra- capital cost required a combination of. these two techniques could reduce the for diffused air aeration. how this the more prom- A considerable amount of research is needed before even scale. Many reser- ising techniques should be applied to reservoirs on a large state water quality voirs will need such systems in the near future to meet investigated, standards, but unless the possible aeration methods are properly tives, ap- vast sums of money may be spent unwisely. ns in the ist not be CONCLUSIONS ince and releases had been 1. After various methods of increasing DO levels in reservoir hat clog- in the tailrace or investigated, TVA concluded that diffused air aeration (2) ap- methods diffused oxygen injection into the releases are the most promising attained Oxygen injection for correcting the problem at Fort Patrick Henry Dam. :tion is at of the high power. was selected for further investigation primarily because requirement and high initial cost of diffused air aeration. Fort Pat- than downstream 2. Upstream diffuser injection probably is more economical most eco- , sys- however, if upstream injection adversely affects the turbine c physical :" injection; tem, downstream injection may be more economical. large an- economical than 3. Liquid oxygen storage tanks probably would be more i its high highly vari- oxygen generation because the oxygen requirement is osts. Oxy- on-site able, seasonal, and relatively low. :, is more in Fort Patrick 4. Fourteen commercially available diffusers were operated ;-term op- period, Henry Reservoir for various lengths of time during an eight-month ly will be and no apparent clogging or plugging occurred.

309

SW ,-;~"~ 7 5. The diffuser with the smallest pore size that is commercially available did not clog after being operated for ovcr 300 hours in water that contained a relatively high concentration of ferrous iron. a STRI 6. To compare the relative oxygen transfer efficiencies for various diffusers, 45-foot high tank was constructed. Although installation of auxiliary equipment is not yet complete, preliminary tests indicate that high absorp- tion efficiencies can be attained at a low gas flow rate per unit of diffuser. costs, 7. Selection of a diffuser should include considerations for total system and possible installation problems, operation and maintenance problems, transfer efficiency. 8. Large-scale field studies are proposed for 1973 in which the best location will be determined for an oxygen injection system. • ~Pilot

REFERENCES years ha A. S. Rosenfelt, and R. 0. Blosser. "Re- Oxygeimolecuk 1. Amberg, H. R., L. F. Cormack, W. Funk, 4, 15 Oxygen," Industrial Water Engineering, aeration of Streams with Molecular ciency ul 2.(February 1967). Transfer in side strei 2. Ippen, Arthur T. and Charles E. Carver, Jr. "Basic Factors of Oxygen n Aeration Systems," Sewage and Industrial IWastes, 26, 813 (July 1954). efficienci Aeration Tanks," r 3. King, Henry R. "Mechanics of Oxygen Absorption in Spiral Flow nver tfl 1955). Sewage and Industrial Wastes, 27, 894 (August water b paper presented at the Sanitary Engi- 4. Nicholas, IV. R. "Aeration System Design," March 12, 1965. neering Institute, University of Wisconsin, Madison, Wisconsin, Dissolved Oxygen Concentra- reaerati 5. Ruane, R. J. "Investigations of Methods to Increase Environ- conditi tions Downstream from Reservoirs," Proceedings of the Eleventh Annual Nashville, Tennessee: ditions. mental and Water Resources Engineering Conference. mate Vanderbilt University, June 1972 (In preparation). Aeration," of D 6. Speece, R. E. "The Use of Pure Oxygen in River and Impoundment In. Proceedings of the 24th Industrial Water Conference-Part One. Lafayette, diana: Purdue University, May 1969.

iCr L .drou ing t gene (D ma bas ýexp cat tre

310

t.